| *** woleium has quit IRC | 00:01 | |
| *** ramd has quit IRC | 00:03 | |
| *** maplebed has quit IRC | 00:07 | |
| *** MarkAtwood has joined #openstack | 00:10 | |
| *** Dumfries has quit IRC | 00:14 | |
| *** adjohn has joined #openstack | 00:14 | |
| *** adjohn has quit IRC | 00:19 | |
| *** adjohn has joined #openstack | 00:20 | |
| *** MarkAtwood has quit IRC | 00:23 | |
| jaypipes | vishy: still around? | 00:23 |
|---|---|---|
| vishy | jaypipes: yup | 00:24 |
| *** mahadev has quit IRC | 00:25 | |
| jaypipes | vishy: got a sec to discuss a path for making the integration of Nova(-objectstore) and Glance cleaner? | 00:25 |
| jaypipes | vishy: specifically re: the issues raised in, for example, https://answers.launchpad.net/nova/+question/155250 | 00:26 |
| vishy | yes | 00:26 |
| *** thingee has joined #openstack | 00:28 | |
| vishy | ah yes that is an issue | 00:28 |
| vishy | be nice to have some sort of clean cache command | 00:28 |
| *** dragondm has quit IRC | 00:28 | |
| vishy | alternatively we could cache based on name_checksum or something like that | 00:28 |
| jaypipes | vishy: well, I'm a bit unclear as to what the root of the problem really is.. | 00:29 |
| vishy | jaypipes: images are cached on compute hosts | 00:29 |
| vishy | jaypipes: based on the image id | 00:29 |
| jaypipes | vishy: I'm wondering if, once a euca-bundle-image completes, perhaps we should delete the image parts from the nova-objectstore? | 00:30 |
| vishy | jaypipes: doesn't really have anything to do with euca / ec2 | 00:30 |
| jaypipes | oh.. | 00:30 |
| *** gregp76 has quit IRC | 00:30 | |
| jaypipes | vishy: sorry, I thought this was cached in nova-objectstore. :( | 00:30 |
| vishy | jaypipes: there is no way for the compute host to know that you uploaded a new image and reused the id | 00:30 |
| jaypipes | vishy: right... one more case for using UUIDs for identifying images ;) | 00:31 |
| vishy | agreed | 00:32 |
| jaypipes | vishy: is there a way to make the errors coming back from the virt layer (qemu-ndb, etc) more descriptive of what's actually going on? | 00:32 |
| *** zenmatt has quit IRC | 00:33 | |
| vishy | they usually are | 00:33 |
| vishy | just the error log also says something like "Is virtualization enabled in the bios?" | 00:33 |
| vishy | so for some reason everyone ignores the traceback above... | 00:34 |
| *** kashyap has quit IRC | 00:34 | |
| vishy | usually there is a very specific ProcessExecutionError with a traceback just above that line | 00:34 |
| *** nelson has quit IRC | 00:35 | |
| *** nelson has joined #openstack | 00:35 | |
| jaypipes | vishy: like the traceback in this question? https://answers.launchpad.net/glance/+question/154442 | 00:36 |
| *** joearnold has quit IRC | 00:37 | |
| vishy | jaypipes: that might require a little investigation | 00:39 |
| vishy | it seems like there should have been an error raised before that if it is trying to pull a non-existant image | 00:39 |
| *** zenmatt has joined #openstack | 00:39 | |
| vishy | existent | 00:40 |
| vishy | libvirtError: operation failed: failed to retrieve chardev info in qemu with 'info chardev' is definitely not a very useful error message | 00:40 |
| jaypipes | right | 00:40 |
| vishy | and I'm actually surprised it got that far... | 00:41 |
| vishy | it seems like it should have bombed in the glance image service code somewhere | 00:41 |
| *** bblair48 has joined #openstack | 00:42 | |
| jaypipes | vishy: Glance shows the images fine, just as in the queued status, not active... | 00:42 |
| vishy | right, but the download failed somehow right? | 00:43 |
| *** jakedahn has quit IRC | 00:45 | |
| *** jakedahn has joined #openstack | 00:45 | |
| jaypipes | vishy: hmm, looking at the code, there actually shouldn't be a way for the status to be manually set to 'queued'. Glance's API server does that automatically, and should not be possible to override that. I'll look more into it. | 00:45 |
| vishy | i think it happens if you put an image with only metadata and no image data right? | 00:46 |
| Ryan_Lane | it also happens if you put in an image without some of the metadata | 00:46 |
| vishy | and then when you put the image data it is supposed to automatically flip to active? or are you supposed to manually post active? | 00:46 |
| Ryan_Lane | for example, if you use uec-publish-tarball | 00:47 |
| vishy | Ryan_Lane: that is the issue we are discussing | 00:47 |
| jaypipes | vishy: should automatically handle updating status field in all cases. and no caller should be able to manually set it. | 00:47 |
| vishy | jaypipes: 51 metadata = image_service.get(elevated, image_id, image_file) | 00:48 |
| vishy | in the above traceback that call is failing to write image data for some reason | 00:48 |
| vishy | i don't see exactly how that can fail without an exception being thrown | 00:49 |
| vishy | so I'm not totally sure that the traceback is actually a problem with the image/glance | 00:50 |
| *** kashyap has joined #openstack | 00:50 | |
| vishy | jaypipes: the thing is I'd like to be able to display in the ec2 api the state of the image even if it isn't available... | 00:51 |
| jaypipes | vishy: what would you display? | 00:52 |
| vishy | jaypipes: so if we filter out queued images, I have to manually hack around it in painful ways | 00:52 |
| vishy | jaypipes: it shows the current state of the upload / conversion process -> decrypting -> untarring -> available | 00:52 |
| vishy | and allows us to put a state in if something fails | 00:53 |
| vishy | jaypipes: I could make that info only available if you specify the id (which uses show instead of details), but that is different from the way it used to work so it might be confusing to users | 00:53 |
| *** jakedahn has quit IRC | 00:54 | |
| jaypipes | vishy: thinking... | 00:54 |
| vishy | (if there is a way for me to create an image in the active state with just metadata and then send the image data later, that would also work) | 00:56 |
| jaypipes | vishy: there is a way to register an image and send the image data later. You are currently doing that in the S3ImageService. | 00:58 |
| jaypipes | vishy: on line 159: image = self.service.create(context, metadata) | 00:58 |
| jaypipes | vishy: it isn't passing the actual image data at all. | 00:58 |
| vishy | jaypipes: right, but didn't you state above that that will put it in the queued state? | 00:59 |
| *** jakedahn has joined #openstack | 01:00 | |
| jaypipes | vishy: yep. | 01:00 |
| *** obino has quit IRC | 01:00 | |
| jaypipes | vishy: ok, so when we changed glance index to only return active images, we screwed you... | 01:00 |
| vishy | jaypipes: right so that doesn't work because then I can't show the state... | 01:00 |
| vishy | :) | 01:00 |
| jaypipes | vishy: gotcha. OK, so our "fix" was to do this: | 01:01 |
| jaypipes | filter_by(status='active').\ | 01:01 |
| jaypipes | vishy: looks like we need to instead do: | 01:01 |
| jaypipes | filter_by(status is not 'killed').\ | 01:01 |
| vishy | jaypipes: that would work (as long as it doesn't break other people) | 01:02 |
| jaypipes | or: filter_by(status!='killed').\ | 01:02 |
| vishy | i like killed? is it user settable? | 01:02 |
| jaypipes | vishy: well, the original bug that forced that change was that killed images were showing up in glance index, but the status was not shown in glance show, so nobody knew why killed images were showing up :) | 01:02 |
| vishy | (in general i'm a fan of lazy deletes for security reasons) | 01:02 |
| jaypipes | vishy: we never actually delete anything ;) | 01:03 |
| jaypipes | vishy: are you asking for us to also show killed images? | 01:03 |
| jaypipes | in glance index? | 01:03 |
| vishy | jaypipes: :) so I think the status != 'killed' works for me | 01:03 |
| jaypipes | vishy: ah, ok. bcwaldon's work on https://blueprints.launchpad.net/glance/+spec/api-results-filtering will give some more filtering ability too... | 01:04 |
| vishy | jaypipes: no, I just like the idea of glance keeping images around even if they are deleted by a user (very important for nasa for example) | 01:04 |
| jaypipes | vishy: gotcha. yeah, we don't get rid of anything. | 01:04 |
| jaypipes | vishy: OK, I will change the project of this bug to Glance: https://bugs.launchpad.net/nova/+bug/773388 | 01:05 |
| uvirtbot | Launchpad bug 773388 in nova "EC2 images controller not setting image status to 'active' properly" [Undecided,New] | 01:05 |
| vishy | as to the caching problem, UUIDs take care of it so I don't think we really need to worry about a fix for it at the moment | 01:05 |
| *** maplebed has joined #openstack | 01:06 | |
| vishy | jaypipes: I'm also curious why it doesn't go to active when the actual data is uploaded? | 01:06 |
| vishy | jaypipes: perhaps it fails before it actually uploads the data because it can't list the image anymore? | 01:06 |
| jaypipes | vishy: the reason that is happening is because the way that S3ImageService is calling glance.client.add_image() is by specifying only metadata, and then trying to upload an image later with glance.client.update_image(). Only problem is that the call to glance.client.update_image() never occurs... | 01:07 |
| jaypipes | vishy: heh, you beat me to the answer there.. | 01:08 |
| vishy | ok but why isn't occurring? is 177 self.service.update(context, image_id, metadata) failing because it depends on being able to list the image? | 01:08 |
| jaypipes | vishy: prolly worth adding a try: except around the upload and change the Glance image status to killed in a problem? | 01:08 |
| jaypipes | vishy: but then the image won't be listed in glance index and we're back to square one ;) | 01:09 |
| vishy | it should probably update the image_status to "error" at least | 01:09 |
| jaypipes | vishy: line 177 doesn't actually update the image data, though :( | 01:10 |
| jaypipes | vishy: so there's no upload at all... | 01:10 |
| vishy | yes the upload happens later | 01:10 |
| vishy | so it updates the state twice then finally uploads the image | 01:10 |
| vishy | is one of the intermediate updates failing? Am I not allowed to update metadata for a queued image without giving it data? | 01:11 |
| vishy | because it used to work... | 01:11 |
| jaypipes | vishy: until line 199, if an error occurs, the image will still be in queued... | 01:11 |
| jaypipes | vishy: no, you can update metadata just fine without an image file. | 01:11 |
| jaypipes | vishy: but if an error occurs in, say, decrypting, then the image will be stuck in queued status in Glance... | 01:12 |
| vishy | jaypipes: yes traditionally in that case it would show the image status as decrypting or untarring etc. So you could see where it failed. It would definitely be better to update it to say error decrypting or error untarring etc. | 01:12 |
| jaypipes | vishy: so I was suggesting putting a try block around those operations and sending an update to Glance if an error occurs. | 01:12 |
| *** jdurgin has quit IRC | 01:13 | |
| vishy | i think leaving it in queued with a reasonable error message is the clearest for users...agree on try except for sure, but setting it to killed just means it disappears from a users perspective | 01:13 |
| jaypipes | right... | 01:14 |
| jaypipes | vishy: k, I'll get on this fix right away. | 01:14 |
| vishy | but i have that handy image_status field for a more useful error message | 01:14 |
| vishy | jaypipes: cool thanks. I'll put some kind of try, finally in there for better error messages on the nova side | 01:15 |
| *** maplebed has quit IRC | 01:16 | |
| jaypipes | vishy: coolio. | 01:16 |
| *** hallyn has joined #openstack | 01:19 | |
| vishy | jay, submitted bug776825 | 01:20 |
| *** woleium has joined #openstack | 01:21 | |
| *** obino has joined #openstack | 01:22 | |
| *** clauden has quit IRC | 01:28 | |
| *** mahadev has joined #openstack | 01:30 | |
| uvirtbot | New bug: #776825 in nova "When register_image fails, the state of the image isn't updated" [Wishlist,New] https://launchpad.net/bugs/776825 | 01:32 |
| *** erik-s is now known as erik-s|detached | 01:32 | |
| *** mahadev has quit IRC | 01:35 | |
| *** hvaldivia has joined #openstack | 01:37 | |
| *** mahadev has joined #openstack | 01:42 | |
| *** KnightHacker has quit IRC | 01:42 | |
| *** jero_ has joined #openstack | 01:48 | |
| *** jero_ has left #openstack | 01:48 | |
| *** KnightHacker has joined #openstack | 01:50 | |
| *** Ryan_Lane has quit IRC | 01:50 | |
| *** reldan has quit IRC | 01:51 | |
| *** Ryan_Lane has joined #openstack | 01:51 | |
| *** nelson has quit IRC | 01:59 | |
| *** msivanes has quit IRC | 02:02 | |
| *** msivanes has joined #openstack | 02:03 | |
| *** Ryan_Lane has quit IRC | 02:10 | |
| *** eca has joined #openstack | 02:11 | |
| *** eca has left #openstack | 02:16 | |
| *** miclorb_ has quit IRC | 02:16 | |
| *** dmi_ has quit IRC | 02:17 | |
| *** hadrian has quit IRC | 02:20 | |
| *** dmi_ has joined #openstack | 02:28 | |
| *** cloudgroups has joined #openstack | 02:31 | |
| *** reldan has joined #openstack | 02:38 | |
| *** bblair48 has quit IRC | 02:39 | |
| *** dmi_ has quit IRC | 02:43 | |
| *** AimanA has quit IRC | 02:44 | |
| *** AimanA has joined #openstack | 02:45 | |
| *** reldan has quit IRC | 02:45 | |
| *** reldan has joined #openstack | 02:45 | |
| *** reldan has quit IRC | 02:53 | |
| *** rchavik has quit IRC | 03:04 | |
| *** koolhead17 has joined #openstack | 03:10 | |
| *** pLr has joined #openstack | 03:14 | |
| pLr | cool project :) | 03:14 |
| *** MarkAtwood has joined #openstack | 03:15 | |
| *** RoAkSoAx has quit IRC | 03:17 | |
| *** stewart has quit IRC | 03:35 | |
| *** hvaldivia has left #openstack | 03:46 | |
| *** zns has joined #openstack | 03:57 | |
| *** jamiec_ has joined #openstack | 03:59 | |
| zns | What's the status of openstack-common? I'd like to use wsgi.py in Keystone and I see slightly different versions in nova, swift, and glance. Seems like there could be one shared version… any thoughts on which one to use? | 04:00 |
| *** Vek has quit IRC | 04:05 | |
| *** kashyap has quit IRC | 04:06 | |
| *** freeflying has quit IRC | 04:06 | |
| *** kashyap has joined #openstack | 04:06 | |
| *** freeflying has joined #openstack | 04:07 | |
| *** miclorb has joined #openstack | 04:07 | |
| *** stewart has joined #openstack | 04:13 | |
| *** koolhead17 has quit IRC | 04:20 | |
| *** zenmatt has quit IRC | 04:21 | |
| *** zenmatt has joined #openstack | 04:21 | |
| *** khash has joined #openstack | 04:27 | |
| *** nijaba has quit IRC | 04:29 | |
| *** nijaba has joined #openstack | 04:32 | |
| *** nijaba has joined #openstack | 04:32 | |
| *** jakedahn has quit IRC | 04:36 | |
| *** reldan has joined #openstack | 04:37 | |
| *** reldan_ has joined #openstack | 04:43 | |
| *** reldan has quit IRC | 04:43 | |
| *** reldan_ is now known as reldan | 04:43 | |
| *** khash has quit IRC | 04:45 | |
| *** f4m8_ is now known as f4m8 | 04:45 | |
| *** reldan_ has joined #openstack | 04:46 | |
| *** reldan has quit IRC | 04:46 | |
| *** reldan_ is now known as reldan | 04:46 | |
| *** cloudgroups has left #openstack | 04:46 | |
| *** koolhead11|afk is now known as koolhead11 | 04:50 | |
| *** Zangetsue has joined #openstack | 04:58 | |
| *** santhosh has joined #openstack | 05:02 | |
| *** johnson81385 has joined #openstack | 05:03 | |
| *** stewart has quit IRC | 05:06 | |
| *** johnson81385 has quit IRC | 05:07 | |
| *** jakedahn has joined #openstack | 05:08 | |
| *** AimanA has quit IRC | 05:15 | |
| *** AimanA has joined #openstack | 05:15 | |
| *** gregp76 has joined #openstack | 05:16 | |
| *** gregp76 has left #openstack | 05:16 | |
| *** zns has quit IRC | 05:20 | |
| *** MarkAtwood has quit IRC | 05:24 | |
| *** foxtrotgulf has quit IRC | 05:28 | |
| *** zns has joined #openstack | 05:40 | |
| zykes- | wasn't there a goal to host a 2011 event for openstack in Europe ? | 05:42 |
| *** zaitcev has quit IRC | 05:43 | |
| *** JayLeeeee has joined #openstack | 05:48 | |
| zykes- | was there a mail from vish about some post summit stuff ? | 05:49 |
| *** reldan has quit IRC | 05:49 | |
| *** woleium has quit IRC | 05:52 | |
| *** RoAkSoAx has joined #openstack | 05:59 | |
| *** RoAkSoAx has joined #openstack | 05:59 | |
| *** Vek has joined #openstack | 06:01 | |
| *** jamesurquhart has quit IRC | 06:03 | |
| *** stewart has joined #openstack | 06:03 | |
| *** zns has quit IRC | 06:06 | |
| *** zenmatt has quit IRC | 06:09 | |
| *** mahadev has quit IRC | 06:12 | |
| *** reldan has joined #openstack | 06:13 | |
| *** reldan has quit IRC | 06:20 | |
| *** reldan has joined #openstack | 06:20 | |
| *** perestrelka has joined #openstack | 06:29 | |
| *** hagarth has joined #openstack | 06:32 | |
| *** miclorb has quit IRC | 06:37 | |
| *** miclorb_ has joined #openstack | 06:37 | |
| koolhead11 | hi all | 06:42 |
| HugoKuo__ | hello kool | 06:45 |
| HugoKuo__ | how do you think once nova-network dead , could user access instance ? | 06:46 |
| koolhead11 | HugoKuo__, its like asking how will your access internet if someone has chopped off internet wire before it enters your home :D | 06:49 |
| HugoKuo__ | ha | 06:49 |
| HugoKuo__ | I'm doing it now | 06:50 |
| HugoKuo__ | I got an idea , testing | 06:50 |
| HugoKuo__ | maybe I'll focus on "keepalive" later | 06:51 |
| *** fabiand__ has joined #openstack | 07:01 | |
| HugoKuo__ | koolheadl : there r 5 india engineer here from your conutry :) | 07:03 |
| *** carlp_ has quit IRC | 07:15 | |
| koolhead11 | HugoKuo__, awesome!! here as in ? | 07:16 |
| *** carlp_ has joined #openstack | 07:20 | |
| *** daveiw has joined #openstack | 07:25 | |
| *** miclorb_ has quit IRC | 07:31 | |
| pgregory | hi | 07:34 |
| *** thingee has quit IRC | 07:35 | |
| pgregory | in one last ditch attempt to get this working, I completely reinstalled Ubuntu 10.10 Server, and followed to the letter the instructions in the manual I was pointed to yesterday. | 07:35 |
| pgregory | I used the "3.4.2 Scripted Installation" method, using the script on github from elasticdog. | 07:36 |
| pgregory | When asked for input I chose the default IP of the machine for all such questions, for the "Network range for ALL projects" I entered 10.0.0.0/12, amount of IPs 24, "Desired network range for the 'pgregory' project" 10.0.0.0/24, number of networks 1, available IPs 24. | 07:38 |
| pgregory | It seemed to install ok, I read through and skipped the section 10.1.1 as it seemed that the script did all that. | 07:41 |
| pgregory | I did 10.1.2, with a bit of changing, as the instructions don't take into account that as a normal user you can't access the /root/creds/novarc, and the appended copy of that script to ~/.bashrc is nonsense, as the first thing it does is set NOVA_KEY_DIR to the same folder as the script, which when appended to .bashrc will be the users home, which is *not* where the key files are. | 07:43 |
| pgregory | So I had to switch to root, source the novarc and then download and publish the tarball. | 07:43 |
| pgregory | The rest of 10.1.2 is weird, it says "We can realize what state an image is in using the 'euca-describe-instaces<sic>' comman." which doesn't make sense as a) that queries instances, not images, and b) there have been no instructions at that point to launch an instance. | 07:45 |
| pgregory | Moving onto 10.1.3 we finally launch an instance, which seems to work, euca-describe-instances shows one instance in the 'running' state, with IP 10.0.0.2. | 07:46 |
| pgregory | All good so far... | 07:46 |
| pgregory | However, pinging 10.0.0.2 just sits there, and so does ssh'ing to it. | 07:47 |
| pgregory | So I've done nothing but follow the instructions to the letter, on a clean system. | 07:47 |
| pgregory | Anyone got any ideas how to track down why I can't see the running instance? | 07:47 |
| pgregory | I've scanned all the /var/log/nova/nova-*.log files, nothing but DEBUG and INFO messages in there at all. | 07:48 |
| HugoKuo__ | it's network issue | 07:49 |
| HugoKuo__ | not error or bug | 07:49 |
| pgregory | HugoKuo__: fine, but how do I fix it, or at least find out what is causing it? | 07:49 |
| HugoKuo__ | if you want somemore info of instance , | 07:49 |
| HugoKuo__ | plz #euca-get-consol-output <instance-id> | 07:50 |
| pgregory | HugoKuo__: did that, nothing showing as an error there. | 07:50 |
| HugoKuo__ | ssh daemon is up ? | 07:51 |
| HugoKuo__ | did you configure euca-authorize | 07:51 |
| pgregory | yep | 07:51 |
| pgregory | even if I hadn't, that wouldn't affect ping would it? | 07:51 |
| pgregory | euca-get-console-output: http://pastebin.com/kMejY0qn | 07:52 |
| HugoKuo__ | 2011-05-03 23:03:15,432 - DataSourceEc2.py[WARNING]: 23:03:15 [ 1\/100]: url error [[Errno 113] No route to host] | 07:53 |
| HugoKuo__ | there's the problem | 07:53 |
| HugoKuo__ | check QA on Launchpad | 07:53 |
| pgregory | ok, but what does it mean in English? | 07:53 |
| HugoKuo__ | your instance can not get metadata | 07:53 |
| pgregory | more's the point, this is clean Ubuntu, following instruction manual to the letter, using image downloaded as per instructions, why would it not work? | 07:54 |
| pgregory | where is the QA? on docs.openstack.org? | 07:54 |
| HugoKuo__ | https://answers.launchpad.net/nova | 07:55 |
| pgregory | Strange that the QA would be in a completely different place to the rest of the docs, why is that? | 07:55 |
| HugoKuo__ | actually , there's a link in openstack.org | 07:56 |
| HugoKuo__ | http://wiki.openstack.org/ | 07:57 |
| pgregory | could you help me to properly identify the real solution amongst all those candidates please. I've found that if you try one solution and it doesn't work with OpenStack you very quickly end up going down a rabbit hole with no chance of return, and after 4 complete reinstalles, I'm just about 'reinstalled out. | 07:59 |
| *** koolhead11 has quit IRC | 08:00 | |
| pgregory | last time I tried the recommeded iptables fix from the manual, I ended up going backwards, and was no longer able to even launch an instance. | 08:00 |
| HugoKuo__ | but , I'm working on different architecture now :\ | 08:03 |
| pgregory | different as in not OpenStack? | 08:04 |
| * pgregory would be very happy to hear about a more workable alternative at this point. | 08:05 | |
| HugoKuo__ | just different topology | 08:05 |
| pgregory | anyone else here able to help be get this damn thing working? | 08:06 |
| pgregory | I had hoped that following the manual to the letter would actually result in a working system, silly me. | 08:07 |
| openstackjenkins | Project dashboard-tarmac build #26,197: FAILURE in 36 sec: http://jenkins.openstack.org/job/dashboard-tarmac/26197/ | 08:07 |
| openstackjenkins | Project nova-bexar-tarmc build #22,009: FAILURE in 36 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22009/ | 08:07 |
| HugoKuo__ | how manu hosts ? | 08:08 |
| pgregory | 1 | 08:08 |
| openstackjenkins | Project nova-tarmac build #83,753: FAILURE in 36 sec: http://jenkins.openstack.org/job/nova-tarmac/83753/ | 08:08 |
| pgregory | single machine install, it does everything. | 08:08 |
| pgregory | this is just a very preliminary test to see if OpenStack will do what we want. | 08:08 |
| *** heden has joined #openstack | 08:09 | |
| HugoKuo__ | how many NICs ? | 08:09 |
| *** rcc has joined #openstack | 08:09 | |
| pgregory | I'm not trying to do anything complex or clever at this point, and it's a little worrying that even the most basic install/setup is failing so badly. | 08:09 |
| pgregory | 1 | 08:09 |
| pgregory | I'm trying to do it all within one machine, I'm not concerned at this stage about anything outside this single machine. | 08:10 |
| pgregory | I just want to run the cloud manager, and launch an instance and ssh into it. | 08:10 |
| pgregory | all within the same machine. | 08:10 |
| pgregory | nothing too complex you'd have thought. | 08:10 |
| HugoKuo__ | fine | 08:10 |
| HugoKuo__ | how many nics do you have in your host | 08:11 |
| *** allsystemsarego has joined #openstack | 08:11 | |
| *** allsystemsarego has joined #openstack | 08:11 | |
| pgregory | 1 | 08:11 |
| *** irahgel has joined #openstack | 08:11 | |
| openstackjenkins | Project dashboard-tarmac build #26,198: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/dashboard-tarmac/26198/ | 08:12 |
| openstackjenkins | Project nova-bexar-tarmc build #22,010: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22010/ | 08:12 |
| HugoKuo__ | what's your host's network | 08:13 |
| openstackjenkins | Project nova-tarmac build #83,754: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-tarmac/83754/ | 08:13 |
| pgregory | sorry, not sure what you mean, can you elaborate? | 08:13 |
| HugoKuo__ | what's your ip on eth0 | 08:14 |
| pgregory | 192.168.1.64 | 08:14 |
| HugoKuo__ | #ip link set dev br100 promisc on | 08:15 |
| HugoKuo__ | wait | 08:15 |
| HugoKuo__ | what's up network mode | 08:15 |
| pgregory | actually, that's not true, just looking at ifconfig output, it seems since the install, eth0 only has an ipv6 address, br100 is showing as 192.168.1.64 | 08:15 |
| HugoKuo__ | sure ... | 08:16 |
| pgregory | this is my ifconfig output...http://pastebin.com/L0Rg1CTj | 08:16 |
| pgregory | before running the install script all I had was eth0 and lo, and I can still ssh into the host from outside on 192.168.1.64 | 08:17 |
| HugoKuo__ | iptables -t nat -A nova-network-POSTROUTING -s 10.0.0.0/x -d 192.168.1.0/24 -j ACCEPT do ya have this rule in your iptable? | 08:17 |
| openstackjenkins | Project dashboard-tarmac build #26,199: STILL FAILING in 37 sec: http://jenkins.openstack.org/job/dashboard-tarmac/26199/ | 08:17 |
| openstackjenkins | Project nova-bexar-tarmc build #22,011: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22011/ | 08:17 |
| pgregory | sorry to be a pain, but what's the easiest way to tell (bit of a noobie I'm afraid, if you hadn't already guessed). | 08:18 |
| openstackjenkins | Project nova-tarmac build #83,755: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-tarmac/83755/ | 08:18 |
| pgregory | sudo iptables --list | grep POSTROUTING doesn't show anything. | 08:19 |
| HugoKuo__ | I don't know XD ........ it's a little complex for beginner @@ | 08:19 |
| HugoKuo__ | iptables-save | 08:19 |
| pgregory | http://pastebin.com/AsbGHcgP | 08:20 |
| HugoKuo__ | http://www.dubsquared.com/?p=43 take a look | 08:21 |
| HugoKuo__ | #iptables -t nat -A nova-network-POSTROUTING -s 10.0.0.0/8 -d 192.168.1.0/24 -j ACCEPT | 08:22 |
| openstackjenkins | Project dashboard-tarmac build #26,200: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/dashboard-tarmac/26200/ | 08:22 |
| HugoKuo__ | try this | 08:22 |
| openstackjenkins | Project nova-bexar-tarmc build #22,012: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22012/ | 08:22 |
| pgregory | ok, when I've dont that should I restart networking or any nova services? | 08:23 |
| pgregory | or just try pinging the instance? | 08:23 |
| openstackjenkins | Project nova-tarmac build #83,756: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-tarmac/83756/ | 08:23 |
| HugoKuo__ | try bootup new instance | 08:24 |
| HugoKuo__ | then check the instance console | 08:24 |
| pgregory | new instance running, still no response to ping. | 08:25 |
| pgregory | http://pastebin.com/7wpNVtr0 | 08:26 |
| openstackjenkins | Project dashboard-tarmac build #26,201: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/dashboard-tarmac/26201/ | 08:27 |
| openstackjenkins | Project nova-bexar-tarmc build #22,013: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22013/ | 08:27 |
| pgregory | then after waiting a little longer and getting the console output again... | 08:27 |
| pgregory | http://pastebin.com/8TGYKWT5 | 08:28 |
| openstackjenkins | Project nova-tarmac build #83,757: STILL FAILING in 36 sec: http://jenkins.openstack.org/job/nova-tarmac/83757/ | 08:28 |
| HugoKuo__ | still failed to get metadata | 08:28 |
| HugoKuo__ | there's a clue | 08:28 |
| HugoKuo__ | instance must be communicated with your host network | 08:29 |
| pgregory | sorry, lost me again I'm afraid. | 08:30 |
| * pgregory thinks he has a lot of reading to do if this ever gets past testing stage. | 08:30 | |
| *** JayLeeeee has quit IRC | 08:31 | |
| openstackjenkins | Project dashboard-tarmac build #26,202: STILL FAILING in 38 sec: http://jenkins.openstack.org/job/dashboard-tarmac/26202/ | 08:32 |
| openstackjenkins | Yippie, build fixed! | 08:33 |
| openstackjenkins | Project nova-bexar-tarmc build #22,014: FIXED in 1 min 12 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22014/ | 08:33 |
| openstackjenkins | Project nova-tarmac build #83,758: STILL FAILING in 37 sec: http://jenkins.openstack.org/job/nova-tarmac/83758/ | 08:33 |
| *** kashyap has quit IRC | 08:35 | |
| *** naehring has joined #openstack | 08:35 | |
| openstackjenkins | Yippie, build fixed! | 08:37 |
| openstackjenkins | Project dashboard-tarmac build #26,203: FIXED in 22 sec: http://jenkins.openstack.org/job/dashboard-tarmac/26203/ | 08:37 |
| openstackjenkins | Yippie, build fixed! | 08:37 |
| openstackjenkins | Project nova-tarmac build #83,759: FIXED in 6.2 sec: http://jenkins.openstack.org/job/nova-tarmac/83759/ | 08:37 |
| openstackjenkins | Project nova-bexar-tarmc build #22,015: FAILURE in 36 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22015/ | 08:37 |
| pgregory | anyone got any pointers how to resolve this? | 08:38 |
| openstackjenkins | Yippie, build fixed! | 08:41 |
| openstackjenkins | Project nova-bexar-tarmc build #22,016: FIXED in 2 sec: http://jenkins.openstack.org/job/nova-bexar-tarmc/22016/ | 08:41 |
| *** santhosh has quit IRC | 08:41 | |
| *** Zangetsue_ has joined #openstack | 08:42 | |
| *** Zangetsue has quit IRC | 08:42 | |
| *** Zangetsue_ is now known as Zangetsue | 08:42 | |
| *** miclorb_ has joined #openstack | 08:45 | |
| *** gholt has quit IRC | 08:48 | |
| *** rchavik has joined #openstack | 08:50 | |
| *** kashyap has joined #openstack | 08:50 | |
| *** miclorb_ has quit IRC | 08:51 | |
| pgregory | anyone got any ideas? | 08:53 |
| *** smoser has quit IRC | 09:03 | |
| *** jeffjapan has quit IRC | 09:05 | |
| *** smoser has joined #openstack | 09:06 | |
| pgregory | I've tried switching to FlatDHCPManager, and restarting/respawning, and now when I can ping, but I get connection refused when I try to ssh into it. | 09:09 |
| pgregory | s/when I can ping/I can ping/g | 09:10 |
| *** kraay has joined #openstack | 09:24 | |
| alekibango | pgregory: euca-authorize? | 09:28 |
| pgregory | alekibango: I've done "euca-authorize -P tcp -p 22 default" and "euca-authorize -P icmp -t -1:-1 default" | 09:29 |
| alekibango | hmm ok | 09:29 |
| pgregory | I've actually hit another strange problem now, I terminated and respawned the instance, it was respawned on 10.0.0.2, and now when I ping it I get From 10.0.0.1 icmp_seq=1 Destination Host Unreachable | 09:30 |
| pgregory | not sure why it would fail to find 10.0.0.1 when I was pinging 10.0.0.2. | 09:30 |
| pgregory | I currently looking at the cloud.com site, it seems to have a much more approachable interface, perhaps that would be better fit for my needs? | 09:30 |
| *** Zangetsue has quit IRC | 09:31 | |
| pgregory | I'd love to use OpenStack, it seems to tick all the boxes, open, active, EC2 compatible (to a degree), but I just cannot get it working, and <deity> knows I've tried. | 09:31 |
| alekibango | pgregory: respawn is not possible | 09:32 |
| alekibango | its either creating new one or rebooting it | 09:32 |
| *** dmi_ has joined #openstack | 09:33 | |
| alekibango | pgregory: i am having hard time with nova too | 09:33 |
| pgregory | alekibango: that's what I did, "euca-terminate-instance i-00000005" followed by "euca-run-instances ami-??? -t m1.tiny -k mykey" | 09:33 |
| alekibango | thats destroying it | 09:33 |
| alekibango | and creating another one | 09:33 |
| pgregory | yeah, that's what I wanted to do, but now I can't even ping the new one. | 09:33 |
| alekibango | its not reboot | 09:33 |
| alekibango | pgregory: rigorously check logs | 09:34 |
| alekibango | that often helped me | 09:34 |
| alekibango | maybe even reboot the server :) | 09:34 |
| pgregory | I'm sure that might solve it, but it's not really practical, rebooting the host everytime I kill an instance and start a new one isn't really going to be that useful :) | 09:35 |
| alekibango | if youare on flat now, do you have your own bridge? | 09:35 |
| alekibango | configured in /etc/network/interfaces? | 09:35 |
| pgregory | I'm now using FlatDHCPManager | 09:35 |
| alekibango | did you remove bridges created before by nova? | 09:35 |
| alekibango | did you reboot? | 09:35 |
| alekibango | maybe try :) | 09:35 |
| pgregory | I'll try | 09:35 |
| alekibango | pgregory: stop all nova services.... | 09:36 |
| alekibango | delete logs | 09:36 |
| alekibango | start services | 09:36 |
| alekibango | check logs. line by line | 09:36 |
| pgregory | I'm rebooting the host altogether. | 09:36 |
| alekibango | and google parts of the log -- you will often find some answers on nova launchpad page | 09:37 |
| alekibango | btw -- i am also for move to github | 09:37 |
| pgregory | ? | 09:37 |
| alekibango | pgregory: there was some discussion about possibilty to move from launchpad to other service -- and i like git much more | 09:38 |
| alekibango | bzr is slow and awkward somehow | 09:38 |
| pgregory | me too, I use git a lot, it's great, and github has some very cool tools. | 09:39 |
| alekibango | we can have answers in launchpad still... | 09:39 |
| alekibango | they are useful | 09:39 |
| *** miclorb has joined #openstack | 09:40 | |
| pgregory | rebooted, spawned a new instance, (10.0.0.2) pinging it responded with 10.0.0.1 Host Unreachable for a time, then I did a euca-get-console-output, and then pinged it again, and I got through. Can't ssh into it though. | 09:43 |
| pgregory | connection refused still. | 09:43 |
| alekibango | pgregory: and btw while nova is hard to use, most other solutions are in similar state too and they also add some level of missing features and adaptability... | 09:43 |
| pgregory | alekibango: possibly, but my major concern is that a) I'm not doing anything complex at the moment, just running a single machine, clean Ubuntu 10.10 install, and trying to launch a single instance on the machine, nothing outside, and it's failing at that simplest of levels, and b) I'm following the published documentation to the letter, and it's still not working. | 09:45 |
| alekibango | bugs everywhere... last time i tried archipel... very nice javascript gui, talking to servers using XMPP (which is also used for ralating users to servers)... but missing lots of features.... the gui is buggy... (tells me i cannot add drive as the machine is running while it is not, etc) and worst of all, the browser uses 2+GB ram and is often laggish when using it... | 09:46 |
| alekibango | still the project might be fun for someone with lower needs | 09:46 |
| pgregory | I'm looking at cloud.com at the moment. | 09:46 |
| naehring | sandywalsh, around? | 09:46 |
| alekibango | i just need nova to be production ready :) | 09:47 |
| alekibango | it has most functions i need | 09:47 |
| alekibango | i can adapt it to fill the need | 09:47 |
| alekibango | i just need it to work | 09:47 |
| pgregory | I just need it to work at the most basic level :) | 09:47 |
| alekibango | by which i mean, i cant delete images from glance right now etc | 09:47 |
| pgregory | by which I mean launch an instance locally on the host and connect to it, not much to ask really. | 09:48 |
| alekibango | i understand your pain pgregory :) its becoming much better after cactus | 09:48 |
| alekibango | but its still not yet there | 09:48 |
| *** rchavik has quit IRC | 09:48 | |
| alekibango | pgregory: are you using glance? | 09:48 |
| pgregory | "I have this great car for you, it has no engine at the moment, and you can't actually open the doors and get it, but it looks great!" | 09:49 |
| alekibango | :) i know how that feels | 09:49 |
| pgregory | no, just following the basic install docs on Ubuntu, using the install script from github. | 09:49 |
| alekibango | pgregory: without glance, it worked for me... now... its pain again | 09:50 |
| *** miclorb has quit IRC | 09:50 | |
| *** miclorb__ has joined #openstack | 09:50 | |
| pgregory | I just ran the script on a clean Ubuntu to get nova installed and configured, then I downloaded an image referenced also in the docs, and published it, launched it and..... | 09:50 |
| RichiH | what ram to ssd to spindle ratios make sense for swift? | 09:51 |
| pgregory | nothing, nada, zilch, neitt. | 09:51 |
| pgregory | sooooo frustrating. | 09:52 |
| alekibango | we need to make deployment easier... | 09:52 |
| alekibango | for now 90 % of people fail to run nova on one machine | 09:52 |
| pgregory | need to make deployment *possible* | 09:52 |
| alekibango | and from the rest, most fail to run it on more machines | 09:52 |
| *** onlany has quit IRC | 09:52 | |
| alekibango | pgregory: but it is possible | 09:52 |
| alekibango | in some setups | 09:53 |
| pgregory | I'm sure it is, but I've not seen it yet, and it's not possible following the guide AFAICT | 09:53 |
| alekibango | just people getting creative get in all sorts of horrible problems | 09:53 |
| pgregory | if the basic steps in the guide to get the simplest setup working fail so miserably, what chance is there of doing anything *useful* with it? | 09:54 |
| alekibango | pgregory: i believe automated functional tests which will come soon will help a lot in this | 09:54 |
| pgregory | If I were getting creative and mucking about with stuff, I'd fully understand it failing, but I'm not, I'm working from a clean install and following the setup to the letter, that at least should always work. | 09:54 |
| alekibango | please dont get angry, nova is still very young and big project | 09:55 |
| pgregory | I run an OS project, and we alway always always try to make sure that at any one time, the stable trunk works and you can follow the basic introductory documentation to get a simple test working. | 09:55 |
| alekibango | i still have faith in it | 09:55 |
| pgregory | and we only have a handful of developers doing it. | 09:56 |
| alekibango | pgregory: with nova, problem is that you can have many different configs | 09:56 |
| alekibango | many strange things which will change the result | 09:56 |
| alekibango | just little ones can do | 09:56 |
| pgregory | I do understand that, and I'm sure there are some really obscure cases, but I say again, there should be at least one straightforward install that just works. | 09:56 |
| soren | pgregory: We do have an automated test setup. It's just not super comprehensive as of yet. | 09:57 |
| pgregory | I should be able to go to the site, follow a "Getting Started" guide that says, if you do this and this and this and this, you will get a working system. | 09:57 |
| soren | pgregory: There most certainly is a straightforward install that just works. | 09:57 |
| soren | pgregory: The Ubuntu packages should be exactly that. | 09:57 |
| pgregory | soren: could you please point me at it then? | 09:57 |
| pgregory | soren: I've done that, I've installed Ubuntu 10.10 Server 4 times on this machine, and followed the install guide in the manual to the letter, and still it doesn't work. | 09:58 |
| soren | pgregory: Which install guide? | 09:58 |
| soren | pgregory: We (unfortunately) have a few. | 09:58 |
| alekibango | :) | 09:58 |
| pgregory | http://docs.openstack.org/cactus/openstack-compute/admin/content/scripted-ubuntu-installation.html | 09:59 |
| alekibango | pgregory: as i said, we are moving fast. I gave up holding hats, trying to hold my hair instead... | 09:59 |
| soren | Oh, good god, don't follow that. | 09:59 |
| alekibango | annegentle: ^^ please note | 09:59 |
| pgregory | you are kidding me! that is the exact one that someone else insisted I follow here yesterday! | 10:00 |
| *** naehring is now known as naehring|away | 10:00 | |
| alekibango | i think annegentle needs more help from devs... | 10:00 |
| alekibango | we need to document the deployment more... | 10:00 |
| alekibango | sane architectures - examples with whole configs (etc.tgz) | 10:01 |
| soren | pgregory: Hang on, let me find something for you. | 10:01 |
| alekibango | for more machines | 10:01 |
| pgregory | so, I'm willing to give it one more try, I'll reinstall Ubuntu 10.10 from clean on this machine, if someone can point me at a Ubuntu 10.10 guide that will just *work*. | 10:01 |
| pgregory | please | 10:02 |
| soren | pgregory: This one should be in order with just one change: | 10:02 |
| soren | http://fnords.wordpress.com/2010/12/02/bleeding-edge-openstack-nova-on-maverick/ | 10:02 |
| soren | pgregory: Just skip the bits about adding the PPA. | 10:02 |
| soren | pgregory: Oh, wait, no, not for maverick. | 10:02 |
| soren | For maverick, add the ppa:nova-core/release ppa. | 10:02 |
| *** reldan has quit IRC | 10:02 | |
| pgregory | right, I'm just kicking off a reinstall of Ubuntu, fingers crossed, I'll follow that to the letter and see. | 10:03 |
| * pgregory is not at all convinced, very jaded by now. | 10:04 | |
| soren | If you follow it to the letter, you'll get get nova-core/trunk ppa which is not what you want right now, probably. | 10:04 |
| soren | pgregory: Who told you to follow that guide, by the way? | 10:05 |
| alekibango | soren: when i want to use glance and i want to use uec-publish-tarball --- > do i need nova-objectstore too ? why? | 10:06 |
| soren | uec-publish-tarball works by pushing that images to our S3 implementation. That's nova-objectstore. | 10:07 |
| pgregory | hold on... | 10:07 |
| pgregory | Ryan_Lane told me to read that guide. | 10:08 |
| soren | Fascinating. | 10:08 |
| *** dirkx_ has joined #openstack | 10:08 | |
| soren | I'll have to talk to him later. I can't remember anyone reporting any sort of success with those scripts for a looong time. | 10:08 |
| * pgregory has his fingers, toes, legs, arms, and everything else he can find crossed... | 10:09 | |
| alekibango | soren: but nova-objectstore is not scalable right? | 10:09 |
| alekibango | so why is your howto telling users to use tools which need taht one? | 10:10 |
| alekibango | i am confused too :) | 10:10 |
| soren | alekibango: Well, you can run a bajillion of them on different hosts all connected to the same nfs store. | 10:10 |
| soren | But Swift is the way to go for production. | 10:10 |
| soren | Swift+glance. | 10:10 |
| soren | I'm just trying to get him started. | 10:11 |
| pgregory | just to be clear while this install is happening, I replace "sudo add-apt-repository ppa:nova-core/trunk" with "sudo add-apt-repository ppa:nova-core/release", right? | 10:11 |
| alekibango | well i have swift + glance | 10:11 |
| soren | With something that just works. | 10:11 |
| HugoKuo__ | I'm testing an unregular topology , but instance alway get "404 not found" metadata response from API server ..... | 10:11 |
| soren | ...and then we can work out from that. | 10:11 |
| alekibango | but i would like to have way for users to upload images | 10:11 |
| alekibango | how ??? | 10:11 |
| soren | pgregory: That's the idea, yes. | 10:11 |
| pgregory | thanks | 10:11 |
| soren | alekibango: I don't actually use glance myself yet. | 10:11 |
| alekibango | aaaaa | 10:11 |
| alekibango | can nova-objectstore somehow store images into glance(swift), while providing api for eucatools?? | 10:12 |
| alekibango | that might be best of both worlds | 10:13 |
| alekibango | shared nfs really is not really nice idea for me | 10:13 |
| soren | Swift supports the S3 interface. | 10:14 |
| soren | No, shared NFS is dreadful. | 10:15 |
| alekibango | ha, swift can be used directly? | 10:15 |
| soren | Sure. | 10:15 |
| alekibango | wow | 10:15 |
| * alekibango needs to find out how.. but will eat something first :) | 10:15 | |
| soren | Yeah, /me heads lunchwards, too. | 10:15 |
| *** miclorb__ has quit IRC | 10:23 | |
| *** rohit_ has joined #openstack | 10:25 | |
| *** Pyro_ has joined #openstack | 10:27 | |
| * soren lunches | 10:31 | |
| *** rchavik has joined #openstack | 10:36 | |
| *** adjohn has quit IRC | 10:40 | |
| *** zns has joined #openstack | 10:41 | |
| *** TibshoOT has quit IRC | 10:52 | |
| *** TibshoOT has joined #openstack | 10:53 | |
| *** soosfarm has quit IRC | 10:54 | |
| *** soosfarm_ has joined #openstack | 10:55 | |
| pgregory | oh well, so much for that. | 10:55 |
| sandywalsh | naehring|away, here now ... ping me when you're back | 10:56 |
| pgregory | soren: I did that, clean Ubuntu 10.10, followed the instructions on that page to the letter, apart from changing nova-core/trunk to nova-core/release, and I had to change nova.conf to include --libvirt-type=qemu as my h/w doesn't support VT-x, I can launch an instance, but can't ping it. | 10:57 |
| Pyro_ | hi, im trying to install nova-cactus using Pkg installation and i'm at the step(3) that builds and installs nova services | 10:58 |
| pgregory | s/--libvirt-type/--libvirt_type/g | 10:58 |
| Pyro_ | the step 'sudo pip install -r tools/pip-requires' | 10:58 |
| alekibango | btw is here anyone from czech republic? | 10:58 |
| Pyro_ | fails with url timeouts | 10:59 |
| pgregory | Pyro_: good luck with that, I've been trying solidly for 2 days now, no success at all. | 10:59 |
| Pyro_ | yes, it seems that the urls are incorrect, i've been struggling to install many of those tools manually! | 10:59 |
| alekibango | Pyro_: being you, i would first test it on ubuntu | 11:00 |
| alekibango | with packages | 11:00 |
| Pyro_ | im on a ubuntu 10.10 | 11:00 |
| alekibango | 2011-05-04 12:02] <soren> http://fnords.wordpress.com/2010/12/02/bleeding-edge-openstack-nova-on-maverick/ | 11:02 |
| pgregory | soren: if you get a mo, I'm *exactly* where I was before after trying that guide. I can launch an instance, can't ping it, (get error 10.0.0.1 unreachable, irrespective of the instance IP), then I try euca-get-console-output, and ping it again, and magically it works, but when I try to ssh into it I get connection refused. | 11:02 |
| alekibango | 2011-05-04 12:02] <soren> For maverick, add the ppa:nova-core/release ppa. | 11:02 |
| pgregory | Deja-bloody-Vu, again, again, again.... | 11:02 |
| alekibango | Pyro_: try that ... | 11:02 |
| pgregory | Pyro_: and if you get it working, I'd be interested to hear, that's what I've just tried and failed with. | 11:02 |
| Pyro_ | ok, let me try my best, because the Pylint URL that failed with pip just opened (and downloaded the tgz) manually | 11:04 |
| pgregory | I'm going to wait until soren returns from lunch and see if there's any solution to this situation, then I'm going to give cloud.com a try, see if I have any more luck with that. | 11:06 |
| soren | pgregory: Are you using the image from that blog post | 11:06 |
| soren | pgregory: ? | 11:06 |
| soren | pgregory: Can you pastebin the console output (from euca-get-console-output), please? | 11:07 |
| Pyro_ | lunch? which timezone is soren in | 11:07 |
| pgregory | soren: yes, and just a sec... | 11:08 |
| soren | Pyro_: CEST. | 11:08 |
| soren | Pyro_: UTC+2. | 11:08 |
| pgregory | soren: BTW there is a small error in that guide, the downloading of the image uses $image in the publish command, but it's never defined, small issue I know, but worth fixing. | 11:10 |
| Pyro_ | thanks :) are you soren from the openstack intro video? that was a nice talk | 11:10 |
| pgregory | http://pastebin.com/p188H6zV | 11:10 |
| soren | Pyro_: The one from FOSDEM? | 11:11 |
| Pyro_ | yes i think, it was last year, july i think, i watched this video two days ago | 11:11 |
| soren | pgregory: Ah. That's the image's fault. Plymouth is exploding because it doesn't have a console to talk to and then it gets lonely. | 11:11 |
| Pyro_ | jul/feb | 11:12 |
| soren | pgregory: I fixed that in Natty. | 11:12 |
| pgregory | soren: is there a working image I can try with then? | 11:12 |
| soren | pgregory: The natty one should work well. I fixed Plymouth to not throw its toys out of the pram if there's no console for it to play with. | 11:13 |
| pgregory | so same place but with natty in place of maverick? | 11:13 |
| soren | pgregory: Yes. | 11:13 |
| soren | You're rather unlucky. It only happens every once in a while. | 11:14 |
| soren | init: mountall main process (54) killed by PIPE signal | 11:14 |
| soren | ...is the indication that it did this time. | 11:14 |
| alekibango | which image are u talking about? | 11:14 |
| soren | Pyro_: Err... So Feb of this year? At FOSDEM? | 11:14 |
| soren | Pyro_: Anyways, if it's soeone named soren who works on OpenStack, it's quite likely that it's me :) | 11:15 |
| alekibango | hint: he is not the golf player :) | 11:15 |
| pgregory | soren: mind if I ask where you're based? | 11:15 |
| Pyro_ | cool, soren hansen | 11:16 |
| alekibango | pgregory: i am the same timezone as you two... (cz).. americans will come in 2 hours | 11:16 |
| soren | pgregory: Denmark. | 11:17 |
| pgregory | I'm UK | 11:17 |
| * soren waves westwards | 11:17 | |
| Pyro_ | i'm a noob at openstack, so may ask lot of nooby questions in the coming days, please bear | 11:17 |
| alekibango | :) i thought you are italy :) | 11:17 |
| soren | What, me? Really? | 11:17 |
| alekibango | Pyro_: we like questions! | 11:17 |
| soren | I'd make a dreadful Italian. | 11:18 |
| alekibango | soren: no, pgregory :) | 11:18 |
| Pyro_ | thnx,im Rohit from India | 11:18 |
| *** rostik has joined #openstack | 11:18 | |
| *** rchavik has quit IRC | 11:18 | |
| Pyro_ | and i feel soo bad to have missed the design summit by a whisker! | 11:18 |
| pgregory | alekibango: :) what made you think I was from Italy? My bad English ;) | 11:19 |
| alekibango | quick glance over line in whois: pgregory is online via calvino.freenode.net (Milan, IT). | 11:19 |
| alekibango | heh | 11:19 |
| soren | Pyro_: Oh, but thanks! I'm glad you liked the talk. | 11:19 |
| alekibango | soren: btw, thanks a lot for pointing something obvious for me | 11:19 |
| soren | alekibango: hm? | 11:20 |
| alekibango | i was somehow interpreting the text line """For production use, you should use OpenStack Storage, aka. Swift.""" as 'TODO item' | 11:20 |
| pgregory | I hop all over the place, I have a 'screen' running on a server in Oregon State University which I use to maintain a constant connection to IRC. | 11:20 |
| pgregory | need it for my OS project. | 11:20 |
| rostik | is it possible in swift to specify that data belonging to accountA should be stored in specific zone(s)? | 11:20 |
| * soren is connected from... Dallas, I think. Or Chicago. | 11:20 | |
| alekibango | pgregory: what project are u working on? | 11:20 |
| *** ctennis has quit IRC | 11:21 | |
| pgregory | I run the Aqsis project, an OS RenderMan compliant renderer, but my interest in OpenStack is something else, sort of related. | 11:21 |
| alekibango | ah... ic :) | 11:21 |
| alekibango | aqsis is nice | 11:21 |
| pgregory | thanks | 11:21 |
| alekibango | btw do you know makehuman? | 11:22 |
| alekibango | i hope nova will launch faster than makehuman hehe | 11:22 |
| pgregory | yes, very well, we've worked with the guys for some time on that project. | 11:22 |
| alekibango | i am waiting for makehuman to be useable for long time :) i would love to make some dilbert like stories using it | 11:23 |
| pgregory | :) | 11:24 |
| pgregory | hmm, qemu is going to be sllooooow right? | 11:28 |
| pgregory | I'm monitoring the console output, and it's chugging along now, Plymouth gave some errors, but it seems to be making progress beyond the last one. | 11:28 |
| pgregory | yeee-bloody-har, I'm in!!! | 11:29 |
| pgregory | seems that when "euca-describe-instances" says "running", it's not 'ready' yet, takes quite a while after that before enough of the instance is up to be able to connect to it, understandable I guess with qemu | 11:30 |
| pgregory | soren: thanks loads, what a relief to finally get something working. | 11:30 |
| *** rchavik has joined #openstack | 11:30 | |
| *** ctennis has joined #openstack | 11:38 | |
| *** ctennis has joined #openstack | 11:38 | |
| pgregory | next question, is it possible to run a web dashboard on this configuration? | 11:39 |
| *** markvoelker has joined #openstack | 11:42 | |
| soren | pgregory: Yeah, what I changed in plymouth is just to make it die, not take out mountall and whatnot with it. | 11:44 |
| soren | pgregory: Yeah, euca-describe-instances will say running quite early. | 11:44 |
| soren | pgregory: In physical-hardware terms it's once it's been switched on and has been determined to keep drawing power. It may still take a while before it's actually booted. | 11:45 |
| soren | pgregory: And yes, QEMu will be sloooow. | 11:45 |
| soren | pgregory: Even though "Q" stands for quick. | 11:45 |
| soren | Actually, QEMu is a pretty quick emulator. It's just that emulation is slow. | 11:46 |
| soren | Kind of like how the world's fastest snail still isn't very fast at all. | 11:47 |
| *** hadrian has joined #openstack | 11:48 | |
| *** beebrox has joined #openstack | 11:49 | |
| *** hadrian_ has joined #openstack | 11:53 | |
| *** hadrian__ has joined #openstack | 11:55 | |
| *** hadrian has quit IRC | 11:56 | |
| *** hadrian__ is now known as hadrian | 11:56 | |
| *** dmi_ has quit IRC | 11:57 | |
| alekibango | pgregory: can you please give me link to the image you were talking about? | 11:57 |
| pgregory | alekibango: the one that didn't work? | 11:58 |
| *** hadrian_ has quit IRC | 11:58 | |
| *** naehring|away is now known as naehring | 11:58 | |
| naehring | sandywalsh, here I am. ping ;) | 11:58 |
| alekibango | pgregory: i would rather see one which works | 11:59 |
| alekibango | :) | 11:59 |
| pgregory | http://uec-images.ubuntu.com/natty/current/natty-server-uec-amd64.tar.gz | 11:59 |
| alekibango | ty | 11:59 |
| pgregory | soren: do you know if it's possible to run Dashboard on something like this? | 12:00 |
| pgregory | although I suppose I ought to get public IP's and so on working first. | 12:00 |
| *** gondoi has joined #openstack | 12:00 | |
| soren | pgregory: "something like this"? | 12:01 |
| pgregory | soren: this setup, the Ubunutu 10 with maverick release ppa | 12:02 |
| soren | pgregory: Sure it's possible. IT's what it was made for, pretty much. | 12:02 |
| pgregory | any pointers to documentation? | 12:03 |
| pgregory | I have a meeting on Monday, and it'd be fantastic to show the rest of the team a running cloud host that they can connect to and view/run instances in a web browser. | 12:03 |
| pgregory | we're probably 6-8 months away from *needing* the cloud resources, but it'd be nice to say "this is our solution". | 12:04 |
| soren | pgregory: I think the repository includes documentation, but I'm not entire sure. HAve you deployed Django apps before? | 12:05 |
| soren | entirely. Man, I type poorly today. | 12:05 |
| pgregory | played with Django, yeah. | 12:05 |
| soren | Ok, then this shouldn't be hard at all. | 12:06 |
| soren | It's not packaged yet, unfortunately. | 12:06 |
| pgregory | I'm a software architect by trade, so I'm more than happy getting my hands dirty with that sort of thing (prefer Rails thought ;) ) | 12:06 |
| alekibango | i cant find docs on how to use swift s3 with nova... | 12:06 |
| soren | I've not yet come up with a good way to package django apps. :( | 12:06 |
| alekibango | soren: can you give some hint please? | 12:06 |
| *** hadrian has quit IRC | 12:06 | |
| soren | alekibango: Doesn't it "just work2? | 12:07 |
| alekibango | how to start swift providing s3 ? | 12:07 |
| alekibango | i cant find s3 in nova docs :) | 12:07 |
| *** hagarth has quit IRC | 12:07 | |
| soren | alekibango: Again: Doesn't it "just work"? | 12:09 |
| soren | alekibango: If you speak S3 to it, does it not work? | 12:09 |
| alekibango | soren: how? | 12:09 |
| alekibango | i am not speaking s3 myself :) i never used amazon | 12:10 |
| soren | You've completely lost me. Can we start over? What are you trying to do? | 12:10 |
| zykes- | when will there be any design summit to europe ? | 12:10 |
| alekibango | soren: i would love to store images in swift | 12:11 |
| alekibango | in a way that endusers of public cloud will be able to use | 12:11 |
| zykes- | it's kind of hopeless for folks here to try to launch it when all the focus is in the states | 12:11 |
| fabiand__ | zykes-, +1 from me for this. | 12:12 |
| soren | zykes-: While I do agree we should do on in Europe, I don't see how the absence of one in Europe is holding anyone back from launching OpenStack? | 12:12 |
| soren | It's not a training event. | 12:12 |
| alekibango | so in nova flag file i should use : --image_service nova.image.s3.S3ImageService , and --s3_host $SWIFT_PROXY_IP -- right? what more should i do? | 12:13 |
| alekibango | shouldnt nova need some account? | 12:13 |
| alekibango | password, etc? | 12:13 |
| zykes- | soren: just that the company i work for thinks that there's not focus towards .eu for it | 12:13 |
| alekibango | soren: just one documented example would make this clear | 12:14 |
| notmyname | alekibango: I can't help with using nova, but to enable the s3 compatibility in swift, enable the swift3 middleware on the swift proxy server (see the proxy server sample config). also, can nova speak to swift using the swift api? perhaps this is where glance comes in? | 12:14 |
| alekibango | notmyname: glance fails when i use uec-publish-tarballl | 12:14 |
| alekibango | and euca-... commands | 12:15 |
| alekibango | it tries to connect to s3 | 12:15 |
| soren | aliguori: uec-publish-tarball has no idea about glance. | 12:15 |
| soren | errr... | 12:15 |
| soren | aliguori: Not you :) | 12:15 |
| soren | alekibango: uec-publish-tarball has no idea about glance. | 12:15 |
| alekibango | soren: yes i know this now :) | 12:15 |
| alekibango | but how to store images for uec-publish-tarball in glance? | 12:15 |
| alekibango | er.. in swift | 12:15 |
| soren | Why do you want to? | 12:16 |
| alekibango | soren: i don not like storing image in nfs | 12:16 |
| alekibango | i want users to be able to upload images somehow | 12:16 |
| soren | alekibango: This whole mess is a hack to support people who are stuck with tools that are tied to AWS. | 12:17 |
| zykes- | there's no events so far after 1 year then for the project over here. | 12:17 |
| soren | alekibango: If you can use the OpenStack tools instead, you should. | 12:17 |
| alekibango | soren: i see... if i have other option i would use | 12:17 |
| alekibango | how? | 12:17 |
| alekibango | is that documented/... and working? | 12:17 |
| soren | alekibango: Use glance. | 12:17 |
| alekibango | i failed to do so horribly | 12:17 |
| alekibango | you mean let endusers of public cloud to use public glance api? | 12:18 |
| alekibango | is that good idea considered that glance is not secure much and has no auth? | 12:18 |
| soren | alekibango: Talk to jaypipes. | 12:18 |
| alekibango | soren: nova with -- (s3 + swift) would be nice -- but i fail to find example of that | 12:19 |
| sandywalsh | http://aws.typepad.com/aws/2011/05/iam-identity-access-management.html | 12:21 |
| alekibango | i wanted to go into production after bexar... but even testing it is so crazy still.... | 12:21 |
| alekibango | sandywalsh: ? | 12:22 |
| *** burris has quit IRC | 12:22 | |
| *** zns has quit IRC | 12:23 | |
| sandywalsh | alekibango, just interesting | 12:23 |
| alekibango | it is :) | 12:23 |
| pgregory | Wow! | 12:26 |
| *** burris has joined #openstack | 12:26 | |
| pgregory | it just *worked*, does exactly what it says on the tin! | 12:26 |
| pgregory | what a pleasant surprise that is. | 12:26 |
| alekibango | ?? what worked pgregory? | 12:26 |
| pgregory | OpenStackDashboard | 12:27 |
| alekibango | :) | 12:28 |
| pgregory | I just installed it and launched it exactly as described in the docs, and now I can access the host machine via a web browser on a separate machine and look at running instances, projects, images etc. | 12:28 |
| pgregory | sooooo cool. | 12:28 |
| alekibango | :) | 12:28 |
| alekibango | congrats | 12:28 |
| pgregory | not to me, to whomever is responsible for setting up the project and docs, couldn't have been easier. | 12:29 |
| alekibango | i am glad you have some good result finally | 12:29 |
| *** dirkx_ has quit IRC | 12:30 | |
| *** posulliv has quit IRC | 12:33 | |
| pgregory | I can even launch instances in it. | 12:33 |
| * pgregory is a very happy bunny now. | 12:33 | |
| alekibango | the car started moving hehe | 12:33 |
| pgregory | :D | 12:33 |
| *** kashyap has quit IRC | 12:34 | |
| *** kashyap has joined #openstack | 12:34 | |
| *** naehring has quit IRC | 12:38 | |
| *** naehring has joined #openstack | 12:38 | |
| *** ivan has quit IRC | 12:44 | |
| *** ivan has joined #openstack | 12:45 | |
| soren | I hate holidays. It takes *days* to catch up on e-mail once you're back. | 12:46 |
| rostik | swift.common.Ring.get_nodes('path') -> will the number of nodes be equal to number of replicas where the object is to be saved? does node == replica here? | 12:51 |
| notmyname | rostik: yes | 12:53 |
| notmyname | you will get the partition and recplica_count nodes | 12:54 |
| Pyro_ | why do command line installations from pypi time out so often? | 12:54 |
| notmyname | use get_more_nodes(partition) to get handoff nodes, up to a total of zone_count nodes | 12:54 |
| *** zns has joined #openstack | 12:56 | |
| rostik | +notmyname: thanks. maybe you can also clarify the following. Is it possible to impose a restriction that data belonging to specific account has to be stored in specific zone(s)? | 12:57 |
| *** dirkx_ has joined #openstack | 12:57 | |
| notmyname | no. we talked about modifying the ring to allow for those types of constraints. currently, you are only guaranteed each copy in a unique zone | 12:57 |
| notmyname | s/we talked about/last week at the conference we talked about | 12:58 |
| *** hadrian has joined #openstack | 13:00 | |
| RichiH | are there any recommended command line tools to get data into & out of swift? | 13:03 |
| soren | Swift comes with a client tool. | 13:05 |
| soren | It's called st. | 13:05 |
| *** beebrox has quit IRC | 13:08 | |
| notmyname | st, as soren said, is nice. but swift has a REST-ful/HTTP API, so anything that speaks HTTP can be used (like curl and wget). st simply takes care of some of the extra typing | 13:11 |
| *** posulliv has joined #openstack | 13:16 | |
| zul | soren: easy to solve...never go on holidays | 13:16 |
| alekibango | or stop using email :) | 13:18 |
| *** naehring has quit IRC | 13:18 | |
| *** dirkx_ has quit IRC | 13:20 | |
| *** msivanes has quit IRC | 13:27 | |
| *** msivanes has joined #openstack | 13:28 | |
| pgregory | hmm, anyone know how I make it possible for a user added via the Dashboard admin to be able to launch instances? | 13:28 |
| alekibango | soren: https://answers.launchpad.net/glance/+question/146519 -- jaypipes thinks we do not need s3 at all | 13:28 |
| pgregory | I just get permission denied, I've created a new key. | 13:28 |
| *** lborda has joined #openstack | 13:29 | |
| *** santhosh has joined #openstack | 13:32 | |
| Pyro_ | I just got a 'Command failed, please check log for more info' | 13:33 |
| Pyro_ | after running nova-manage | 13:33 |
| Pyro_ | question is ...where is the log file? | 13:33 |
| *** naehring has joined #openstack | 13:33 | |
| pgregory | Pyro_: on Ubuntu, /var/log/nova/nova*.log | 13:33 |
| Pyro_ | ok | 13:34 |
| Pyro_ | no nova/ for me in /var/log | 13:34 |
| *** amccabe has joined #openstack | 13:35 | |
| Pyro_ | my installation is in /opt/nova-2011.2 | 13:35 |
| alekibango | Pyro_: try makng the directory, owned by nova | 13:35 |
| Pyro_ | ok | 13:35 |
| alekibango | --logdir=/var/log/nova | 13:35 |
| alekibango | and configure it in nova.conf file | 13:35 |
| Pyro_ | don't know what's wrong but my installation did not create a 'nova' user | 13:36 |
| Pyro_ | none in /etc/passwd | 13:36 |
| Pyro_ | nova-manage user create and project create worked fine | 13:37 |
| alekibango | hmm installing from source its imho on you | 13:37 |
| Pyro_ | and I could get the access keys | 13:37 |
| alekibango | u need to make sure it runs as user you want | 13:37 |
| alekibango | nova works good when runned as root | 13:37 |
| alekibango | :) | 13:37 |
| alekibango | but its not that secure | 13:37 |
| alekibango | you can run nova as user johndoe if you want... just make all rights ok | 13:38 |
| alekibango | (which is sometimes 'fun') | 13:38 |
| Pyro_ | i did not run anything as root, but with sudo | 13:39 |
| Pyro_ | and my user has privs in /etc/sudoers | 13:39 |
| alekibango | ps faxw |grep nova ---- will reveal who is running it | 13:39 |
| alekibango | ps fauxw |grep nova ---- will reveal who is running it | 13:40 |
| Pyro_ | yikes, no nova running! | 13:40 |
| Pyro_ | is there a service start thing here? | 13:40 |
| alekibango | i dont know if there is in src | 13:41 |
| Pyro_ | http://wiki.openstack.org/NovaInstall/Cactus this is what i followed till the dot | 13:41 |
| *** dmi_ has joined #openstack | 13:41 | |
| alekibango | but you definitelly need to run few services | 13:41 |
| *** yair has joined #openstack | 13:41 | |
| soren | alekibango: That's what I keep saying! | 13:41 |
| *** dmi_ has quit IRC | 13:42 | |
| alekibango | soren: if you are talking about glance/s3 now, ok... but glance still is buggy and obscure a lot | 13:42 |
| alekibango | and not auth? | 13:42 |
| soren | alekibango: I thought glance worked rather well. | 13:42 |
| alekibango | watch bug list | 13:42 |
| alekibango | you cant delete image | 13:42 |
| alekibango | ... no auth | 13:42 |
| Pyro_ | i remember installing using nova.sh and at the end i had 6-7 services running on the single host | 13:43 |
| *** mahadev has joined #openstack | 13:43 | |
| alekibango | which means anyone can add mirriade of images | 13:43 |
| soren | alekibango: I also keep saying: Talk to jaypipes. :) | 13:43 |
| *** zenmatt has joined #openstack | 13:43 | |
| alekibango | i can imagine content sharing network -- which will misuse it | 13:43 |
| Pyro_ | and each on a different screen (?) | 13:43 |
| soren | Anyways, I need to run guys. My daughter is in the hospital. | 13:43 |
| alekibango | soren: i will, sorry for bugging you so much :) | 13:43 |
| alekibango | and thanks for answers, you helped me a lot | 13:43 |
| Pyro_ | take care soren, hope she is better | 13:44 |
| alekibango | +1 | 13:44 |
| alekibango | get her some vitamines, food in hospitals is generally not good | 13:45 |
| *** dmi_ has joined #openstack | 13:45 | |
| *** f4m8 is now known as f4m8_ | 13:47 | |
| *** agarwalla has quit IRC | 13:47 | |
| zns | alekibango: on auth, we're working with jaypipes to integrate glance with keystone soon to provide auth. Keystone is proposed identity & auth - see launchpad.net/keystone. | 13:49 |
| *** santhosh has quit IRC | 13:50 | |
| jaypipes | alekibango: once we have an auth tenant from keystone, we will have shared image groups. expect that for the D2 milestone in this release series. | 13:50 |
| Pyro_ | i dont the log dir should be manuall created | 13:50 |
| *** yair has quit IRC | 13:50 | |
| Pyro_ | the installation must create it | 13:50 |
| *** santhosh has joined #openstack | 13:51 | |
| alekibango | Pyro_: you can configure it to store logs in other place | 13:52 |
| alekibango | jaypipes: thanks for info | 13:52 |
| alekibango | zns: also | 13:52 |
| alekibango | thanks | 13:52 |
| alekibango | ... so when i could think about nova ready for production on 20-100 servers? | 13:52 |
| alekibango | i mean in HA mode | 13:52 |
| alekibango | and with sane storage and auth and stuff | 13:53 |
| pgregory | I've created a new project for a colleague to test, he is project manager, so is able to launch instances (all this done throught he dashboard, which is cool), but when he launches is stick in 'networking' state, and the logs have a NoMoreNetworks error. | 13:53 |
| pgregory | do I have to allocate a new network to the project? if so, any pointers to the docs for doing so? | 13:53 |
| alekibango | try looking on mova-manage network | 13:54 |
| *** santhosh has quit IRC | 13:55 | |
| pgregory | alekibango: I did, but I can't find anything in there that associates a network with a project. | 13:55 |
| pgregory | or doesn't it matter, will it just pick a network that isn't in use? | 13:55 |
| *** med_out is now known as medberry | 13:56 | |
| *** mahadev has quit IRC | 13:57 | |
| *** deshantm_laptop has joined #openstack | 13:59 | |
| *** troytoman-away is now known as troytoman | 13:59 | |
| pgregory | no worries, sussed it, just create a couple of networks and the projects will pick them up it seems. | 14:00 |
| pgregory | cheers | 14:00 |
| alekibango | yep.. | 14:01 |
| *** troytoman is now known as troytoman-away | 14:01 | |
| alekibango | jaypipes: i have seen another maybe related problem with glance - image URI seems to be missing port in details. not sure if its connected to the hottest bug yuo have... | 14:02 |
| alekibango | but i would sure love to have this fixed | 14:03 |
| *** mahadev has joined #openstack | 14:04 | |
| alekibango | so... when you think will nova be ready for production (managing 100 servers)? | 14:07 |
| *** MarkAtwood has joined #openstack | 14:09 | |
| Pyro_ | hope its not n(e)va | 14:09 |
| alekibango | i am not happy with nova release schedule. i would love it being more agile... and working all the time... | 14:10 |
| jaypipes | alekibango: hmm... unrelated to the hot bug. Could you file a bug for that? Thanks much! | 14:11 |
| alekibango | jaypipes: i might even try to fix it | 14:11 |
| jaypipes | alekibango: even better! :) | 14:12 |
| alekibango | jaypipes: but i first need to fix few things with my nova... may i have few questions on you? | 14:13 |
| pgregory | what's the best way to get access to instances from outside the host? | 14:14 |
| jaypipes | alekibango: I will try my best :) | 14:14 |
| alekibango | here i think i have similar problem -- https://answers.launchpad.net/swift/+question/154177 | 14:15 |
| alekibango | do you think i should not use s3 backend + swift | 14:16 |
| alekibango | and i should rather use glance | 14:16 |
| alekibango | ... i cant find howto (docs) to help me store nova images in swift s3 | 14:16 |
| alekibango | simple example would be nice | 14:16 |
| alekibango | but i am not sure if its worth the job | 14:17 |
| alekibango | i am afraid to open glance on public ip | 14:17 |
| alekibango | people will start uploading anything really | 14:17 |
| alekibango | using it as content sharing | 14:17 |
| alekibango | :) | 14:17 |
| alekibango | i cant wait few months to get it fixed - i need people to be able to upload images asap | 14:18 |
| alekibango | into public cloud | 14:18 |
| alekibango | and using nfs storage for images doesnt look good for me... i want to use swift | 14:18 |
| alekibango | how? | 14:18 |
| *** santhosh has joined #openstack | 14:19 | |
| *** mahadev has quit IRC | 14:20 | |
| *** mahadev has joined #openstack | 14:21 | |
| alekibango | jaypipes: ^^ | 14:21 |
| jaypipes | alekibango: can I ask why you want to do that? (as opposed to just using Glance with Swift backend?) | 14:22 |
| alekibango | public glance api without auth ==> swift full of anything people will feed into it | 14:23 |
| jaypipes | alekibango: why do you need to open glance IP to public? | 14:23 |
| alekibango | jaypipes: so how will endusers of the cloud upload images? | 14:23 |
| alekibango | i dont get this part... maybe i am not smart enough :) | 14:23 |
| jaypipes | alekibango: via Nova :) | 14:23 |
| alekibango | ...?? | 14:24 |
| alekibango | nova-manage image ? | 14:24 |
| alekibango | oh ****. i forgot about that one | 14:24 |
| jaypipes | alekibango: I mean, I understand your concern, I do. Until we get auth done, the way I'm suggesting interacting with Glance is through Nova and the tools you normally use for your API. If you use EC2, then you would use the euca-bundle/upload stuff, and if you use the OpenStack API, you would use python-novaclient. | 14:24 |
| *** mahadev has quit IRC | 14:25 | |
| alekibango | euca- commands failed for me without objectstore | 14:25 |
| jaypipes | alekibango: in other words, if you set up Nova to use Glance, you simply don't need to interact with Glance manually... | 14:25 |
| *** mgoldmann has joined #openstack | 14:25 | |
| jaypipes | alekibango: yes, you need objectstore in order to use the EC2 api. there's no way around that currently, since the eucatools work over the S3 api. | 14:26 |
| alekibango | so i will end up having objectstore with images stored in filesystem? | 14:26 |
| alekibango | using swift would fix that | 14:26 |
| alekibango | (as s3) | 14:26 |
| jaypipes | alekibango: nova-objectstore will store the intermediate image parts (kernel, ramdisk), and glance will store the decrypted image (AMI) | 14:26 |
| alekibango | i dont think i understand your last line well, can you please explain? | 14:27 |
| *** Pyro_ has quit IRC | 14:27 | |
| alekibango | jaypipes: i am sorry, i just fail to understand this from docs | 14:27 |
| jaypipes | alekibango: ok, no worries, let me walk you through it. | 14:27 |
| *** YHershko has joined #openstack | 14:27 | |
| alekibango | (iam not user of amazon so some things are new to me) | 14:28 |
| alekibango | so nova-objectstore might store some parts which glance will comebine and store in swift? | 14:29 |
| alekibango | is that what you mean? | 14:29 |
| *** zenmatt has quit IRC | 14:30 | |
| jaypipes | alekibango: ok, so when someone does image bundling using euca2ools (euca-bundle-image/uec-publish-tarball, etc), what happens is that euca-bundle-image does a few things: a) create the kernel and ramdisk image pieces (intermediate image files) b) encrypt these intermediate files with the user's key, c) create the manifest.xml file that describes the image and d) tar up a kernel, ramdisk, and machine image into an encrypt | 14:30 |
| jaypipes | ed tarball. | 14:30 |
| alekibango | jaypipes: may i start some etherpad which we can edit together? i will try to insert it into wiki somewhere. it would help a lot to newbies | 14:31 |
| jaypipes | alekibango: euca-bundle-image does not talk to Glance directly. It talk to the EC2 API, issuing image-related commands as well as the S3 API for storing the intermediate image files and the manifest.xml file. | 14:31 |
| jaypipes | alekibango: that's a good idea, yes. | 14:31 |
| alekibango | http://etherpad.openstack.org/glance-walkthrough | 14:31 |
| jaypipes | alekibango: awesome. | 14:32 |
| jaypipes | alekibango: OK, so open up the file nova/image/s3.py | 14:32 |
| *** imsplitbit has joined #openstack | 14:32 | |
| alekibango | k | 14:33 |
| *** jkoelker has joined #openstack | 14:33 | |
| jaypipes | alekibango: you will see in the _s3_create() method the various interactions between nova-objectstore, nova, and Glance. Basically, the S3ImageService reads the manifest.xml file from nova-objectstore, then registers metadata about the kernel and ramdisk image parts in Glance. It then decrypts that machine image and stores the decrypted machine image file in Glance, along with a set of metadata that tells it which kernel | 14:34 |
| jaypipes | and ramdisk are in use for that machine image. | 14:34 |
| *** photron has joined #openstack | 14:34 | |
| alekibango | jaypipes: write that into the etherpad directly :) | 14:34 |
| *** jamesurquhart has joined #openstack | 14:35 | |
| jaypipes | alekibango: when you do things like euca-describe-images, what is happening is euca-describe-images is talking to the EC2 API controller, which is routing the request to the S3ImageService, which then queries Glance for image information... | 14:35 |
| jaypipes | alekibango: I was writing in case anyone else was interested... ;) I'll copy to etherpad | 14:35 |
| alekibango | i am copying right now :) | 14:36 |
| jaypipes | alekibango: ah :) | 14:36 |
| *** zenmatt has joined #openstack | 14:36 | |
| jaypipes | alekibango: heh, and that's just the EC2 API... for OpenStack API, it's completely different. :( | 14:37 |
| alekibango | jaypipes: what is shim ? | 14:43 |
| alekibango | a small library which transparently intercepts an API, changes the parameters passed, handles the operation itself, or redirects the operation elsewhere. :) | 14:44 |
| *** rnirmal has joined #openstack | 14:45 | |
| jaypipes | alekibango: you got it :) | 14:45 |
| jaypipes | alekibango: hey, running to a meeting... I'll catch up to you in a short while. Please add questions to the etherpad that I can answer! | 14:46 |
| alekibango | k, thansk! | 14:46 |
| *** knorthcut has joined #openstack | 14:46 | |
| alekibango | this clears it a lot for me | 14:46 |
| alekibango | but there will be still few questions :) | 14:46 |
| imsplitbit | why hello fellow openstackers | 14:47 |
| imsplitbit | it's a great day to be alive | 14:47 |
| alekibango | morning, imsplitbit :) -- jaypipes is helping me understand glance on http://etherpad.openstack.org/glance-walkthrough | 14:48 |
| alekibango | imsplitbit: enjoy the radioactive fallout... dont worry, be happy! | 14:49 |
| imsplitbit | alekibango: nice, jaypipes is freakin awesome. He is helping me learn pymox/nosetests. | 14:49 |
| *** hub_cap has joined #openstack | 14:50 | |
| * alekibango runs away from java... | 14:50 | |
| imsplitbit | alekibango: LOL that is a good thing to do | 14:50 |
| alekibango | :) | 14:51 |
| imsplitbit | I have been a ruby developer for about 8 years so I am not yet pythonic, but I am learning. and no, I didn't do ruby on rails. I curse the day rails was created, it has destroyed the ruby community and the public image of a great language. | 14:52 |
| alekibango | imsplitbit: why it was such problem? | 14:54 |
| alekibango | i do not understand | 14:54 |
| alekibango | (i avoided ruby also, but it is not freaking me out) | 14:55 |
| *** santhosh has quit IRC | 14:55 | |
| alekibango | i was friends with perl... hehe | 14:55 |
| alekibango | now for 10 last years, python is the love story | 14:56 |
| imsplitbit | hmm I wish I could find love for python, it is more utilitarian for me. I don't dislike it. | 14:56 |
| imsplitbit | rails devs ruined the ruby community for the most part with their religious zealotry. No one language is right for everything and they didn't really see that. It made people making good honest ruby based apps have trouble trying to sell their project. | 14:58 |
| imsplitbit | People would say "well yeah but ruby doesn't scale" which is completely idiotic. Rails has problems at scale but that is due to the rails infrastucture, not because the ruby language somehow doesn't have the ability to run on several machines. | 14:59 |
| imsplitbit | anyway I shouldn't get myself started | 14:59 |
| imsplitbit | so far my python experience is with writing a virt connector for nova and it is treating me ok thus far | 15:00 |
| *** dmi_ has quit IRC | 15:01 | |
| *** daveiw has quit IRC | 15:02 | |
| *** dragondm has joined #openstack | 15:03 | |
| *** mray has joined #openstack | 15:04 | |
| *** mray has left #openstack | 15:04 | |
| *** zns has quit IRC | 15:06 | |
| *** mray has joined #openstack | 15:06 | |
| alekibango | imsplitbit: :) | 15:06 |
| alekibango | imsplitbit: pythons power is in simplicity.. it reads like english poem | 15:06 |
| *** zns has joined #openstack | 15:07 | |
| alekibango | with python, mostly there is 'one really obvious way to do it' :) | 15:08 |
| *** rchavik has quit IRC | 15:09 | |
| *** ewindisch has quit IRC | 15:10 | |
| *** jdurgin has joined #openstack | 15:13 | |
| *** zns has quit IRC | 15:14 | |
| *** nelson has joined #openstack | 15:15 | |
| *** yamahata_lt has joined #openstack | 15:18 | |
| HugoKuo__ | does anyone knows where's docs about instance metadata ? | 15:19 |
| *** cole has joined #openstack | 15:22 | |
| *** zns has joined #openstack | 15:23 | |
| *** burris has quit IRC | 15:23 | |
| *** mancdaz has joined #openstack | 15:24 | |
| *** YHershko has quit IRC | 15:25 | |
| *** rostik has quit IRC | 15:26 | |
| *** jmckind has joined #openstack | 15:27 | |
| *** hiddentoken has joined #openstack | 15:30 | |
| RichiH | just to make sure: current swift is 100% api compatible to swift, correct? /me is trying to hack Net::Amazon::S3 to speak to swift | 15:31 |
| imsplitbit | alekibango: I think we will have to disagree on that but I am working towards becoming more proficient | 15:31 |
| *** zenmatt has quit IRC | 15:31 | |
| *** hiddentoken has quit IRC | 15:31 | |
| *** jmckind has quit IRC | 15:31 | |
| *** jmckind has joined #openstack | 15:31 | |
| *** obino has quit IRC | 15:32 | |
| creiht | RichiH: yes, swift is 100% api compatible with swift :) If you were intending to ask about cloudfiles, then that is also mostly true :) | 15:34 |
| *** yamahata_lt has quit IRC | 15:34 | |
| creiht | things like the CDN features are not in swift | 15:34 |
| *** dirkx_ has joined #openstack | 15:38 | |
| *** tomeff has joined #openstack | 15:38 | |
| pgregory | is it possible to get remote access to an instance with having public IP's? | 15:40 |
| *** czajkowski has joined #openstack | 15:41 | |
| *** kbringard has joined #openstack | 15:41 | |
| HugoKuo__ | associate a floatip or fixed_ips in same network of your pulic internet | 15:42 |
| RichiH | creiht: in that case, i have yet to figure out what the correct access key & secret key are. for st, i am using -U foo -K bar and thought those would be my access & secret key | 15:43 |
| creiht | RichiH: are you instead trying to ask if swift is compatible with s3? | 15:45 |
| RichiH | erm | 15:46 |
| RichiH | yes | 15:46 |
| *** zns has quit IRC | 15:46 | |
| creiht | hehe | 15:46 |
| RichiH | ..so.. :) | 15:46 |
| *** naehring has quit IRC | 15:46 | |
| creiht | While there are similarities they are not exactly the same | 15:46 |
| creiht | that said there is an s3 compatibility layer | 15:47 |
| *** zaitcev has joined #openstack | 15:47 | |
| pgregory | HugoKuo__: but doesn't that involve public IP's? | 15:47 |
| creiht | for basic features | 15:47 |
| RichiH | is that enabled by default? | 15:47 |
| creiht | no it is not | 15:47 |
| creiht | Is there a reason that you need s3 compatibility? | 15:47 |
| RichiH | and as i am at the "try to list buckets" stage, i would say basic is fine | 15:47 |
| RichiH | creiht: there are a few tools that rely on s3 which i'd like to be able to use | 15:48 |
| creiht | hrm... just realized there isn't an example swift3 config in the example proxy conf | 15:49 |
| creiht | RichiH: the compatibility layer is a bit experimental, but you can give it a try | 15:50 |
| creiht | http://swift.openstack.org/misc.html#module-swift.common.middleware.swift3 | 15:50 |
| creiht | is the only doc about it at the moment | 15:50 |
| creiht | to enable it... | 15:50 |
| *** koolhead17 has joined #openstack | 15:50 | |
| RichiH | ta | 15:50 |
| pgregory | I'd like to know if there is a way to SSH into a VM from outside the same way I can with an Amazon EC2 instance, without it having a unique public IP. | 15:50 |
| RichiH | fwiw, google did not find that one | 15:50 |
| creiht | RichiH: yeah we need a better overall doc for that | 15:50 |
| creiht | RichiH: http://paste.openstack.org/show/1283/ | 15:52 |
| creiht | you will need to change your proxy server conf file to add the swift3 middleware | 15:52 |
| creiht | my pipeline has a lot of extra stuff, but should give you an idea how to add swift3 to yours | 15:52 |
| creiht | RichiH: the doc that I sent you should have enough info to give you an idea of what you will need for credentials | 15:53 |
| creiht | It has only really been tested with the boto library | 15:53 |
| *** zns has joined #openstack | 15:53 | |
| RichiH | yah, i guessed those, anyway | 15:54 |
| creiht | hehe | 15:54 |
| RichiH | that part is easy/obvious :) | 15:54 |
| *** koolhead17 has quit IRC | 15:54 | |
| *** koolhead17 has joined #openstack | 15:55 | |
| RichiH | creiht: the only thing i need to restart is the swift-proxy, correct? | 15:55 |
| *** dendrobates is now known as dendro-afk | 15:56 | |
| *** zenmatt has joined #openstack | 15:56 | |
| creiht | RichiH: correct | 15:56 |
| hallyn | having trouble spawning instances with openstack on natty - nova-compute.log shows: http://paste.ubuntu.com/603329/ | 15:59 |
| hallyn | shoudl i just hack the .py file to ignore that key? | 16:00 |
| hallyn | or make a change to libvirt setup? | 16:00 |
| *** dendro-afk is now known as dendrobates | 16:00 | |
| *** MarkAtwood has left #openstack | 16:01 | |
| RichiH | creiht: in that case, i need to dig deeper as the perl lib is still unhappy. tcpdump it is | 16:03 |
| creiht | RichiH: someone else had problems trying to get s3curl to work with it as well | 16:03 |
| RichiH | creiht: any idea who that someone was or where to get that info? | 16:04 |
| creiht | my first guess is maybe it is having issues with the extra : in the key | 16:04 |
| RichiH | maybe they fixed it | 16:04 |
| creiht | ? | 16:04 |
| RichiH | who had that problem | 16:04 |
| creiht | RichiH: I don't think so, or at least I hadn't heard anything else | 16:04 |
| creiht | I'm digging up the link... | 16:04 |
| creiht | RichiH: https://answers.launchpad.net/swift/+question/154177 | 16:05 |
| *** zns has quit IRC | 16:06 | |
| RichiH | ta | 16:06 |
| *** zns has joined #openstack | 16:08 | |
| *** fabiand__ has quit IRC | 16:09 | |
| RichiH | cool, my amazonaws pcap makes most current wireshark segfault | 16:10 |
| creiht | lol | 16:11 |
| *** patcoll has joined #openstack | 16:11 | |
| *** RickB17 has joined #openstack | 16:13 | |
| * RichiH is lucky, today | 16:14 | |
| *** deshantm_laptop has quit IRC | 16:17 | |
| *** obino has joined #openstack | 16:17 | |
| *** Vladimir_ has joined #openstack | 16:18 | |
| jaypipes | alekibango: ok, I'm back. going to work on answering your questions now. | 16:18 |
| RichiH | creiht: assuming the colon does indeed cause this error, wouldn't it be possible to simply teach swift3 to expect an underscore or similar? | 16:18 |
| RichiH | hopefully, underscores are not allowed in openstack name | 16:19 |
| RichiH | s | 16:19 |
| hallyn | ok, adding '--use-ipv6=true' to nova.conf worked. Still it looks like a bug in libvirt_conn.py:_get_network_info() to not set mapping['ip6s'] without checking for has_key before using it later | 16:19 |
| kbringard | hallyn: yea | 16:19 |
| kbringard | that's been fixed in trunk, but it's not been pushed to the ppa repos from what I can tell | 16:19 |
| hallyn | d'oh | 16:19 |
| kbringard | yea | 16:20 |
| hallyn | ok, good to know, thanks! | 16:20 |
| kbringard | https://answers.launchpad.net/nova/+question/154771 | 16:20 |
| hallyn | I'm in business, so I"m happy :) | 16:20 |
| kbringard | my problem was deeper... I don't have IP6 stuff set in the DB | 16:20 |
| ccooke | ... Okay, what on earth went wrong with the forum? | 16:20 |
| kbringard | so when I'd go to launch instances it would still fail because the ip addr command would be trying to assign the address "None" | 16:21 |
| kbringard | hah | 16:21 |
| *** nacx has quit IRC | 16:21 | |
| kbringard | no idea why it's not been pushed to the repo though | 16:21 |
| jaypipes | kbringard: heyo. | 16:22 |
| kbringard | jaypipes: what's up homie | 16:22 |
| hallyn | kbringard: phooi :) | 16:22 |
| *** gregp76 has joined #openstack | 16:22 | |
| jaypipes | kbringard: hope all is well with you? :) | 16:22 |
| jaypipes | kbringard: care to give https://code.launchpad.net/~jaypipes/glance/bug773388/+merge/59858 a review? | 16:22 |
| kbringard | jaypipes: yessir, I made it back to CO alive | 16:23 |
| kbringard | sure, let me go check it out | 16:23 |
| jaypipes | coolio. | 16:23 |
| *** ewindisch has joined #openstack | 16:23 | |
| kbringard | how's life in the OH? | 16:23 |
| jaypipes | kbringard: cold and rainy :( | 16:23 |
| *** ewindisch has quit IRC | 16:23 | |
| kbringard | haha, yea, it's cold and snowy here | 16:23 |
| *** ewindisch has joined #openstack | 16:23 | |
| jaypipes | the golf clubs are calling me... | 16:24 |
| *** maplebed has joined #openstack | 16:24 | |
| nhm | Are any of you guys looking at security scanning for your network? Nessus? | 16:24 |
| *** obino has quit IRC | 16:24 | |
| RichiH | creiht: if i changed line 403 in /usr/share/pyshared/swift/common/middleware/swift3.py and /usr/lib/pymodules/python2.6/swift/common/middleware/swift3.py to split on _ instead of :...? | 16:24 |
| kbringard | that is an ephing monster test | 16:25 |
| *** Shentonfreude has joined #openstack | 16:25 | |
| RichiH | though i have no idea how to create a .pyc from a .py -- i will need to find out how to do that tomorrow | 16:25 |
| *** obino has joined #openstack | 16:25 | |
| RichiH | creiht: fwiw, i don't speak python, but i do speak perl. we should be able to make this work :) | 16:26 |
| *** woleium has joined #openstack | 16:26 | |
| ccooke | RichiH: you don't generally need to do anything to get pyc files | 16:27 |
| kbringard | jaypipes: does python have anything like VCR in ruby... it automagically records the network transaction and then replays it in subsequent calls | 16:27 |
| *** throughnothing has joined #openstack | 16:27 | |
| kbringard | for offline testing... could help with your expected stuff | 16:27 |
| kbringard | just a thought | 16:27 |
| jaypipes | kbringard: nothing comes to mind, but creiht might know of something. | 16:27 |
| RichiH | ccooke: so i just move them away and that's it? | 16:27 |
| * RichiH tries | 16:27 | |
| ccooke | RichiH: when you include a .py file from a script, it saves the byte-compiled inclded .py as a .pyc to speed up subsequent loads | 16:28 |
| *** jakedahn has quit IRC | 16:28 | |
| *** jakedahn has joined #openstack | 16:28 | |
| ccooke | kbringard: VCR was a nice tool. I wish that existed on a few other languages | 16:28 |
| kbringard | yea, I love it | 16:29 |
| ccooke | Generally, I love ruby. If only the packaging and threading models could be fixed, it would be my favourite language for anything - and the threading *will* be sorted soonish | 16:30 |
| RichiH | ccooke: k | 16:30 |
| RichiH | creiht: simply changing line 403 and using an underscore in my auth string does not work | 16:31 |
| *** rcc has quit IRC | 16:32 | |
| *** sdadh011 has joined #openstack | 16:32 | |
| kbringard | jaypipes: looks like it should work fine, so long as things below that aren't relying on glance returning an image to determine if it's "ready" | 16:33 |
| nhm | ah, looks like openVAS is the way to go now for opensource vulnerability scanning. | 16:33 |
| kbringard | nhm: yea, they vastly improved how it works | 16:33 |
| kbringard | it was such a pain in the ass in like, version 2 | 16:33 |
| kbringard | but it's way easier now | 16:33 |
| *** kashyap has quit IRC | 16:34 | |
| nhm | kbringard: I need to plan out my production openstack deployment now that I've got a proof-of-concept working. | 16:34 |
| jaypipes | kbringard: cool, thx for the review. | 16:34 |
| kbringard | heh, no problem :-) | 16:35 |
| RichiH | creiht: ah, if i understand correctly, the split is done several times and thus i can't simply replace : with _ | 16:35 |
| kbringard | it was such an inconvenience ;-) | 16:35 |
| nhm | So vulnerability scanning, ldap integration, remote file system mounts, configuring vlans, blah balh blah. :) | 16:35 |
| kbringard | nhm: sounds like a good time | 16:36 |
| nhm | At least I've got the puppet/tftp/kickstart stuff more or less done for the underlying cluster. | 16:36 |
| *** sdadh01 has quit IRC | 16:36 | |
| nhm | kbringard: this kind of stuff is actually something that the community could use more documentation on. It's (relatively) easy to get a test setup up and going. | 16:37 |
| kbringard | nhm: have you looked at Dell's crowbar stuff? | 16:37 |
| kbringard | http://en.community.dell.com/dell-blogs/enterprise/b/inside-enterprise-it/archive/2011/03/16/openstack-installer-demo-at-sxsw.aspx | 16:37 |
| kbringard | they were demoing it at the design conference last week | 16:37 |
| kbringard | looks pretty nifty | 16:37 |
| antonym | I yanked a list of all the etherpads created during the Diablo Summit and posted them here: http://wiki.openstack.org/DiabloSummitEtherpads | 16:38 |
| nhm | kbringard: I've just kind of glanced at it. It would be more useful if we didn't already have tons of clusters and remotely managed workstations here. | 16:39 |
| kbringard | yea, makes sense | 16:39 |
| *** troytoman-away is now known as troytoman | 16:39 | |
| *** _h0c1n_ has joined #openstack | 16:40 | |
| *** _h0c1n_ has joined #openstack | 16:40 | |
| *** _h0c1n_ has quit IRC | 16:40 | |
| nhm | If all you want to do is deploy cloud infrastructure it could be great though. My task is to figure out how to integrate an openstack cloud with our traditional HPC resources in a way that makes sense. ;) | 16:40 |
| *** joearnold has joined #openstack | 16:40 | |
| *** Vladimir__ has joined #openstack | 16:40 | |
| kbringard | anyone know why the trunk ppa repo is on 1039 still? | 16:42 |
| *** sdadh011 has quit IRC | 16:42 | |
| OutBackDingo | kbringard: yeah too bad dells only allowing testing of crowbar on their hardware solution also | 16:44 |
| *** gondoi has quit IRC | 16:44 | |
| *** cole has quit IRC | 16:45 | |
| kbringard | OutBackDingo: yea, agreed | 16:46 |
| OutBackDingo | kbringard: itll heavily limit their scope | 16:46 |
| kbringard | although, the guy said that the idea is that they opened up the... APIs? I guess for lack of a better word... so that other hardware manufacturers can write little plugins for their BIOSs | 16:47 |
| OutBackDingo | kbringard: which is kind of a joke | 16:47 |
| kbringard | so if HP for instance, wanted to provide crowbar support, they could just write a little thing | 16:47 |
| kbringard | yea, *shrug* | 16:47 |
| kbringard | who knows | 16:47 |
| *** jamesurquhart has quit IRC | 16:47 | |
| *** sdadh01 has joined #openstack | 16:48 | |
| kbringard | it's pretty obvious that the ultimate point is to sell Dell hardware | 16:48 |
| OutBackDingo | though im biased ive never had 1 good experience with dell | 16:48 |
| OutBackDingo | ill stick with Intel systems | 16:48 |
| jaypipes | alekibango, vishy, sirp_, kbringard: I just created http://etherpad.openstack.org/GlanceSharedImageGroups. Let me know what you think... | 16:49 |
| kbringard | wow, that's weird to see it change while I'm reading it | 16:49 |
| kbringard | haha | 16:49 |
| *** jamesurquhart has joined #openstack | 16:49 | |
| jaypipes | alekibango, vishy, sirp_, kbringard: obviously dependent on Keystone integration, but I think it's worth exploring the API now. | 16:50 |
| *** kashyap has joined #openstack | 16:51 | |
| *** Dumfries has joined #openstack | 16:53 | |
| *** photron has quit IRC | 16:53 | |
| *** mahadev has joined #openstack | 16:53 | |
| kbringard | one real quick ?... so anyone can still do a GET /images/<IMAGE ID>/shares... or a GET /images/<IMAGE ID> or whatever? or will it do an auth check before returning any of this stuff? | 16:53 |
| KnightHacker | jaypipes: How can I get all the images that are shared under my tenant? | 16:53 |
| kbringard | I see you have a 401 unauthorized deal, but it's under the POST, so I wasn't sure if that cascaded up to all the calls | 16:53 |
| jaypipes | KnightHacker: GET /images/<ID>/shares | 16:54 |
| jaypipes | KnightHacker: oh, sorry... | 16:54 |
| jaypipes | KnightHacker: good point. | 16:54 |
| KnightHacker | :) thanks | 16:54 |
| kbringard | yea, that was my next question :-) | 16:54 |
| jaypipes | KnightHacker, kbringard: think it should be something like: /images/<ID>/shared-images? | 16:54 |
| jaypipes | or should I try to adapt the return of GET /images/<ID>/shares to also return an images collection? | 16:55 |
| kbringard | hmm, it makes more sense to me | 16:55 |
| kbringard | or rather... hrmm | 16:55 |
| KnightHacker | Well, I was thinking it would be something like GET /<tenantID>/images | 16:55 |
| kbringard | sorry, let me stop and think | 16:55 |
| KnightHacker | This should return all images collection | 16:55 |
| KnightHacker | Then we can have a filter on the "shared" ones. | 16:56 |
| KnightHacker | GET /<tenantID>/images?shared or whatever. | 16:56 |
| jaypipes | KnightHacker: you mean GET /tenants/<TENANT_ID/images? | 16:56 |
| KnightHacker | If you don't provide <tenantID> then you are asking for the public ones. | 16:56 |
| KnightHacker | Sure | 16:56 |
| jaypipes | KnightHacker: yeah, that makes sense. lemme add that to the spec. | 16:56 |
| KnightHacker | That way you can authorize that the requestor is asking for images under the tenant that he belogs to .. or delegated to. | 16:57 |
| *** zns has quit IRC | 16:57 | |
| *** GasbaKid has joined #openstack | 16:57 | |
| kbringard | does anyone know the status of live migration with KVM? | 16:59 |
| *** GasbaKid has quit IRC | 17:00 | |
| *** thingee has joined #openstack | 17:00 | |
| *** GasbaKid has joined #openstack | 17:02 | |
| *** zns has joined #openstack | 17:02 | |
| kbringard | ooooooo, it looks like tmpfile needs to be on the nfs mount as well? | 17:03 |
| Dumfries | hey all, attempting to get persistent storage going, when I start nova-volumes I am getting the "no such table: volumes ..." I have ensure it is using proper db. any one bump into this ? | 17:03 |
| *** Vladimir__ has quit IRC | 17:06 | |
| *** winston-d has quit IRC | 17:06 | |
| *** bcwaldon has joined #openstack | 17:07 | |
| *** Vladimir_ has quit IRC | 17:07 | |
| Dumfries | My nova.conf looks like .... http://pastie.org/1864814 | 17:07 |
| *** GasbaKid has quit IRC | 17:09 | |
| *** johnpur has joined #openstack | 17:10 | |
| *** ChanServ sets mode: +v johnpur | 17:10 | |
| *** GasbaKid has joined #openstack | 17:11 | |
| *** tblamer has joined #openstack | 17:12 | |
| *** amccabe has quit IRC | 17:14 | |
| *** GasbaKid has quit IRC | 17:14 | |
| *** Ryan_Lane has joined #openstack | 17:15 | |
| *** bcwaldon has quit IRC | 17:16 | |
| *** GasbaKid has joined #openstack | 17:16 | |
| *** hub_cap has quit IRC | 17:18 | |
| *** msivanes has quit IRC | 17:19 | |
| *** GasbaKid has quit IRC | 17:20 | |
| *** deshantm_laptop has joined #openstack | 17:21 | |
| *** dirkx_ has quit IRC | 17:22 | |
| *** GasbaKid has joined #openstack | 17:22 | |
| *** foxtrotgulf has joined #openstack | 17:22 | |
| *** dendrobates is now known as dendro-afk | 17:22 | |
| kbringard | anyone familiar with this: | 17:24 |
| kbringard | Cannot confirm tmpfile at /var/lib/nova/instances is on same shared storage | 17:24 |
| kbringard | when attempting to live migrate a vm? | 17:24 |
| kbringard | that path is on an nfs mount which both nodes have mounted | 17:24 |
| kbringard | to the same logical path | 17:24 |
| HugoKuo__ | how to check the correction of instance metadata ? | 17:25 |
| nhm | kbringard: alas, you are farther ahead than me. :) | 17:25 |
| kbringard | someone awhile back said live migration was bugged | 17:25 |
| kbringard | but I've not been able to get anyone to confirm it, nor can I find a bug report about it | 17:26 |
| *** bcwaldon has joined #openstack | 17:26 | |
| nhm | kbringard: probably not enough people testing it yet. | 17:26 |
| nhm | kbringard: though I would think the NASA guys must have it working. | 17:26 |
| *** GasbaKid has quit IRC | 17:26 | |
| kbringard | one would think | 17:26 |
| *** GasbaKid has joined #openstack | 17:28 | |
| j05h | kbringard: looks like it is having a problem creating a file on the destination host. | 17:29 |
| *** dprince has joined #openstack | 17:29 | |
| j05h | "This method enables compute nodes to recognize that they mounts same shared storage." | 17:29 |
| j05h | but maybe you knew that. | 17:29 |
| kbringard | j05h: yea, I saw that, but they do have the same shared storage mounted | 17:29 |
| kbringard | so either I'm missing a setting or it's not doing what it's supposed to :-) | 17:30 |
| *** GasbaKid has quit IRC | 17:31 | |
| *** deshantm_laptop_ has joined #openstack | 17:31 | |
| dprince | mtaylor: around? | 17:31 |
| *** irahgel has left #openstack | 17:32 | |
| *** dmi_ has joined #openstack | 17:33 | |
| *** deshantm_laptop has quit IRC | 17:34 | |
| *** deshantm_laptop_ is now known as deshantm_laptop | 17:34 | |
| *** evil_e has quit IRC | 17:35 | |
| *** dprince has quit IRC | 17:35 | |
| *** evil_e has joined #openstack | 17:36 | |
| *** evil_e has quit IRC | 17:40 | |
| jaypipes | alekibango: added some answers to that etherpad on images... I'll work on adding examples for the openstack API. | 17:41 |
| *** evil_e has joined #openstack | 17:48 | |
| *** MarkAtwood has joined #openstack | 17:49 | |
| jaypipes | sirp_: added some commentary to http://etherpad.openstack.org/GlanceSharedImageGroups about ownership vs. membership... | 17:56 |
| jaypipes | KnightHacker, sirp_: feel free to comment/update | 17:57 |
| sirp_ | jaypipes: cool will checkout | 17:57 |
| *** zns has quit IRC | 17:58 | |
| *** hub_cap has joined #openstack | 17:59 | |
| KnightHacker | jaypipes: Will do. Thanks. | 17:59 |
| *** jamesurquhart has quit IRC | 18:00 | |
| *** zns has joined #openstack | 18:03 | |
| *** jonkelly has joined #openstack | 18:03 | |
| *** deshantm_laptop has quit IRC | 18:04 | |
| *** hiddentoken has joined #openstack | 18:04 | |
| *** jmckind has quit IRC | 18:04 | |
| *** obino has quit IRC | 18:05 | |
| *** obino has joined #openstack | 18:06 | |
| *** bcwaldon has quit IRC | 18:07 | |
| jaypipes | zns: ready for me in the keystone meeting? | 18:07 |
| *** MarkAtwood has quit IRC | 18:07 | |
| KnightHacker | jaypipes: I was about to ping you on that. | 18:07 |
| *** MarkAtwood has joined #openstack | 18:08 | |
| *** msivanes has joined #openstack | 18:08 | |
| KnightHacker | Did you get the invite? | 18:08 |
| zns | jaypipes: yes. | 18:08 |
| jaypipes | zns: link to meeting agenda? | 18:09 |
| alekibango | re... thanks jaypipes... will look on it now | 18:12 |
| *** photron_ has joined #openstack | 18:13 | |
| *** hub_cap has quit IRC | 18:14 | |
| *** hub_cap has joined #openstack | 18:14 | |
| vishy | alekibango: the images end up in glance if you use glance service and ec2 api | 18:15 |
| alekibango | vishy: i am nto sure if i understand how | 18:17 |
| alekibango | is there some url you can point me to? | 18:17 |
| alekibango | vishy: we are editing http://etherpad.openstack.org/glance-walkthrough | 18:17 |
| jdurgin | creiht: do you know when you plan on having the driver api settled so other backends can be added to lunr? | 18:19 |
| creiht | kbringard: I'm not aware of any tool like VCR in python | 18:22 |
| *** jamesurquhart has joined #openstack | 18:23 | |
| creiht | jdurgin: it may still be a bit before we get to that point | 18:24 |
| vishy | alekibango: one sec. In a meeting will help soon | 18:24 |
| alekibango | vishy: np. i can wait. thanks alot | 18:24 |
| creiht | RichiH: yeah it would probably be easier to figure out how to fix that library if it doesn't work :) | 18:24 |
| RichiH | creiht: well, i confirmed that i am sending everything _exactly_ as i am sending it to s3 | 18:25 |
| RichiH | creiht: i can extract a text dump with removed passwords if you want | 18:26 |
| *** johnpur has quit IRC | 18:26 | |
| *** clauden_ has joined #openstack | 18:27 | |
| jdurgin | creiht: is that a few months, or not until diablo? | 18:27 |
| alekibango | jdurgin: http://wiki.openstack.org/DiabloReleaseSchedule | 18:28 |
| *** throughnothing has quit IRC | 18:28 | |
| creiht | jdurgin: I would hope sooner than a few months :) | 18:28 |
| alekibango | September 22nd - diablo release :) | 18:29 |
| Dumfries | is there a simple way to migrate from sqlite to mysql while preserving existing data | 18:29 |
| creiht | RichiH: The problem with the s3curl.pl script was that it wasn't signing the request correctly | 18:29 |
| *** throughnothing has joined #openstack | 18:29 | |
| RichiH | creiht: you mean base64 encode the password? | 18:30 |
| RichiH | creiht: if yes, as i am using s3ls to connect to s3, it has to encode correctly | 18:31 |
| creiht | well each request is signed (and the backend knows the secret key so it can verify the signature) | 18:31 |
| *** santhosh has joined #openstack | 18:32 | |
| creiht | My first hunch was that the : messed up the parsing of the secret key, so it signed the request incorrectly | 18:32 |
| vishy | alekibango: FYI, i gave a pretty detailed description of the interactions here: https://answers.launchpad.net/glance/+question/153638 | 18:32 |
| creiht | but I'm not very proficient in Perl, so I couldn't debug further | 18:32 |
| alekibango | vishy: i think i have seen that before... but i will read it vigorously again | 18:33 |
| vishy | alekibango: it doesn't specify how you could replace nova-objectstore with swift/s3 front end. I don't know if anyone has tried it yet | 18:34 |
| vishy | alekibango: but glance/swift back end means the images all end up in swift | 18:34 |
| alekibango | vishy: jaypipes suggested that it should not be neeed to use s3 | 18:34 |
| alekibango | that we can use openstack tools for client uploading images | 18:35 |
| alekibango | but apart from glance tool i fail to see how | 18:35 |
| vishy | alekibango: unfortunately, we don't really have any control of ec2 api so it will need an s3 clone | 18:35 |
| alekibango | vishy: if i can provide endusers *some* way to upload images | 18:36 |
| vishy | alekibango: but it is possible to upload images without using euca commands. nova-manage image commands work just fine for example | 18:36 |
| alekibango | without putting glance api on public network unprotected.... i would be happy | 18:36 |
| vishy | alekibango: well glance on public will be fine when auth is in place... | 18:36 |
| vishy | :) | 18:36 |
| alekibango | heh right | 18:36 |
| *** alandman has joined #openstack | 18:36 | |
| *** zenmatt has quit IRC | 18:36 | |
| alekibango | but that will take how long? months? | 18:37 |
| *** zenmatt has joined #openstack | 18:37 | |
| alekibango | i need to start public demo soon | 18:37 |
| alekibango | and i sure do not want users to be able to put whatever content into my swift backend by using glance api | 18:37 |
| vishy | alekibango: i would suspect it is a couple months off | 18:38 |
| *** obino has quit IRC | 18:38 | |
| alekibango | i was said diablo D2... June 30th | 18:38 |
| alekibango | which is 2 months... hmm | 18:38 |
| creiht | alekibango: the diablo release is now in September | 18:39 |
| *** agarwalla has joined #openstack | 18:39 | |
| j05h | dammit…my brain read that as the Diablo 3 release date is June 30th. Now I am disappoint. | 18:39 |
| creiht | lol | 18:40 |
| alekibango | http://wiki.openstack.org/DiabloReleaseSchedule | 18:40 |
| *** MarkAtwood has quit IRC | 18:40 | |
| vishy | sorry alekibango, must hit another meeting | 18:40 |
| creiht | hah... didn't realize they had split it into separate releases like that | 18:40 |
| vishy | I'll be back on later this afternoon | 18:41 |
| vishy | creiht: they aren't releases | 18:41 |
| vishy | they are milestones | 18:41 |
| creiht | vishy: ahh, well that isn't exactly evident from that page | 18:41 |
| *** jamesurquhart has quit IRC | 18:42 | |
| *** jamesurquhart has joined #openstack | 18:42 | |
| mtaylor | dprince: morning hombre! | 18:43 |
| *** MarkAtwood has joined #openstack | 18:44 | |
| *** alekibango has quit IRC | 18:44 | |
| *** Joelio has joined #openstack | 18:45 | |
| *** matt25 has joined #openstack | 18:48 | |
| *** throughnothing has quit IRC | 18:49 | |
| *** alekibango has joined #openstack | 18:52 | |
| *** troytoman is now known as troytoman-away | 18:55 | |
| *** santhosh has quit IRC | 18:58 | |
| *** jmckind has joined #openstack | 19:01 | |
| *** hiddentoken has quit IRC | 19:02 | |
| *** cjonrun has quit IRC | 19:04 | |
| *** dendro-afk is now known as dendrobates | 19:05 | |
| kbringard | creiht: yea, no biggie, I was just looking at Jay's expect return stuff and thinking that it'd be sweet to have the test use real life transactions that are captured and replayed | 19:10 |
| alandman | Hello I'm new to using Openstack and am trying to get it running on HyperV | 19:11 |
| alandman | I've followed the steps at http://wiki.openstack.org/HypervInstall and running into an error and hoping someone can point me in the right direction | 19:11 |
| alandman | Why I try to launch python bin\nova-compute I get an error loading module wsgi.py | 19:12 |
| alandman | The error is "ImportError: No module named routes" | 19:12 |
| alandman | Am I missing a python module? | 19:13 |
| kbringard | alandman: yea, looks like the routes module isn't installed | 19:14 |
| kbringard | what OS are you running? | 19:14 |
| alandman | Windows 2008 R2 | 19:14 |
| kbringard | Windows? | 19:14 |
| kbringard | ah, hrmm, sorry, I have little experience with Windows in general, and none in regards to OpenStack | 19:15 |
| kbringard | in theory when you install python it should install an easy_install command, and/or a pip command | 19:15 |
| kbringard | you should be able to do easy_install wsgi | 19:15 |
| kbringard | which should in theory install all the deps | 19:16 |
| kbringard | if it says it's already installed, then you should be able to do easy_install routes | 19:16 |
| kbringard | ... at least that's how it works in *nix | 19:16 |
| *** openstackjenkins has quit IRC | 19:16 | |
| *** sward_ has joined #openstack | 19:17 | |
| *** openstackjenkins has joined #openstack | 19:17 | |
| alandman | I will just step throught depdencies then thanks | 19:17 |
| *** sward_ has quit IRC | 19:17 | |
| *** MarkAtwood has left #openstack | 19:18 | |
| *** sward_ has joined #openstack | 19:18 | |
| *** sward_ is now known as namaqua | 19:18 | |
| kbringard | alandman: no problem, best of luck sir | 19:18 |
| *** namaqua has joined #openstack | 19:18 | |
| *** openstackjenkins has quit IRC | 19:23 | |
| *** openstackjenkins has joined #openstack | 19:24 | |
| *** Shentonfreude has quit IRC | 19:26 | |
| *** dendrobates is now known as dendro-afk | 19:26 | |
| kbringard | anyone have experience with the CloudPipe stuff? | 19:35 |
| *** joearnold has quit IRC | 19:35 | |
| kbringard | it all makes sense, but I'm wondering how you get the VPN client side data? | 19:35 |
| *** ianweller has quit IRC | 19:36 | |
| *** ewindisch has quit IRC | 19:37 | |
| *** jakedahn has quit IRC | 19:37 | |
| *** ianweller has joined #openstack | 19:39 | |
| *** alandman has quit IRC | 19:39 | |
| *** ianweller is now known as Guest73664 | 19:40 | |
| xtoddx | kbringard: nova-mange can spit out the credentials. then you just email them or whatever is convenient | 19:41 |
| *** dendro-afk is now known as dendrobates | 19:42 | |
| kbringard | ah, cool, I figured it would be in there somewhere | 19:42 |
| kbringard | do you happen to know which sections it's under? I've checked vpn and project, but don't see it there | 19:43 |
| *** DigitalFlux has quit IRC | 19:43 | |
| *** DigitalFlux has joined #openstack | 19:44 | |
| *** DigitalFlux has joined #openstack | 19:44 | |
| *** namaqua has quit IRC | 19:46 | |
| *** namaqua has joined #openstack | 19:47 | |
| *** jmeredit has joined #openstack | 19:49 | |
| *** obino has joined #openstack | 19:49 | |
| *** rdw has quit IRC | 19:50 | |
| *** rdw has joined #openstack | 19:50 | |
| *** lborda has quit IRC | 19:51 | |
| *** lborda has joined #openstack | 19:53 | |
| *** _nil has joined #openstack | 19:54 | |
| _nil | hey | 19:54 |
| xtoddx | kbringard: i thought it was project zipfile? | 19:55 |
| kbringard | hmmmm, I thought that was just your nova creds and the certs to bundle images, but let me reread this, it may be both | 19:55 |
| *** openstack has joined #openstack | 19:58 | |
| namaqua | any idea when the openstack wiki for release cycle will be updated to reflect any changes discussed at the design summit? | 20:00 |
| *** obino has quit IRC | 20:00 | |
| *** obino has joined #openstack | 20:00 | |
| xtoddx | kbringard: yea, looks like project zipfile calls AuthManager's get_credentials method, which checks for vpn information, so I'd look there to understand what's happening | 20:00 |
| *** erik-s|detached is now known as erik-s | 20:00 | |
| kbringard | ah, you know, you may be right... come to think of it I've not exported a zipfile since setting up cloudpipe, so it's likely that if the VPN data exists then it'll export it | 20:01 |
| kpepple | namaqua: Diablo release schedule is updated at http://wiki.openstack.org/DiabloReleaseSchedule | 20:01 |
| kbringard | thanks, you rock | 20:02 |
| *** pothos_ has joined #openstack | 20:08 | |
| *** n1md4 has left #openstack | 20:09 | |
| *** pothos has quit IRC | 20:09 | |
| *** pothos_ is now known as pothos | 20:10 | |
| *** amccabe has joined #openstack | 20:14 | |
| *** dendrobates is now known as dendro-afk | 20:15 | |
| jaypipes | sandywalsh: lmao. | 20:16 |
| *** jakedahn has joined #openstack | 20:18 | |
| _nil | where is the information about starting to contribute? | 20:20 |
| _nil | is there any OpenMP or MPI in the codebase? | 20:20 |
| sandywalsh | :) | 20:21 |
| kpepple | _nil: http://wiki.openstack.org and no | 20:21 |
| *** deepy_ is now known as deepy | 20:23 | |
| *** deepy has joined #openstack | 20:24 | |
| *** photron_ has quit IRC | 20:28 | |
| *** allsystemsarego has quit IRC | 20:28 | |
| *** mahadev_ has joined #openstack | 20:32 | |
| *** mahadev has quit IRC | 20:34 | |
| hallyn | I dunno, uploading my own kernel image (following https://help.ubuntu.com/community/UEC/CreateYourImage as a guide) is not working out for me. the images never get out of 'decrypting' state, and I keep these in nova-api.log: | 20:38 |
| hallyn | 2011-05-04 15:37:07,260 DEBUG nova.utils [-] Running cmd (subprocess): openssl rsautl -decrypt -inkey /var/lib/nova/CA/private/cakey.pem from (pid=2400) execute /usr/lib/pymodules/python2.7/nova/utils.py:141 | 20:38 |
| hallyn | 2011-05-04 15:37:07,270 DEBUG nova.utils [-] Result was 1 from (pid=2400) execute /usr/lib/pymodules/python2.7/nova/utils.py:157 | 20:38 |
| kbringard | does anyone know why the ppa trunk maverick repo is still showing build 1039? | 20:39 |
| *** tblamer has quit IRC | 20:41 | |
| hallyn | hm, maybe a combination of chowning and rebooting will help | 20:42 |
| hallyn | nope | 20:45 |
| *** jakedahn has quit IRC | 20:48 | |
| *** jakedahn has joined #openstack | 20:48 | |
| *** ewindisch has joined #openstack | 20:49 | |
| *** joearnold has joined #openstack | 20:53 | |
| *** mahadev_ has quit IRC | 20:54 | |
| *** mahadev has joined #openstack | 20:55 | |
| *** jakedahn_ has joined #openstack | 20:56 | |
| *** ctennis has quit IRC | 20:57 | |
| vishy | kbringard: the project zip includes vpn creds | 20:58 |
| vishy | kbringard: and you can get it from the dashboard as well | 20:58 |
| kbringard | ah, and if it doesn't, it's because I don't have VPN stuff setup, yea? | 20:58 |
| kbringard | that's perfect, thank you sir | 20:59 |
| *** Guest73664 is now known as ianweller | 20:59 | |
| *** ianweller has joined #openstack | 20:59 | |
| *** koolhead17 has quit IRC | 21:00 | |
| *** jakedahn has quit IRC | 21:00 | |
| *** jakedahn_ is now known as jakedahn | 21:00 | |
| _nil | kpepple: thanks. | 21:01 |
| _nil | i'd like to help if i can | 21:01 |
| *** h0cin has quit IRC | 21:02 | |
| *** mgoldmann has quit IRC | 21:02 | |
| *** OutBackDingo has quit IRC | 21:03 | |
| *** kbringard_ has joined #openstack | 21:09 | |
| *** kbringard has quit IRC | 21:09 | |
| *** kbringard_ is now known as kbringard | 21:09 | |
| *** mahadev_ has joined #openstack | 21:10 | |
| *** mahadev has quit IRC | 21:10 | |
| *** mahadev has joined #openstack | 21:11 | |
| *** Ryan_Lane1 has joined #openstack | 21:12 | |
| *** Ryan_Lane1 has joined #openstack | 21:12 | |
| *** Ryan_Lane has quit IRC | 21:12 | |
| *** Ryan_Lane1 is now known as Ryan_Lane | 21:12 | |
| *** ctennis has joined #openstack | 21:13 | |
| *** ctennis has joined #openstack | 21:13 | |
| *** deshantm has quit IRC | 21:17 | |
| *** msivanes has left #openstack | 21:24 | |
| *** imsplitbit has quit IRC | 21:27 | |
| *** jdurgin has quit IRC | 21:28 | |
| *** namaqua has quit IRC | 21:29 | |
| kpepple | _nil: cool, there are lots of place to help (docs, code, etc) | 21:30 |
| *** deshantm has joined #openstack | 21:38 | |
| *** zns has quit IRC | 21:40 | |
| openstackjenkins | Project nova build #880: SUCCESS in 2 min 40 sec: http://jenkins.openstack.org/job/nova/880/ | 21:44 |
| openstackjenkins | Tarmac: Don't fail the test suite in the absence of VCS history. | 21:44 |
| kbringard | hallyn: are you using the prebuilt images? | 21:47 |
| *** jmeredit has quit IRC | 21:48 | |
| Dumfries | running "euca-describe-instances" yields "Unexpected error raised: invalid literal for int() with base 10: 'ami-u0ndzalr'" after upgrading. | 21:53 |
| kpepple | Dumfries: did you convert your images via the release notes instructions ? | 21:54 |
| Dumfries | kpepple: I did, yes. | 21:54 |
| Dumfries | I even went as far as removing the old images, and upload using the new method. However I am still getting the same errro. | 21:55 |
| Dumfries | error* | 21:55 |
| *** clauden_ has quit IRC | 21:55 | |
| kpepple | Dumfries: do you get a stack trace in nova-api.log that you can paste on paste.openstack.org ? this sounds familiar ... | 21:57 |
| Dumfries | kpepple: sure one moment. | 21:57 |
| *** mahadev_ has joined #openstack | 21:58 | |
| Dumfries | kpepple: http://paste.openstack.org/show/1286/ | 21:58 |
| * kpepple fires up vim to peer into cloud.py | 21:59 | |
| Dumfries | :) | 21:59 |
| *** mahadev_ has quit IRC | 22:00 | |
| *** hub_cap has quit IRC | 22:00 | |
| *** mahadev has quit IRC | 22:01 | |
| *** mahadev has joined #openstack | 22:01 | |
| *** jmckind has quit IRC | 22:02 | |
| *** GasbaKid has joined #openstack | 22:02 | |
| *** _vinay has joined #openstack | 22:02 | |
| _vinay | Hi | 22:02 |
| _vinay | I am using the nova_adminclient module | 22:02 |
| _vinay | and I need to increase the quota for a project | 22:03 |
| _nil | kpepple: i'm mostly a C/Go programmer now | 22:03 |
| _nil | so i'll just need to read some stuff | 22:03 |
| *** agarwalla has quit IRC | 22:03 | |
| kpepple | Dumfries: you are using Cactus release version correct ? | 22:04 |
| Dumfries | kpepple: 2011.3~bzr1039 | 22:04 |
| *** OutBackDingo has joined #openstack | 22:04 | |
| _vinay | It looks like the method to call this isnt there | 22:04 |
| _vinay | any ideas ? | 22:04 |
| *** tomeff has quit IRC | 22:07 | |
| *** tomeff has joined #openstack | 22:07 | |
| *** clauden_ has joined #openstack | 22:08 | |
| *** tomeff has quit IRC | 22:08 | |
| openstackjenkins | Project nova build #881: SUCCESS in 2 min 40 sec: http://jenkins.openstack.org/job/nova/881/ | 22:09 |
| openstackjenkins | Tarmac: Fixes for nova-manage vpn list | 22:09 |
| kpepple | Dumfries: this seems like it is a bug ... it's passing the image_id where it looks like it should be passing the internal (non-ec2) instance_id | 22:10 |
| Dumfries | kpepple: lovely, rekon there is a way around it? | 22:10 |
| *** patcoll has quit IRC | 22:11 | |
| *** dendro-afk is now known as dendrobates | 22:13 | |
| kpepple | Dumfries: hold on .. let me see if vishy is around | 22:13 |
| Dumfries | kpepple: cheers | 22:13 |
| kpepple | vishy: got a question on some code in nova/api/ec2/cloud.py that you changed just before Cactus RC (cloud.py:709) ... you around ? | 22:14 |
| *** knorthcut has quit IRC | 22:19 | |
| kpepple | Dumfries: soooo ... i *believe* the problem here is that we are passing the image_id to ec2utils.id_to_ec2_id where we should be passing the instance_id ... this could be because we have a bug in the code but it also could be that your nova-manage image_convert went south | 22:20 |
| *** dirkx_ has joined #openstack | 22:21 | |
| kpepple | Dumfries: what image service are your using ? | 22:21 |
| Dumfries | I betting on the conversion going south. | 22:21 |
| uvirtbot | New bug: #777460 in nova "OS API xml deserialization missing imageRef and flavorRef" [Undecided,New] https://launchpad.net/bugs/777460 | 22:21 |
| Dumfries | kpepple: I was using the default in bexar, haven't moved up the glance yet. | 22:21 |
| *** kbringard has quit IRC | 22:22 | |
| *** dirkx_ has quit IRC | 22:22 | |
| kpepple | Dumfries: if you are using local imagestore, list the directory where the images are stored ... | 22:22 |
| Dumfries | sure | 22:22 |
| Dumfries | kpepple: I removed them manually previously. | 22:23 |
| Dumfries | they were under "buckets" but are now gone. | 22:23 |
| kpepple | Dumfries: my machine just crashed :( | 22:25 |
| Dumfries | kpepple: hell of a day for the both of us it would seem ;) | 22:25 |
| kpepple | Dumfries: try uploading a new image and then launching that | 22:25 |
| Dumfries | I can bring up a new instance with the new image however the same error is yielded when running "euca-describe-instances"/ | 22:26 |
| Dumfries | kpepple: ^^ | 22:30 |
| kpepple | Dumfries: i missed that one ... switching between iphone and mac ... | 22:31 |
| Dumfries | Ah, got it ;) | 22:31 |
| pgregory | is it possible to have a single network (FlatDHCPManager) and have all projects associate with that? | 22:32 |
| kpepple | Dumfries: do you have any instances running that don't have images in your image directory ? | 22:32 |
| *** jamesurquhart1 has joined #openstack | 22:33 | |
| *** jamesurquhart has quit IRC | 22:33 | |
| Dumfries | kpepple: Yes | 22:33 |
| *** lborda has quit IRC | 22:36 | |
| uvirtbot | New bug: #777467 in swift "No command line utility exists for setting container ACL's" [Undecided,New] https://launchpad.net/bugs/777467 | 22:38 |
| *** jakedahn has quit IRC | 22:38 | |
| *** rnirmal has quit IRC | 22:39 | |
| *** jmckind has joined #openstack | 22:40 | |
| pgregory | anyone? | 22:42 |
| *** miclorb has joined #openstack | 22:45 | |
| *** donkdonk has joined #openstack | 22:45 | |
| *** donkdonk has left #openstack | 22:45 | |
| *** mray has quit IRC | 22:45 | |
| vishy | pgregory: yes, that is the way flatdhcp works | 22:46 |
| vishy | pgregory: they only get their own networks in vlan mode | 22:46 |
| pgregory | vishy: I've created a new project, and when I try to run an instance on it, I get an error in nova-compute.log NoMoreNetwork. | 22:47 |
| pgregory | any idea why? | 22:47 |
| vishy | pgregory: probably you are using vlan mode somewhere? | 22:47 |
| kpepple | Dumfries: sorry, interrupted by phone call. | 22:47 |
| Dumfries | kpepple: no worries, appreciate the help. | 22:48 |
| pgregory | vishy: hmm, looks like you might be right, I'll check it out. | 22:48 |
| *** jamesurquhart1 has quit IRC | 22:48 | |
| kpepple | Dumfries: it could be that the describe instances call is trying to lookup the image_id of something that doesn't exist anymore ... but it's strange that it's failing there | 22:49 |
| Dumfries | kpepple: Is there a way to go about removing the reference or updating it to point to another image id? | 22:50 |
| kpepple | Dumfries: i don't think so ... not on a running instance | 22:51 |
| vishy | pgregory: you have to be careful switching modes | 22:51 |
| Dumfries | well at least to get me to a point where I could spawn new instances to replace the old ones. | 22:51 |
| kpepple | Dumfries: unless you hacked it in the database | 22:51 |
| vishy | pgregory: generally the best strategy is to recreate the db | 22:51 |
| Dumfries | kpepple: I am not against that to be honest. | 22:52 |
| vishy | pgregory: (although you can manually go in and try to edit it, it is easy to mess things up) | 22:52 |
| pgregory | vishy: seems to be ok, I switched mode and can now launch instances in either project (via Dashboard) | 22:52 |
| kpepple | Dumfries: you'd have to update the instance's image_id to the new image's image_id ... it's an integer number, but i am not sure how it is generated. you can probably find out by reading nova-manage source code | 22:53 |
| pgregory | vishy: for future reference though, how do I clear and recreate the db? | 22:53 |
| vishy | pgregory: delete the sqlite file (or drop the db in mysql) | 22:53 |
| Dumfries | kpepple: it is an option, although my python is more than a little rusty | 22:53 |
| vishy | pgregory: then run all the basic nova-manage commands you did when starting... | 22:53 |
| vishy | nova-manage db sync | 22:53 |
| vishy | nova-manage network create ... | 22:54 |
| *** markvoelker has quit IRC | 22:54 | |
| pgregory | vishy: thanks, I'll make a note of that. | 22:54 |
| vishy | nova-manage user admin ... | 22:54 |
| vishy | etc. | 22:54 |
| * pgregory can't beleive what a difference a day makes. | 22:54 | |
| vishy | pgregory: sorry the docs were so tough to get through | 22:54 |
| vishy | pgregory: I'll have a chat with anne and see if we can get some people testing all of the ones that are in the docs and remove any that are outdated | 22:55 |
| pgregory | this time yesterday I was frustrated to death, now I have an excellent proof of concept running with Web interface, multiple users, and multiple projects. | 22:55 |
| *** MarkAtwood has joined #openstack | 22:55 | |
| pgregory | vishy: if you want any help doing that I'd be happy to try some stuff out, or at least share my findings thus far. | 22:55 |
| pgregory | I'm beginning to get a proper feel for the whole system now. | 22:56 |
| pgregory | I've got VNC proxy working too, but am interested if there are any other options for remotely accessing an instance apart from public IP's? | 22:57 |
| *** MotoMilind has quit IRC | 22:57 | |
| pgregory | for example, Amazon allows you to access an instance via a url, without having to have a costly public IP associated with it. | 22:57 |
| vishy | pgregory: interesting. what can you access? Do you ssh to the url? | 22:58 |
| alekibango | vishy: ajax terminal session? | 22:59 |
| alekibango | console using ajax is fun | 22:59 |
| vishy | pgregory: yes we have ajax console suport as well, but I find it less useful than vnc console as it requires special setup in the guest | 23:01 |
| pgregory | vishy: yes, you get a url which seems to have the ip encoded in it, so I presume they have some script hackery that extracts the ip and redirects you to the right instance. | 23:02 |
| vishy | so i guess they have some sort of ssh proxy? That could be useful, but we don't have anything like that | 23:03 |
| *** jkoelker has quit IRC | 23:03 | |
| vishy | if your cloud is private you can give out rfc1918 ips and add a route to reach them | 23:03 |
| pgregory | vishy: so you might do ssh -i myec2_key.priv root@ec2-192-168-1-1.eu-west-1.compute.amazonaws.com | 23:04 |
| pgregory | vishy: where can I find information about the ajax console? | 23:05 |
| vishy | pgregory: the bottom of this doc, shows you how you can add a route to get to private ips: http://cloudbuilders.github.com/deploy.sh/multi-node-nova-installation-using-vagrant-and-chef.html | 23:05 |
| *** medberry is now known as med_out | 23:05 | |
| vishy | although keep in mind that you will need to use euca-authorize to add in rules if you want to ping / ssh | 23:06 |
| *** agarwalla has joined #openstack | 23:07 | |
| alekibango | vishy: i would be very thankfull if you could help a bit here: http://etherpad.openstack.org/glance-walkthrough (many people stumble on this, we need to it in docs) | 23:07 |
| vishy | pgregory: not sure that there is good docs about the ajax-console-proxy | 23:07 |
| alekibango | vishy: :) you need to be cloned | 23:08 |
| pgregory | vishy: ok, I'll stick with VNC for now, I don't suspect we'll need to access the instances externally for much anyway. | 23:08 |
| *** amccabe has quit IRC | 23:08 | |
| *** jonkelly has quit IRC | 23:10 | |
| vishy | alekibango: believe me I know | 23:12 |
| vishy | :) | 23:12 |
| alekibango | :)))) | 23:12 |
| alekibango | using swift with s3 could work as an alternative for now? | 23:13 |
| alekibango | instead of nova-objectstore..? | 23:13 |
| *** joearnold has quit IRC | 23:14 | |
| *** joearnold has joined #openstack | 23:14 | |
| alekibango | vishy: imho using agile development could help a bit | 23:15 |
| alekibango | please think about it for nova | 23:15 |
| *** MotoMilind has joined #openstack | 23:17 | |
| vishy | alekibango: we have been agile from the beginning, but agile * 10 teams and 70+ devs is a bit hard | 23:18 |
| alekibango | i know, maybe even impossible | 23:18 |
| vishy | alekibango: I don't know if you saw that we are doing 1 month milestones, but this is to help make us a bit more agile | 23:19 |
| alekibango | i have seen it | 23:19 |
| vishy | so that teams can line up milestones with sprints | 23:19 |
| alekibango | vishy: sounds nice | 23:19 |
| alekibango | i would like it to be even shorter ... :) | 23:19 |
| alekibango | but i dont see it from your perspective | 23:20 |
| vishy | alekibango: when our automated testing is up to snuff, we'll revisit shorter milestones :) | 23:21 |
| alekibango | vishy: thats what i wanted to hear, thanks a lot | 23:21 |
| alekibango | :) | 23:21 |
| vishy | alekibango: in theory yes, although no one i know has actually tried it, it may require a few hacks in the code for auth/etc. | 23:21 |
| *** matt25 has quit IRC | 23:21 | |
| alekibango | ??? in theory?? what? | 23:22 |
| vishy | alekibango: in theory, actual s3 should work too | 23:22 |
| alekibango | aha | 23:22 |
| *** GasbaKid has quit IRC | 23:22 | |
| vishy | alekibango: I don't know how the s3 front end for swift handles auth | 23:22 |
| alekibango | i dont know this too lol | 23:22 |
| vishy | alekibango: i do know that the nova code that is pulling data out of s3 doesn't pass any auth creds | 23:22 |
| *** GasbaKid has joined #openstack | 23:23 | |
| vishy | alekibango: so I think that would need to be added | 23:23 |
| alekibango | hmm so best what i can do is make glance api public? | 23:23 |
| alekibango | if i want to run public demo cloud | 23:23 |
| alekibango | i am afraid of getting unwanted 'images' | 23:24 |
| alekibango | 2nd best tip: let them mail it to someone who will put it into glance for them (lol, mail) | 23:24 |
| *** thingee has quit IRC | 23:25 | |
| alekibango | hmm but having popular warez/porn site would make good demo for the cloud hehe | 23:26 |
| alekibango | i mean, guys, we need auth asap! | 23:26 |
| alekibango | for glance | 23:26 |
| vishy | alekibango: agreed... | 23:31 |
| vishy | alekibango: i added a bunch of comments to the etherpad | 23:31 |
| *** MarkAtwood has quit IRC | 23:33 | |
| alekibango | vishy: i see them, thanks a ton ! | 23:34 |
| alekibango | how to use python-novaclient for image upload <--- that would be very nice to have | 23:34 |
| alekibango | i think adding few commands into it would be best idea | 23:35 |
| alekibango | anyone interested in glance, please join editing this little document http://etherpad.openstack.org/glance-walkthrough | 23:35 |
| *** namaqua has joined #openstack | 23:36 | |
| *** namaqua has quit IRC | 23:42 | |
| *** dendrobates is now known as dendro-afk | 23:44 | |
| *** jero has joined #openstack | 23:46 | |
| *** jero has left #openstack | 23:46 | |
| alekibango | btw xnbd looks like interestng project https://bitbucket.org/hirofuchi/xnbd/wiki/Home | 23:47 |
| *** rchavik has joined #openstack | 23:47 | |
| *** _vinay has quit IRC | 23:48 | |
| *** dendro-afk is now known as dendrobates | 23:49 | |
| *** londo_ has quit IRC | 23:49 | |
| *** londo_ has joined #openstack | 23:50 | |
| *** icarus901 has joined #openstack | 23:50 | |
| vishy | agreed, interesting | 23:51 |
| alekibango | still not 'stable' but i would like to try this one too | 23:52 |
| vishy | alekibango: yes, seems like it could allow an alternative live-migration scheme | 23:53 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!