| *** kevnfx has quit IRC | 00:00 | |
| nelson__ | question about /etc/swift/swift.conf#swift_hash_path_suffix | 00:03 |
|---|---|---|
| nelson__ | why does it have to be a "random string of text"? | 00:04 |
| nelson__ | is it for security reasons? Or just to ensure that short object names get well hashed? | 00:04 |
| *** anotherjesse has joined #openstack | 00:04 | |
| *** dubsquared has quit IRC | 00:06 | |
| *** schisamo has quit IRC | 00:13 | |
| *** dfg_ has quit IRC | 00:21 | |
| *** rnirmal has joined #openstack | 00:21 | |
| *** allsystemsarego has quit IRC | 00:24 | |
| *** anotherjesse has quit IRC | 00:24 | |
| *** Ryan_Lane has quit IRC | 00:24 | |
| notmyname | nelson__: heh. funny story about that | 00:34 |
| notmyname | ok, maybe not really funny | 00:34 |
| nelson__ | haha, I'm laughing already and we haven't even gotten to the punch line. | 00:34 |
| notmyname | we has some people internally concerned with customers generating hash collisions. this was before we even talked about open-source and it would require a preimage attack (which has never been done, even with md5) | 00:35 |
| notmyname | so we offered persuasive arguments on why we didn't need it | 00:36 |
| notmyname | and then we added it because it was easier and doesn't really hurt anything | 00:36 |
| notmyname | so it's prepended to every request before the request is hashed to do the ring lookup | 00:36 |
| notmyname | so keep it secret and you can't ever change it (or every single object in the ring get's moved with replication) | 00:37 |
| notmyname | so it's kinda for security reasons, but only for vulnerabilities that have never happened. it's not a bad thing, but.... | 00:39 |
| *** ccustine has quit IRC | 00:39 | |
| *** dendrobates is now known as dendro-afk | 00:40 | |
| vishy | considering using lockfile | 00:41 |
| notmyname | the best argument for adding it was given by redbo. if one could generate many object names that hashed near to one another, and attacker could fill up one partition very quickly. by adding the prefix, you are protected because the attacker can't find the proper hash | 00:41 |
| *** rlucio has quit IRC | 00:45 | |
| *** rlucio has joined #openstack | 00:45 | |
| nelson__ | so it shouldn't be *too* awful long. Like maybe the first part of a uuid? | 00:47 |
| *** dendro-afk is now known as dendrobates | 00:48 | |
| nelson__ | annoyingly, Ubuntu doesn't install `uuid` by default. | 00:49 |
| notmyname | ya, that would be good. or even a whole uuid in hex. don't make it huge. I'm not a security expert, but there may be something about being able to precompute the first blocks of a hash (not sure about md5, I know sha1 works that way), so it may be good to keep it less than the md5 block size | 00:49 |
| nelson__ | dd if=/dev/random count=4 | od -x :-) | 00:51 |
| notmyname | actually, thinking about it more, I think that since it's a prefix, that's not a problem (the block size thing). just some random characters or the name of your first pet or the ssn of your ceo or something ;-) | 00:51 |
| notmyname | ya, that would be perfect :-) | 00:51 |
| nelson__ | oh fuck, why can't programs just shut up and do their work?? | 00:52 |
| notmyname | gotta go | 00:53 |
| nelson__ | this is better: od -t u4 -N 4 -A n </dev/random | 00:55 |
| nelson__ | Okay, I think I have that section of the docs in good shape. Thanks, notmyname. | 01:00 |
| *** dirakx has quit IRC | 01:02 | |
| *** olivier_ has quit IRC | 01:05 | |
| *** olivier_ has joined #openstack | 01:05 | |
| *** sophiap_ has joined #openstack | 01:17 | |
| *** Ryan_Lane has joined #openstack | 01:21 | |
| *** sophiap has quit IRC | 01:22 | |
| *** sophiap_ is now known as sophiap | 01:22 | |
| *** jdurgin has quit IRC | 01:23 | |
| *** kevnfx has joined #openstack | 01:24 | |
| *** PiotrSikora has quit IRC | 01:31 | |
| *** MikeZ1 has quit IRC | 01:33 | |
| *** maplebed has quit IRC | 01:33 | |
| *** skrusty has quit IRC | 01:35 | |
| *** schisamo has joined #openstack | 01:35 | |
| *** Jordandev has quit IRC | 01:41 | |
| *** MarkAtwood has quit IRC | 01:41 | |
| *** MikeZ has joined #openstack | 01:44 | |
| *** skrusty has joined #openstack | 01:47 | |
| *** joearnold has quit IRC | 01:49 | |
| *** PiotrSikora has joined #openstack | 02:04 | |
| *** ccustine has joined #openstack | 02:04 | |
| *** MarkAtwood has joined #openstack | 02:05 | |
| *** Ryan_Lane has quit IRC | 02:09 | |
| *** sandywalsh has quit IRC | 02:16 | |
| *** Jordandev has joined #openstack | 02:18 | |
| *** masumotok_ has quit IRC | 02:24 | |
| *** sandywalsh has joined #openstack | 02:29 | |
| *** MikeZ has quit IRC | 02:32 | |
| *** Cybo has joined #openstack | 02:42 | |
| *** MarkAt2od has joined #openstack | 02:44 | |
| *** ccustine has quit IRC | 02:56 | |
| *** daleolds has quit IRC | 02:57 | |
| *** trin_cz has quit IRC | 03:19 | |
| *** jbryce has quit IRC | 03:19 | |
| Jordandev | i am using RabbitMQ. Please, any advice about the best publisher/consumer: pika (I like it), carrot (others)...? | 03:25 |
| *** Ryan_Lane has joined #openstack | 03:27 | |
| *** sophiap has quit IRC | 03:28 | |
| Jordandev | what you think about kombu for AMQP? | 03:36 |
| Jordandev | any d | 03:36 |
| Jordandev | any feedback ? :) | 03:36 |
| sandywalsh | Jordandev, we're using carrot in nova currently | 03:40 |
| sandywalsh | but jk0 didn't exactly rave about it :) | 03:40 |
| Jordandev | from carrot site: "discontinued; Use Kombu instead" | 03:41 |
| *** aliguori has quit IRC | 03:46 | |
| Jordandev | sandywalsh, what is the main advantage of carrot compared to others like Pika? | 03:47 |
| winston-d | My swift-bench test results: http://paste.openstack.org/show/395/ | 03:47 |
| winston-d | No much improvement compared to yesterday's. | 03:48 |
| sandywalsh | Jordandev, I personally haven't compared them. Carrot is a pretty thin wrapper. | 03:49 |
| sandywalsh | Jordandev, that's interesting about the carrot note. I'll pass that along. | 03:49 |
| Jordandev | thanks sandywalsh :) | 03:50 |
| sandywalsh | Jordandev, np, wish I could be of more help | 03:50 |
| Jordandev | :) | 03:51 |
| *** mdomsch has joined #openstack | 03:56 | |
| *** Ryan_Lane has quit IRC | 04:03 | |
| *** Jordandev has quit IRC | 04:13 | |
| *** shermanboyd has quit IRC | 04:28 | |
| *** shermanboyd has joined #openstack | 04:28 | |
| *** shermanboyd has quit IRC | 04:30 | |
| *** dirakx has joined #openstack | 04:40 | |
| notmyname | winston-d: I'm running swift all-in-one on a 1024 slice. no tuning. xfs mounts are done on a loopback. basic config for dev only. absolutely zero effort spent on performance. http://paste.openstack.org/show/396/ | 04:41 |
| winston-d | notmyname: what do you mean by 1024 slice? | 04:47 |
| notmyname | slicehost 1GB slice | 04:48 |
| winston-d | my swift setup is: 1 proxy, 1 auth (same machine); 6 account/container/object | 04:48 |
| winston-d | 1GbE network connection, 3 replicate | 04:49 |
| winston-d | according to creiht, my result is "incredibly slow" | 04:50 |
| notmyname | I just wanted to give you a baseline of numbers to expect. more machines should == more performance. we can get about 70 req/s on a VM on a laptop. you see my numbers for a dedicated VPS | 04:50 |
| notmyname | yes. "incredibly slow" is a nice way to say it ;-) | 04:50 |
| winston-d | any idea what is wrong with my system? or what should I look at first? | 04:52 |
| notmyname | that's something that creiht can probably help with more than me (he's got a lot more experience walking people through setting up swift) | 04:52 |
| notmyname | I would look at your worker count (I'm using the default | 04:53 |
| notmyname | do you know where you are slowing down? network? cpu? | 04:53 |
| winston-d | you are right. yesterday creiht suggest the same thing. I did modified worker count based on his recommendation. | 04:53 |
| notmyname | actually, I remember him saying something about that. which is why I think it's a good place to start :-) | 04:54 |
| notmyname | bur, really, it's hard to tune something when you don't know where it's slow. perhaps you could run a longer swift-bench run and do some profiling of the machines? | 04:55 |
| winston-d | good idea. | 04:55 |
| notmyname | since you are doing a multi node setup, I would look at the network utilization. are you routing through some switch that is limiting you? or maybe you are maxing one core on a node and need to increase workers | 04:56 |
| winston-d | no, my nodes are connected to the same GbE switch | 04:57 |
| winston-d | and I've already increase workers to 1x core count | 04:57 |
| notmyname | creiht wrote swift-bench to give everyone a common set of tests so we could compare apples to apples. but you could do something like curl commands in a loop | 04:58 |
| *** krish has joined #openstack | 04:58 | |
| *** Lcfseth has joined #openstack | 04:59 | |
| *** EdwinGrubbs has quit IRC | 04:59 | |
| winston-d | switch-bench is suppose to delete everything it creates after test (in default setting), right? | 04:59 |
| notmyname | that won't give you good numbers to compare, but it may highlight a bottleneck | 04:59 |
| notmyname | yes, I think so | 04:59 |
| *** krish has left #openstack | 05:01 | |
| notmyname | if you look in the logs on the storage nodes, you will see what requests to them look like. you could also do requests yourself from the proxy node to the storage nodes to check the connectivity there | 05:01 |
| winston-d | alrite | 05:02 |
| notmyname | are you running memcache? | 05:02 |
| notmyname | could it be something as simple as the auth token not being cached (or account/container existence) | 05:02 |
| winston-d | yes | 05:04 |
| notmyname | hmm..that would have been simple to fix :-) | 05:04 |
| winston-d | let me restart 'memcached' service and try again | 05:04 |
| winston-d | I followed the instructions here to setup my testbed: http://swift.openstack.org/howto_installmultinode.html | 05:04 |
| winston-d | didn't find any configuration for 'memcached' | 05:05 |
| winston-d | is there any? | 05:05 |
| notmyname | [filter:cache] section of the proxy server conf | 05:06 |
| winston-d | OK. that's what i did | 05:06 |
| winston-d | another question, when swift-bench is done, i found there's some directories and files let on XFS disks. is that normal? | 05:07 |
| notmyname | mayber? | 05:07 |
| notmyname | maybe? | 05:07 |
| notmyname | I'm not sure. actually, tonight is the first time I've run swift-bench | 05:08 |
| winston-d | even when i try st delete --all, there's still something | 05:08 |
| winston-d | although 'st list' didn't show any | 05:08 |
| notmyname | deleting objects actually only creates tombstone files. the background processes take care of reclaiming the space on disk | 05:09 |
| notmyname | if you look in the proxy log, the last field is the time the request took | 05:11 |
| winston-d | OK. let me check | 05:11 |
| notmyname | you could upload a 1K file and then fetch it and look at the times in the logs | 05:11 |
| notmyname | things to look for will be times very close to 3 seconds or 10 seconds | 05:12 |
| notmyname | both are used for various timeouts in the system | 05:12 |
| *** EdwinGrubbs has joined #openstack | 05:12 | |
| *** EdwinGrubbs has joined #openstack | 05:12 | |
| notmyname | actually, you should probably do zero byte files rather than 1K files. either way, adjust the math and you can get throughput rates | 05:13 |
| winston-d | I see. | 05:13 |
| notmyname | zero byte objects should highlight timeout errors. obviously it's hard to see timeouts if some requests actually take 3 or 10 seconds because of the amount of data transferred | 05:14 |
| *** blpiatt has joined #openstack | 05:15 | |
| *** Cybo has quit IRC | 05:16 | |
| *** shermanboyd has joined #openstack | 05:19 | |
| winston-d | notmyname: how can i configure swift component to use its own log file instead of syslog? | 05:23 |
| winston-d | proxy server log is now messing-up with rest of system | 05:23 |
| notmyname | check out http://swift.openstack.org/overview_stats.html#running-the-stats-system-on-saio | 05:24 |
| notmyname | steps 2-5 set up the proxy logs to go to a separate file | 05:24 |
| *** rnirmal has quit IRC | 05:25 | |
| notmyname | oh, and step 7 too | 05:25 |
| *** michaeldreamhost has quit IRC | 05:27 | |
| winston-d | new result after restarted memcached service: http://paste.openstack.org/show/397/ | 05:27 |
| winston-d | PUT still sucks, but GET and DEL are much better | 05:27 |
| winston-d | still much lower than your baseline, though | 05:28 |
| *** guynaor has joined #openstack | 05:31 | |
| *** guynaor has left #openstack | 05:31 | |
| *** kashyapc has joined #openstack | 05:32 | |
| *** Ryan_Lane has joined #openstack | 05:33 | |
| *** michaeldreamhost has joined #openstack | 05:33 | |
| *** hadrian has quit IRC | 05:33 | |
| *** Ryan_Lane is now known as Ryan_Lane|away | 05:39 | |
| *** shermanboyd has quit IRC | 05:39 | |
| *** Lcfseth has left #openstack | 05:41 | |
| *** jbryce has joined #openstack | 05:47 | |
| *** f4m8_ is now known as f4m8 | 06:07 | |
| *** ramkrsna has joined #openstack | 06:12 | |
| *** ramkrsna has quit IRC | 06:17 | |
| *** mdomsch has quit IRC | 06:22 | |
| *** ramkrsna has joined #openstack | 06:29 | |
| *** MarkAt2od has left #openstack | 06:41 | |
| *** MarkAtwood has quit IRC | 06:41 | |
| *** miclorb has quit IRC | 06:53 | |
| *** ramkrsna has quit IRC | 06:54 | |
| *** ramkrsna has joined #openstack | 06:54 | |
| *** kevnfx has quit IRC | 07:00 | |
| *** larstobi has joined #openstack | 07:02 | |
| *** larstobi has quit IRC | 07:19 | |
| *** littleidea has quit IRC | 07:33 | |
| *** zaitcev has quit IRC | 07:35 | |
| *** guigui1 has joined #openstack | 07:35 | |
| *** rcc has joined #openstack | 07:42 | |
| ttx | <rlucio> looks like there is some web url you are supposed to use to create an account | 07:44 |
| ttx | uh? | 07:44 |
| ttx | You're confusing UEC (based on eucalyptus) and the UEC images. the web UI is for eucalyptus credentials | 07:47 |
| ttx | rlucio: If you create a key with euca-add-keypair and pass it as "-k" to the euca-run-instances, it will get imported in the ubuntu user authorized_keys, to allow for SSH login | 07:47 |
| ttx | The password on the serial console is only if you use "uncloudinit" (to run outside a cloud) | 07:49 |
| ttx | dubsquared: ^ | 07:51 |
| *** brd_from_italy has joined #openstack | 07:58 | |
| *** blpiatt has quit IRC | 08:01 | |
| *** rcc has quit IRC | 08:02 | |
| ttx | For those who missed the meeting yesrterday (spectorclan?), see the logs at http://wiki.openstack.org/MeetingLogs | 08:05 |
| *** fabiand has joined #openstack | 08:11 | |
| *** dirakx has quit IRC | 08:30 | |
| *** calavera has joined #openstack | 08:38 | |
| *** miclorb_ has joined #openstack | 08:39 | |
| *** littleidea has joined #openstack | 08:44 | |
| *** irahgel has joined #openstack | 08:59 | |
| *** Firaga has joined #openstack | 09:13 | |
| Firaga | hello | 09:14 |
| Firaga | anyone whose successfully created an image for openstack compute, like CentOS/Windows image for openstack? | 09:14 |
| *** allsystemsarego has joined #openstack | 09:15 | |
| *** allsystemsarego has joined #openstack | 09:15 | |
| Firaga | anyone whose successfully created an image for openstack compute, like CentOS/Windows image for openstack? | 09:22 |
| soren | xtoddx: Did you really intend to remove the "audit=INFO" element from the default_log_levels list? | 09:25 |
| *** arthurc has joined #openstack | 09:27 | |
| fraggeln | anyone have a clue on where to look when I get this error: HTTP/1.1 413 Request Entity Too Large | 09:33 |
| fraggeln | (its a 7gb file that im trying to upload) | 09:34 |
| fabiand | fraggeln: in swift? | 09:37 |
| fraggeln | fabiand: in st | 09:37 |
| fraggeln | im using st to upload a file to my swiftsetup | 09:38 |
| fabiand | fraggeln: You might need to segment your upload, as (single) uploads are limited to 5GB | 09:38 |
| fraggeln | is there any way to change that limit? | 09:38 |
| fraggeln | since we are planning to use switf as a backend for our backupsoftware, and I think it creates larger files thaen 5gb. | 09:39 |
| fabiand | segmenting is a good thing (tm), it is chossen to be 5gb to keep the nodes in a better balance .. | 09:39 |
| fraggeln | okay. | 09:39 |
| fabiand | fraggeln: segmenting means: you split your file into smaller ones ... you can retrieve the file afterwards in one go .. | 09:39 |
| fabiand | fraggeln: there you go http://swift.openstack.org/overview_large_objects.html?highlight=segment | 09:40 |
| fraggeln | fabiand: thanks | 09:40 |
| fabiand | np | 09:40 |
| fraggeln | when you talked about segmenting, the first thing that crossed my mind was using tar/rar or something to split the file at the client itself | 09:42 |
| Firaga | anyone whose successfully created an image for openstack compute, like CentOS/Windows image for openstack? | 09:42 |
| soren | Firaga: Not CentOS or Windows, no. | 09:43 |
| soren | Firaga: Ubuntu, yes. | 09:43 |
| Firaga | ubuntu image only soren? | 09:43 |
| soren | I've only built Ubuntu images, yes. | 09:44 |
| soren | That's about all I really care about. | 09:44 |
| Firaga | need directions on how to create from a Xen CentOS 5.5 VM pls | 09:45 |
| Firaga | I only got an .img file. where can i get the kernel & ramdisk? | 09:46 |
| soren | How do you usually run it? | 09:47 |
| Firaga | also, I just updated my nova and now I can't run my instances. "OperationalError: no such column: | 09:47 |
| soren | Daviey: ^^ How's that db schema migration magic coming along? :) | 09:47 |
| Firaga | I create a VM using Xen/virt-lib | 09:48 |
| soren | I'm not sure what you're saing. | 09:48 |
| soren | saying, even. | 09:49 |
| Firaga | ok, where are the docs to create & publish a Xen CentOS 5.5 VM that can be loaded into OpenStack? | 09:51 |
| Firaga | I need to create my own CentOS Xen image | 09:51 |
| Firaga | ttylinux is all i can run | 09:52 |
| *** MarkAtwood has joined #openstack | 09:52 | |
| *** littleidea has quit IRC | 09:53 | |
| Firaga | do openstack have any documentation on how to create a CentOS.img? | 09:55 |
| Firaga | do openstack have any documentation on how to create a CentOS.img? | 09:58 |
| fabiand | Firaga: repeating your question does not lead to quicker responses. | 10:00 |
| *** trin_cz has joined #openstack | 10:00 | |
| Firaga | sorry about that, i accidentally press enter twice, pc is lagging... hehehe | 10:01 |
| *** Firaga has quit IRC | 10:06 | |
| *** befreax has joined #openstack | 10:06 | |
| *** JohnFury has joined #openstack | 10:16 | |
| JohnFury | hello guys\ | 10:16 |
| JohnFury | i'm looking for any documentation for openstack running a Windows or any other Linux OS? | 10:17 |
| soren | There are Ubuntu images available. | 10:18 |
| * soren has an odd sense of deja vu | 10:19 | |
| JohnFury | alright, where can i get it? i appreciate your answers man... | 10:19 |
| soren | ttx: I'm sure you have the link handy... | 10:19 |
| JohnFury | great... | 10:20 |
| JohnFury | how about any procedure on creating an image for openstack? do we got any documentation about that? | 10:20 |
| ttx | JohnFury: http://wiki.openstack.org/RunningNova | 10:21 |
| ttx | JohnFury: or http://fnords.wordpress.com/2010/12/02/bleeding-edge-openstack-nova-on-maverick/ | 10:21 |
| ttx | creating images: it's a bit out of our scope (the OS vendor should provide that) but I agree we could write some docs on how to do it :) | 10:22 |
| JohnFury | just need HOWTO guide on creating kernel & ramdisk | 10:24 |
| JohnFury | i'm downloading them now... | 10:25 |
| *** JohnFury has quit IRC | 10:30 | |
| *** JohnFury has joined #openstack | 10:32 | |
| JohnFury | got disconnected, sorry about that... | 10:33 |
| JohnFury | yeah, creating a documentatio for a procedure on how to create a image for openstack is a very great idea... | 10:33 |
| JohnFury | ttx or Thierry? how did you guys created the ubuntu image. | 10:36 |
| ttx | JohnFury: I didn't. The Ubuntu Server team did. | 10:36 |
| ttx | JohnFury: we/they used vmbuilder, but that should be pretty Ubuntu specific | 10:37 |
| JohnFury | i see... vmbuilder, i will check on that, great... i hope you guys can post a documentation on how you actually did it... | 10:39 |
| JohnFury | anyway, ttx got any twitter account, so i can be your follower? | 10:40 |
| ttx | JohnFury: @tcarrez | 10:43 |
| ttx | JohnFury: you can talk to smoser, he owns that part on Ubuntu's side | 10:43 |
| *** ArdRigh has quit IRC | 10:47 | |
| JohnFury | alright, great... i will check with him, i hope his online... | 10:47 |
| *** icarus901 has quit IRC | 10:50 | |
| *** trin_cz has quit IRC | 10:51 | |
| *** MarkAtwood has quit IRC | 10:58 | |
| *** icarus901 has joined #openstack | 11:05 | |
| soren | JohnFury: It's 5 in the morning where he is, so unlikely. | 11:11 |
| fraggeln | How do people normaly connect to their "cloud"? I doubt people are using the st-commant in applications and stuff. | 11:12 |
| fraggeln | is there anyway to mount it in ie linux/solaris? | 11:12 |
| fraggeln | or am I thinking all wrong here? | 11:13 |
| JohnFury | +sorren:: yeah that's why i got to finish some cooking and eating and dressing... while waiting for our friend to be awake... got to do dishes now openstack people... see yah in a minute. | 11:13 |
| *** dagger has quit IRC | 11:26 | |
| *** dagger has joined #openstack | 11:26 | |
| sandywalsh | ttx, nice openstack announcements today! | 11:29 |
| *** colinnich has joined #openstack | 11:30 | |
| soren | sandywalsh: What, where? | 11:36 |
| befreax | soren: http://twitter.com/tcarrez/status/22613777916628993 | 11:37 |
| *** colinnich has quit IRC | 11:37 | |
| *** colinnich has joined #openstack | 11:38 | |
| *** colinnich has quit IRC | 11:39 | |
| *** colinnich has joined #openstack | 11:40 | |
| soren | befreax: ta | 11:41 |
| *** miclorb_ has quit IRC | 11:47 | |
| *** colinnich has quit IRC | 11:49 | |
| *** colinnich has joined #openstack | 11:51 | |
| *** colinnich has joined #openstack | 11:52 | |
| *** kashyapc has quit IRC | 11:54 | |
| *** colinnich has quit IRC | 11:54 | |
| *** colinnich has joined #openstack | 11:55 | |
| *** colinnich has joined #openstack | 11:59 | |
| *** trin_cz has joined #openstack | 11:59 | |
| *** trin_cz has quit IRC | 12:02 | |
| *** trin_cz has joined #openstack | 12:06 | |
| *** ctennis has quit IRC | 12:10 | |
| *** skrusty has quit IRC | 12:21 | |
| *** skrusty has joined #openstack | 12:36 | |
| *** scaraffe has joined #openstack | 12:40 | |
| *** ctennis has joined #openstack | 12:57 | |
| *** colinnich_ has joined #openstack | 13:14 | |
| *** colinnich_ has quit IRC | 13:17 | |
| notmyname | fraggeln: swift isn't mountable as a filesystem (cause it's not a filesystem). the way to interact with it is to use HTTP, a language binding, or a third-party tool like Cyberduck and others | 13:23 |
| *** crago has joined #openstack | 13:24 | |
| *** ramkrsna has quit IRC | 13:28 | |
| *** scaraffe has quit IRC | 13:35 | |
| *** jbryce has quit IRC | 13:42 | |
| *** crago has left #openstack | 13:44 | |
| *** hadrian has joined #openstack | 13:49 | |
| *** rogue780 has joined #openstack | 13:53 | |
| rogue780 | howdy. anybody here today? | 13:56 |
| soren | notmyname: Do we know a Ryan Brown? | 13:58 |
| *** aliguori has joined #openstack | 13:59 | |
| soren | notmyname: Someone by that name who according to github works for Rackspace has written a gvfs backend for swift. | 13:59 |
| soren | https://github.com/rbr/gvfs | 14:00 |
| fabiand | wow. cool. | 14:01 |
| *** zul has quit IRC | 14:09 | |
| *** jbryce has joined #openstack | 14:09 | |
| *** zul has joined #openstack | 14:10 | |
| sandywalsh | soren, nice find! | 14:10 |
| *** westmaas has joined #openstack | 14:11 | |
| smoser | JohnFury, kernel and ramdisk should not really need anything specific. Ours are just a kernel and ramdisk like you might have on your desktop or any server installation. | 14:13 |
| smoser | they use our '-virtual' kernel. I'm not aware of any kernel specific things we've done that weren't a result of EC2 (xen) | 14:13 |
| *** JohnFury has quit IRC | 14:17 | |
| ttx | smoser: oh no, you killed him. | 14:20 |
| *** kashyapc has joined #openstack | 14:29 | |
| *** allsystemsarego_ has joined #openstack | 14:30 | |
| *** allsystemsarego has quit IRC | 14:31 | |
| *** littleidea has joined #openstack | 14:32 | |
| zul | soren: http://pastebin.ubuntu.com/550657/ im getting this error when using the upstart job in the packaging branch | 14:37 |
| ttx | twisted ? that sounds so last decade | 14:39 |
| soren | zul: nova-compute hasn't used twised since before christmas. | 14:39 |
| soren | twisted, even. | 14:39 |
| soren | zul: That's your first problem. | 14:39 |
| zul | soren: *sigh* | 14:40 |
| soren | zul: Your second problem is that you've been a bad boy and run nova-compute as root. | 14:40 |
| zul | no im running the upstart job | 14:40 |
| soren | Yes, now. | 14:40 |
| * ttx spanks zul | 14:40 | |
| soren | "have been" | 14:40 |
| soren | Oh, wait. | 14:40 |
| soren | /var/run/nova-compute.pid.. Why would it try to put it in /var/run? | 14:40 |
| * soren has to go pick up daughter at day care... Will be back this evening at some point. | 14:41 | |
| soren | toodles. | 14:41 |
| fabiand | fraggeln: you can mount swift using that gvfs client noted above. | 14:44 |
| *** spectorclan has joined #openstack | 14:45 | |
| *** dirakx has joined #openstack | 14:46 | |
| *** fabiand_ has joined #openstack | 14:48 | |
| *** f4m8 is now known as f4m8_ | 14:48 | |
| *** ppetraki has joined #openstack | 14:49 | |
| uvirtbot | New bug: #697689 in openstack-ci "swift-debian broken since build #104" [Undecided,New] https://launchpad.net/bugs/697689 | 14:51 |
| *** bingoooo has joined #openstack | 14:51 | |
| notmyname | soren: nope. haven't met him. looks like he's in the austin office. good find, though | 14:53 |
| *** fabiand has left #openstack | 14:54 | |
| *** calavera has quit IRC | 15:00 | |
| *** b_erb has joined #openstack | 15:01 | |
| *** nelson__ has quit IRC | 15:03 | |
| *** nelson__ has joined #openstack | 15:03 | |
| *** mdomsch has joined #openstack | 15:04 | |
| *** gondoi has joined #openstack | 15:06 | |
| *** kevnfx has joined #openstack | 15:07 | |
| *** Podilarius has joined #openstack | 15:08 | |
| *** kevnfx has quit IRC | 15:13 | |
| bingoooo | hi | 15:15 |
| notmyname | bingoooo: morning | 15:18 |
| *** rnirmal has joined #openstack | 15:22 | |
| *** mdomsch has quit IRC | 15:26 | |
| *** guigui1 has quit IRC | 15:28 | |
| *** dfg_ has joined #openstack | 15:28 | |
| *** dragondm has joined #openstack | 15:42 | |
| *** Jbain has joined #openstack | 15:55 | |
| *** rogue780 has quit IRC | 15:56 | |
| *** blpiatt has joined #openstack | 15:58 | |
| *** dubsquared has joined #openstack | 16:03 | |
| *** b_erb has left #openstack | 16:06 | |
| *** bingoooo has quit IRC | 16:10 | |
| *** johnpur has joined #openstack | 16:11 | |
| *** ChanServ sets mode: +v johnpur | 16:11 | |
| vishy | Firaga, JohnFury: I successfully created a centos image for openstack. It required a few tweaks to kernel modules and such | 16:12 |
| vishy | but it should be easier know that raw-disk-images is in, you don't have to create a separate kernel and ramdisk | 16:13 |
| dubsquared | vishy: did this get published anywhere? | 16:13 |
| vishy | dubsquared: I sent it to JordanRinke, he was going to clean it up a bit and put it up somewhere | 16:14 |
| vishy | we really need a place to shove testing images | 16:14 |
| dubsquared | alright, ill have to bug him some more…i got focused on xenserver stuff so i didn't have time to ask him about it | 16:14 |
| dubsquared | yeah, that would be great... | 16:15 |
| vishy | dubsquared: I can zip it up and put it somewhere if someone knows where I can stick a few gig file | 16:15 |
| dubsquared | i have a hack site im using that has a mess of images | 16:15 |
| dubsquared | cool, ill drop you an account on my cloudfiles | 16:15 |
| dubsquared | unless i see rinke first... | 16:16 |
| * dubsquared peers over at empty desk | 16:16 | |
| *** Ryan_Lane|away is now known as Ryan_Lane | 16:16 | |
| *** kevnfx has joined #openstack | 16:18 | |
| creiht | winston-d: What timezone are you in? | 16:18 |
| fabiand_ | vishy: rawdisk images are supported now? | 16:24 |
| *** Podilarius has quit IRC | 16:24 | |
| fabiand_ | sorry - you just said it :) | 16:24 |
| vishy | fabiand_: in trunk yes | 16:24 |
| fabiand_ | very ince, | 16:24 |
| fabiand_ | nice | 16:24 |
| creiht | winston-d: And I can walk you through debugging your setup | 16:24 |
| creiht | winston-d: it is usually something simple | 16:25 |
| *** dirakx has quit IRC | 16:26 | |
| creiht | winston-d: and as a reference point, with decent hardware, you should be able to easily get over 1k puts/sec through one proxy | 16:27 |
| creiht | with an upper bound of close to 2k with everything well tuned | 16:28 |
| *** kashyapc has quit IRC | 16:34 | |
| *** dirakx has joined #openstack | 16:35 | |
| jk0 | would anyone mind reviewing https://code.launchpad.net/~jk0/nova/lp697019/+merge/45196 ? it's short & sweet | 16:39 |
| *** Podilarius has joined #openstack | 16:39 | |
| *** opengeard has quit IRC | 16:44 | |
| letterj_ | mtaylor: You around? | 16:45 |
| *** ChanServ sets mode: +v letterj_ | 16:49 | |
| creiht | littleidea: notmyname said that you had some questions? | 16:50 |
| *** singh has joined #openstack | 16:50 | |
| littleidea | creiht: this is often the case | 16:50 |
| creiht | lol | 16:50 |
| creiht | :) | 16:50 |
| littleidea | creiht: I was asking about the design of swift | 16:51 |
| creiht | shoot | 16:51 |
| *** kashyapc has joined #openstack | 16:52 | |
| *** cynb has left #openstack | 16:52 | |
| *** daleolds has joined #openstack | 16:52 | |
| littleidea | If I understand correctly, everything goes through the proxy server both ways and I was specifically wondering if you considered the object servers returning requests directly | 16:52 |
| creiht | littleidea: that question comes up quite often :) | 16:52 |
| creiht | we had considered that in the initial design | 16:53 |
| notmyname | littleidea: sorry I couldn't answer yesterday. meeting and other commitments (I'm sure you can't relate ;-) | 16:53 |
| littleidea | notmyname: no sorrow necessary :) | 16:54 |
| creiht | One thing is that going through the proxy was the simplest, and worked "good enough" :) | 16:54 |
| creiht | The other thing that we liked about it is that we only have to have public addresses on the proxies | 16:54 |
| Ryan_Lane | when allocatng elastic IP addresses, is the IP address allocated to a user, or a project? | 16:55 |
| Ryan_Lane | and are these IPs the "floating" IPs in nova-manage? | 16:56 |
| creiht | Those are the simple answers, I believe there were more technical reasons why we didn't persue the idea, but I can't remember off the top of my head | 16:56 |
| *** MarkAtwood has joined #openstack | 16:57 | |
| littleidea | creiht: the short answer is 'we did the simplest thing and it works' | 17:00 |
| creiht | :) | 17:00 |
| letterj_ | mtaylor: How do these usually get resolved? http://hudson.openstack.org/view/Swift/job/swift-debian/104/console | 17:01 |
| creiht | There are also some smarts in the proxy that can handle failure conditions that would be difficult if we handed the connection off to the storage node | 17:01 |
| creiht | PUTs would also not be easy, since those are streamed out to the 3 storage nodes | 17:02 |
| *** Ryan_Lane has quit IRC | 17:02 | |
| *** Ryan_Lane has joined #openstack | 17:02 | |
| *** enigma has joined #openstack | 17:04 | |
| *** letterj_ is now known as letterj | 17:10 | |
| *** rlucio has joined #openstack | 17:11 | |
| *** dirakx has quit IRC | 17:12 | |
| *** brd_from_italy has quit IRC | 17:12 | |
| *** rogue780 has joined #openstack | 17:20 | |
| *** hggdh has quit IRC | 17:26 | |
| *** dirakx has joined #openstack | 17:33 | |
| *** maplebed has joined #openstack | 17:38 | |
| *** hggdh has joined #openstack | 17:39 | |
| *** jdarcy has joined #openstack | 17:46 | |
| *** befreax has quit IRC | 17:52 | |
| *** trin_cz has quit IRC | 17:53 | |
| *** joearnold has joined #openstack | 17:58 | |
| *** jdurgin has joined #openstack | 17:59 | |
| *** singh has quit IRC | 18:00 | |
| eday | termie, vishy: https://code.launchpad.net/~eday/nova/internal-api-cleanup/+merge/45152 is updated when you have a moment :) | 18:01 |
| vishy | eday: can we be sure each component will only be dealing with one type of resource? | 18:05 |
| *** irahgel has left #openstack | 18:08 | |
| eday | vishy: well, each component has a main resource (instances for compute, volumes for volume, ...) | 18:10 |
| eday | vishy: for example, in compute, we have compute.API.attach_volume, so we can have verbs on other types too | 18:10 |
| vishy | i guess i don't really see the advantage of removing the noun from the methods | 18:10 |
| *** rogue780 has quit IRC | 18:11 | |
| vishy | afk for 10 min | 18:11 |
| eday | redundancy? | 18:11 |
| vishy | nm, back. I think it makes it a bit less clear, but I guess it isn't a huge issue in my mind | 18:12 |
| *** spectorclan has quit IRC | 18:12 | |
| eday | ie, volume_api.get_volume() seems silly to have the extra noun, but if others disagree I can change | 18:12 |
| *** pharkmillups_ has joined #openstack | 18:15 | |
| *** pharkmillups_ has quit IRC | 18:16 | |
| *** adiantum has joined #openstack | 18:22 | |
| *** fabiand_ has quit IRC | 18:23 | |
| *** adiantum has quit IRC | 18:32 | |
| *** jc_smith has joined #openstack | 18:32 | |
| *** deshantm_cosi has quit IRC | 18:36 | |
| *** rlucio has quit IRC | 18:43 | |
| _0x44 | ttx: Are you around? | 18:48 |
| *** anotherjesse has joined #openstack | 18:48 | |
| *** arthurc has quit IRC | 18:56 | |
| *** dfg_ has quit IRC | 19:02 | |
| _0x44 | ttx: Nevermind, sirp- and I figured out the best way to slice our gordian knot. | 19:02 |
| nelson__ | with a sword, usually. | 19:07 |
| *** maplebed has quit IRC | 19:07 | |
| _0x44 | nelson__: We decided to use a nerf bat. | 19:08 |
| nelson__ | Nerf(tm) always my weapon of choice. | 19:08 |
| *** fabiand_ has joined #openstack | 19:08 | |
| nelson__ | Now ... if only they made a jian. | 19:08 |
| *** kevnfx has quit IRC | 19:11 | |
| *** dfg_ has joined #openstack | 19:11 | |
| *** hadrian has quit IRC | 19:13 | |
| *** WonTu has joined #openstack | 19:15 | |
| *** WonTu has left #openstack | 19:15 | |
| *** maplebed has joined #openstack | 19:19 | |
| sandywalsh | ok, got api-parity tests passing again, now the laborious part ... pep8 | 19:22 |
| *** Podilarius has left #openstack | 19:22 | |
| *** littleidea has quit IRC | 19:24 | |
| sandywalsh | so one of the things I discovered was the string-based image_id's from objectstore aren't compatible with RS API numeric image ids. Is there anything like the instances "internal_id" in objectstore/glance? | 19:25 |
| *** hadrian has joined #openstack | 19:26 | |
| sirp- | glance image ids are just ints right now | 19:30 |
| sandywalsh | do you impose those on top of the objectstore id's? | 19:30 |
| sirp- | well, right now glance doesn't work with the objectstore— it works with swift or using it's own filesystem. but yes, it hides object ids and exposes a single sequential id for an image | 19:32 |
| sandywalsh | ah, right. That's great ... when do we switch?! | 19:33 |
| sandywalsh | thx sirp | 19:33 |
| sirp- | well we should able to switch in cactus if we want to, but i'm guessing we'll keep objectstore around for a while as an alternative | 19:34 |
| sirp- | also fyi, i'm working on getting the python-cloudservers image-create cmd to work with glance and the OpenStack API, i'm dealing with some of those issues as well | 19:35 |
| sandywalsh | you should look at my branch. I do the numeric mapping in there | 19:35 |
| sandywalsh | the image-create call goes through, but fails in dom0 | 19:35 |
| sirp- | a numeric mapping for images, or for instances (different from internal_id?) | 19:35 |
| sandywalsh | images | 19:36 |
| sirp- | ah ok | 19:36 |
| sirp- | i was planning on just passing through the glance id for now, down the road, we'll probably change that to add more flexibility | 19:36 |
| xtoddx | soren: yea, the audit=INFO was an artifact that wasn't used anymore, there are no calls to getLogger('audit') anywhere | 19:37 |
| sandywalsh | yeah, I needed it for boot and other things too | 19:37 |
| sirp- | have you published your branch on lp? | 19:37 |
| sandywalsh | yup https://blueprints.launchpad.net/nova/+spec/openstack-api-parity | 19:37 |
| sirp- | cool *looking* | 19:37 |
| sandywalsh | here's the trace from the image-create call http://etherpad.openstack.org/BXOU0TTj9M | 19:37 |
| *** trin_cz has joined #openstack | 19:38 | |
| sirp- | sandywalsh: put_vdis; error; (111, 'Connection refused' … were you running glance-api and glance-registry? | 19:39 |
| sandywalsh | nope ... are those services? | 19:39 |
| sirp- | yep | 19:39 |
| sandywalsh | :) excuse my ignorance | 19:40 |
| sirp- | right now, xs-snaps only supports glance, so it will need to be running for image-create to work | 19:40 |
| sandywalsh | I haven't looked into glance yet, so no, sorry I didn't know what the steps were | 19:40 |
| sirp- | no worries, ill need to update the docs once get my second_cut patch up | 19:40 |
| sandywalsh | (beyond installing the plugin) | 19:40 |
| sandywalsh | so ... maybe it'll work? | 19:40 |
| sandywalsh | hmm, actually no | 19:41 |
| sandywalsh | hmm, | 19:41 |
| sandywalsh | hehe ... dunno | 19:41 |
| sirp- | it's really pretty simple… basically just check it out, make a conf file that overrides the filesystem_data_dir=/tmp, and then run both of the binaries | 19:41 |
| sandywalsh | I was thinking about objectstore and glance running side by side | 19:41 |
| sandywalsh | I think as soon as we get past bexar, we'll need to solidify glance for the new api work. | 19:42 |
| sandywalsh | the mapping of id's and the kernel/ramdisk/machine -> flavor/os stuff is ugly | 19:42 |
| sirp- | exactly, dubs and _0x44 were just talking about that | 19:42 |
| sandywalsh | sweet! look forward to it. go cactus! | 19:43 |
| soren | xtoddx: Oh, ok. It's just the commit message didn't mention it, so I thought I'd check. | 19:50 |
| jk0 | would anyone mind reviewing https://code.launchpad.net/~jk0/nova/lp697019/+merge/45196 please? :) | 19:57 |
| jk0 | (anyone *except* eday!) | 19:57 |
| jk0 | just kidding eday :P | 19:57 |
| sandywalsh | I need a bzr local commit hook that catches my pep8 violations. :) | 19:59 |
| creiht | lol | 19:59 |
| creiht | sandywalsh: ask notmyname about that :) | 19:59 |
| sandywalsh | and sends an electric jolt to my chair | 19:59 |
| jk0 | what makes it even more difficult is when different pep8.py versions are used | 19:59 |
| notmyname | heh | 19:59 |
| sandywalsh | jk0, yeah, already ran into that "feature" | 20:00 |
| notmyname | alias bzr="echo 'DID YOU RUN THE TESTS' && bzr" | 20:00 |
| jk0 | hah | 20:00 |
| sandywalsh | haha | 20:00 |
| *** rnirmal has quit IRC | 20:00 | |
| sandywalsh | does runtests.sh run pep8? | 20:01 |
| sandywalsh | (if not ...) | 20:01 |
| jk0 | no, at least if it does, it's really quiet about it | 20:01 |
| jk0 | s/at least/or/ | 20:01 |
| sandywalsh | if you look at nearly every merge-prop, the last commit it usually "pep8" | 20:02 |
| sandywalsh | *is | 20:02 |
| *** rh91uk has joined #openstack | 20:02 | |
| sandywalsh | well, my merge props anyway | 20:02 |
| jk0 | I find a lot of times that pops up after a trunk merge | 20:02 |
| jk0 | mostly from the various pep8.py versions | 20:02 |
| *** rh91uk has left #openstack | 20:03 | |
| dabo | It's great that we're so anal about pep8, but have no problem with keywords being used as names in code. | 20:05 |
| *** fabiand_ has quit IRC | 20:06 | |
| *** glenc_ has joined #openstack | 20:10 | |
| dabo | sandywalsh: Here's the error in the pause code: http://paste.openstack.org/show/403/ | 20:11 |
| *** glenc has quit IRC | 20:12 | |
| sandywalsh | but it paused it, right? | 20:12 |
| dabo | dunno - cloudservers list just shows it as 'build' | 20:13 |
| sandywalsh | what about xe or xencenter? | 20:14 |
| *** laurensell has joined #openstack | 20:14 | |
| dabo | the point is that it's looking by name, which is defined as 'instance-' + the internal_id by sqlalchemy | 20:16 |
| dabo | I'm looking into the VMHelper code now. | 20:16 |
| sandywalsh | something must have changed, never used to do that | 20:16 |
| sandywalsh | dabo, pause worked for me yesterday, it was just marking the db as "error". But it actually found and paused the instance. unpause worked too | 20:17 |
| vishy | soren: have you used qemu-nbd at all? | 20:18 |
| sandywalsh | dabo, I'm going back to integration testing on api now, so I'll be revisiting all that. | 20:18 |
| dabo | sandywalsh: yeah, I know - I reported a different error trying to pause a vm that was already paused: http://paste.openstack.org/show/389/ | 20:18 |
| sandywalsh | dabo, I'll need to look into that. That's another one that worked previously. After api. | 20:20 |
| sandywalsh | dabo, not having the xen tests is painful | 20:21 |
| dabo | sandywalsh: there is a discrepancy between what dom0 reports and what cloudservers list reports | 20:21 |
| dabo | that's the problem | 20:21 |
| sandywalsh | yes, that's what I was saying, the db is now getting set to an odd state | 20:22 |
| dabo | I think I'll wipe that xs and start again | 20:22 |
| *** colinnich has quit IRC | 20:31 | |
| uvirtbot | New bug: #697841 in nova "Actions query should include date" [Undecided,In progress] https://launchpad.net/bugs/697841 | 20:32 |
| *** daleolds has quit IRC | 20:34 | |
| *** brd_from_italy has joined #openstack | 20:37 | |
| *** ctennis has quit IRC | 20:50 | |
| *** dirakx has quit IRC | 20:50 | |
| anotherjesse | Anyone know why the trunk ppa is stuck at bzr rev 331? | 20:52 |
| nelson__ | gholt: thanks for your patches. I'm not a big believer in splitting strings using backslashes, though. | 20:54 |
| nelson__ | I'm not writing my code using punched cards, and I doubt anybody else is either. | 20:54 |
| Ryan_Lane | nelson__: that's the rules for the project | 20:55 |
| Ryan_Lane | nelson__: they use PEP8 | 20:55 |
| Ryan_Lane | or was this for something else? :) | 20:55 |
| nelson__ | I use PEP8 also ... but it also says that everything in it is optional. | 20:55 |
| vishy | soren, eday, and whoever: https://code.launchpad.net/~vishvananda/nova/cow/+merge/45293 | 20:55 |
| vishy | could use testing and reviews | 20:56 |
| Ryan_Lane | nelson__: their build tools actually enforce compliance. things won't build if it code isn't PEP8 compliant | 20:56 |
| *** fabiand_ has joined #openstack | 20:57 | |
| nelson__ | Ryan_Lane, the last time I used a punched card, you weren't alive. | 20:58 |
| * Ryan_Lane shrugs | 20:58 | |
| gholt | Well, it has nothing to do with punched cards, or paper tape. | 20:58 |
| Ryan_Lane | the rules is the rules :) | 20:58 |
| creiht | or formal proofs | 20:58 |
| Ryan_Lane | they require something, I respect what they require | 20:58 |
| creiht | :) | 20:58 |
| nelson__ | The last time I programmed on a terminal with 80 columns was, um, 1983. | 20:58 |
| gholt | And it'd still pass builds, no problem. It just shows up on "bad" graphs. :) | 20:58 |
| creiht | we should require formal proofs of every check in :) | 20:59 |
| Ryan_Lane | hah | 20:59 |
| Ryan_Lane | that would be rough | 20:59 |
| creiht | I'm glad that didn't catch on | 20:59 |
| nelson__ | and it's easy to require lines <80 characters, which is obviously better code than code with lines >=80 characters. | 21:00 |
| nelson__ | Anybody using a tool to enforce PEP8 compliance hasn't actually READ PEP8. | 21:00 |
| nelson__ | They're more like guidelines than actual rules. | 21:00 |
| *** damon__ has joined #openstack | 21:00 | |
| gholt | Well, you can argue that until whenever you'd like. But, for the purposes of the project... | 21:01 |
| Ryan_Lane | for these projects, it's rules. "Them's the breaks" | 21:01 |
| nelson__ | sheesh. | 21:01 |
| uvirtbot | New bug: #697861 in swift "object copy does not copy content-type" [Undecided,New] https://launchpad.net/bugs/697861 | 21:01 |
| gholt | Wait, uvirtbot, I thought that was an intended side effect? | 21:01 |
| notmyname | not sure. it seemed weird to me | 21:02 |
| gholt | I thought /you/ were the one that wanted that, lol | 21:02 |
| notmyname | heh | 21:02 |
| gholt | So folks could change the content-type of an uploaded file. | 21:02 |
| notmyname | you still can by setting the header in the copy request. but if you don't set it, it falls back to the "guess-the-type" rules | 21:03 |
| gholt | Ah, okay. I should've probably read the bug submission itself, eh? | 21:03 |
| notmyname | maybe :-) | 21:03 |
| creiht | nelson__: Some tools (including launchpad tools) look better if you keep things <80 chars | 21:04 |
| nelson__ | I don't suppose quoting PEP8 would help? | 21:04 |
| creiht | including the way launchpad sends emails | 21:04 |
| nelson__ | creiht: but you do have a good point. | 21:04 |
| creiht | For whatever it is worth, the PEP8 keeps things fairly consistent | 21:05 |
| termie | eday: lookin now :) | 21:05 |
| dabo | creiht: for some values of "consistent" :) | 21:05 |
| creiht | nelson__: and it is a very common practice | 21:05 |
| creiht | more consistent | 21:05 |
| creiht | either way, as for most of your nitpics nelson__, if this is the worst that you can complain about with the openstack project, then I am a happy dev :) | 21:06 |
| gholt | lp | 21:07 |
| gholt | Damn, who said that? | 21:07 |
| creiht | hehe | 21:07 |
| *** joearnold has quit IRC | 21:10 | |
| *** colinnich has joined #openstack | 21:11 | |
| *** Ryan_Lane is now known as Ryan_Lane|food | 21:12 | |
| soren | vishy: Yes, I've used qemu-nbd. | 21:23 |
| soren | anotherjesse: which trunk ppa? | 21:23 |
| * soren screams | 21:24 | |
| dabo | can anyone review this? It's a dependency for other blueprints. https://code.launchpad.net/~ed-leafe/nova/xenstore-plugin/+merge/44931 | 21:24 |
| soren | termie: Another example of stuff that's nitpicky about id's: elasticfox. | 21:25 |
| soren | termie: Check out this little gem: | 21:25 |
| soren | termie: "ami" : new RegExp("^ami-[0-9a-f]{8}$"), | 21:25 |
| termie | soren: we broke that already | 21:26 |
| soren | termie: So far, that's two out of two tools I've looked at that are more strict than the spec asks them to be. | 21:26 |
| soren | termie: I know. | 21:26 |
| termie | soren: but that is an error in the "be tolerant in what you accept" category | 21:26 |
| soren | termie: Hence the scream. | 21:26 |
| soren | termie: I know. I never suggested it wasn't. | 21:26 |
| soren | The spec says it's a string. | 21:27 |
| termie | soren: well, they probably also hardcodeded the aws urls ;) | 21:27 |
| soren | Not much more than that. | 21:27 |
| *** colinnich has quit IRC | 21:27 | |
| soren | termie: No, they're configurable in the ui. | 21:27 |
| soren | You can add additional endpoints. | 21:27 |
| termie | soren: oh, well we'll just have to drink the ocean then | 21:27 |
| *** rogue780 has joined #openstack | 21:29 | |
| soren | I'll add a "--strict-ec2" sort of option and that truncates to 32 bit in the EC2 api. | 21:30 |
| soren | 32 bits, base-16. | 21:30 |
| *** colinnich has joined #openstack | 21:30 | |
| soren | I'm deeply surprised they can fit their stuff in such a small keyspace. | 21:30 |
| *** lvaughn has quit IRC | 21:30 | |
| termie | soren: they aren;t assigining things randomly | 21:31 |
| soren | Clearly not. | 21:31 |
| *** lvaughn has joined #openstack | 21:31 | |
| anotherjesse | soren: fixed the issue with ppa - user error | 21:32 |
| soren | anotherjesse: Using one of the old ppa references? | 21:33 |
| anotherjesse | using maverick, adding right ppa, forgetting to apt-get update, getting bzr 331 :) | 21:34 |
| *** Ryan_Lane|food is now known as Ryan_Lane | 21:34 | |
| *** pharkmillups_ has joined #openstack | 21:38 | |
| termie | eday: what's the current status of the nova/image/ directory? | 21:39 |
| anotherjesse | soren: we found some bugs in the ppa packages - what is your availability to merge fixes today? | 21:39 |
| soren | anotherjesse: I'm right here :) | 21:39 |
| soren | anotherjesse: For at least another 10 minutes. | 21:39 |
| soren | anotherjesse: What's up? | 21:40 |
| termie | eday: i don't remember hearing about it coming into existence but it seems to be having work done on it from time ot time | 21:40 |
| termie | eday: but i am not aware of anything using it | 21:40 |
| eday | termie: not really sure, there's an interface and some implementations that are not consistent, but a lot of that was being reworked with Glance. I think sirp has some plans for it | 21:40 |
| sleepson- | soren - looks like /etc/init/nova-volume.conf is referring to nova-scheduler | 21:40 |
| termie | sirp-: pingaling ^^ | 21:40 |
| soren | sleepson-: :( | 21:41 |
| soren | sleepson-: Gah. | 21:41 |
| soren | sleepson-: Hah, it also says: description "Nova API server" | 21:41 |
| soren | sleepson-: I suck at cut and paste. | 21:41 |
| sleepson- | :) | 21:41 |
| soren | Er.. | 21:42 |
| anotherjesse | soren: should logging of stderr working again? | 21:43 |
| soren | anotherjesse: Yes. | 21:43 |
| soren | Fix pushed. Hang on, I'll generate new packages. | 21:44 |
| sleepson- | soren anotherjesse - this is the contents of my nova-compute.log: "WARNING:root:Starting compute node" - and that's it, so there may be a logging issue still | 21:45 |
| sleepson- | (after running an instance) | 21:45 |
| soren | sleepson-: This should all be better once xtoddx's newlog magic lands, I think. | 21:45 |
| soren | sleepson-: If not, it should be a simple matter of just passing --verbose=true in nova.conf. The default verbosity level changed with the eventlet merge, it seems. | 21:46 |
| sleepson- | soren: gotcha. thanks | 21:46 |
| soren | Sure thing. | 21:46 |
| soren | anotherjesse, sleepson-: Thanks for catching the nova-volume thing. New packages pushed to Launchpad. | 21:46 |
| sleepson- | soren - awesome - thx mucho - I'll give them a shot | 21:47 |
| soren | sleepson-: It'll be a little bit before they're apt-getable, though. | 21:48 |
| *** kevnfx has joined #openstack | 21:48 | |
| soren | sleepson-: I uploaded source packages. They have to get built first, and then published. The queue is empty, though, so it shouldn't really be more than 10 minutes or so. | 21:49 |
| sleepson- | soren: makes sense. I'll wait for 15, and then check for an update | 21:49 |
| sirp- | termie: pong | 21:50 |
| jk0 | could someone please review https://code.launchpad.net/~jk0/nova/lp697019/+merge/45196 for me? | 21:50 |
| termie | sirp-: wondering what the current working status of the nova/image directory is | 21:50 |
| termie | sirp-: from what i can tell it has been being used in the context of 'openstack' api only | 21:51 |
| soren | sleepson-: https://launchpad.net/~nova-core/+archive/trunk/+packages <--- You can follow the progress here. Any second now a 2011.1~bzr515-0ubuntu0ppa1~maverick2 will turn up. | 21:51 |
| soren | sleepson-: Then it'll get built (takes around 7 minutes) and then a publisher run needs to kick in. Yeah, 15 minutes from now sounds likely. | 21:52 |
| termie | sirp-: it has a bunch of things that look like they could or should be being used by the rest of the system, but nobody working on the rest of the system knows anything about it ;) | 21:52 |
| sirp- | termie: right, i'm not exactly sure how it's being used by the rest of the system, i've only been using it in the context of the openstack api | 21:53 |
| *** hadrian has quit IRC | 21:53 | |
| sirp- | termie: i don't find it very useful, and i think glance ultimately will replace it, but i'm currently updating it to use glance for now so that i can get image-create and image-list working correctly | 21:54 |
| soren | g'night, all | 21:54 |
| termie | sirp-: ftr it doesn't seem like those things can be developed separately | 21:54 |
| termie | sirp-: glance's development does not currently track the needs of the nova | 21:54 |
| termie | sirp-: so i am trying to fix some of htat | 21:55 |
| termie | sirp-: but as long as the 'openstack' api is being developed in a vacuum its tricky :/ | 21:55 |
| sirp- | termie: right, that's def something that will need to be addressed; perhaps should be the major goal for glance in the cactus release; making it usuable not just in the context of OpenStackAPI, but also with any other supported apis | 21:55 |
| termie | sirp-: we already have some ways to solve that issue, glance doesn't necessarily need to do anything differently but the nova/image directory can't be serving the 'openstack' api only | 21:57 |
| sirp- | termie: for the bexar release, though; it really has been fairly isolated, some my interim goal is just to get OpenStackAPI working correctly, and then refactor it later to be more general. If we end up being okay with solely using glance for images, then ImageService can go away completely and simplify somethings | 21:57 |
| *** fabiand_ has quit IRC | 22:01 | |
| eday | sirp-, termie: THe goal of nova/image is to have an API like network/volume/compute that compute can depend on (and possibly volume if we want snapshots of volumes to go there). I was thinking Glance would be one provider, much like s3 and local are now | 22:02 |
| termie | eday: i agree fully with that | 22:02 |
| termie | eday: but the development of that directory has been in a very different direction :/ | 22:02 |
| eday | it's used by ec2 API indirectly via image_service in nova.compute.API (choosing the s3 implemnetation) | 22:02 |
| eday | termie: ahh, ok. I don't know what the correct API for that is, which is why I was punting and letting the glance folks figure it out :) | 22:03 |
| termie | eday, sirp-: by different direction i just mean that it has been developed as something directly integrated with the 'openstack' api | 22:04 |
| termie | eday, sirp-: rather than providing the inter-service api handle on top of a manager/service implementation | 22:04 |
| *** brd_from_italy has quit IRC | 22:04 | |
| termie | eday, sirp-: (in the context of compute.API, network.API and so forth) | 22:05 |
| *** westmaas has quit IRC | 22:06 | |
| termie | eday, sirp-: i guess i am just saying the equivelant of what eday did with the compute and network stuff now has to be done for the image stuff | 22:07 |
| eday | termie: I'm guessing it won't be too hard to map ec2 methods onto it, but I've not looked | 22:07 |
| termie | eday, sirp-: and furthermore the stuff underneath it has to be done as well to map to our existing metaphors | 22:07 |
| eday | vishy: you ok with my branch now? we can rename those methods later as termie suggested | 22:08 |
| termie | eday: i think vishy is in transit | 22:09 |
| *** hadrian has joined #openstack | 22:09 | |
| eday | ahh | 22:09 |
| eday | still no excuse, he has a phone :) | 22:09 |
| *** joearnold has joined #openstack | 22:10 | |
| *** dirakx has joined #openstack | 22:10 | |
| termie | sirp-: anyway, i guess the real question i have for you is what do you plan on currently touching, as there are a bunch of things i plan on moving around | 22:10 |
| eday | I don't think there are any laws against driving and approving merge requests yet | 22:10 |
| termie | eday: ;) | 22:10 |
| termie | sirp-: and what functionality is currently working and tested | 22:10 |
| sirp- | termie: are you planning on moving them around for the bexar release, or in cactus? | 22:10 |
| termie | sirp-: i plan on moving them around right now, if you feel done with them for the bexar release then all the better i will expect that the tests are testing everything you care about | 22:11 |
| sirp- | termie: functionality currently under dev is instance-snapshot create, which maps to ImageService.create | 22:12 |
| *** pharkmillups_ has quit IRC | 22:12 | |
| termie | sirp-: okies | 22:12 |
| sirp- | termie: per my email to the list, i was having Glance generate an image record so we could return that immediately to the user (which is how the OpenStack API works) | 22:13 |
| *** troytoman has joined #openstack | 22:13 | |
| nelson__ | gholt: so what's the next step for merging? Do I have to actually merge it myself? | 22:13 |
| termie | sirp-: sounds like it is mostly how you expect it to look then so whatever thinking i am doing related to it can expect that all the functionality you require is at the very least stubbed out | 22:13 |
| creiht | nelson__: yeah you will have to merge his branch into yours | 22:14 |
| termie | sirp-: i.e. i am not going to make any choices that somehow impede your progress | 22:14 |
| nelson__ | urgh, but I've kept on making changes. | 22:14 |
| *** anotherjesse has quit IRC | 22:14 | |
| nelson__ | so I guess I need to check it out again .... merge, and check back. | 22:14 |
| sirp- | temie: 'not impeding progress' sounds great :) | 22:14 |
| termie | sirp-: :) | 22:14 |
| sirp- | termie: but yeah, if you have a direction to go with it, by all means do that, ill make it work with glance; and if there's any fundamental incompatibilities, ill let ya know | 22:15 |
| termie | sirp-: i'll ping you on any merges, what is your launchpad name? | 22:15 |
| sirp- | termie: rconradharris | 22:15 |
| termie | k | 22:15 |
| gholt | nelson__: Yeah, that's a bit of launchpad/bzr fun. Only the trunk merges are on auto, and that's because of this "hudson" thing. | 22:16 |
| *** anotherjesse has joined #openstack | 22:17 | |
| dabo | PyCon tutorials announced: http://j.mp/e6LTbS Some cloud sessions, but not a thing about OpenStack or Rackspace. :( | 22:17 |
| JordanRinke | off the top of their head, does anyone know the name of all the swift/glance services? | 22:19 |
| *** troytoman has quit IRC | 22:19 | |
| *** miclorb has joined #openstack | 22:21 | |
| *** normanrichards has joined #openstack | 22:21 | |
| sirp- | JordanRinke: do you mean the individual components of glance, or its high-level responsibilities? | 22:23 |
| dubsquared | swift-init object-updater start | 22:23 |
| dubsquared | swift-init container-updater start | 22:23 |
| dubsquared | swift-init object-replicator start | 22:23 |
| dubsquared | swift-init container-replicator start | 22:23 |
| dubsquared | swift-init account-replicator start | 22:23 |
| dubsquared | swift-init object-auditor start | 22:23 |
| dubsquared | swift-init container-auditor start | 22:23 |
| dubsquared | swift-init account-auditor start | 22:23 |
| dubsquared | swift-init account-reaper start | 22:23 |
| dubsquared | those? | 22:23 |
| sirp- | to follow suit: glance-api and glance-registry | 22:24 |
| JordanRinke | just found it in the swift doc... looks like, Proxy, Auth, Object, Container, Account | 22:24 |
| *** ctennis has joined #openstack | 22:24 | |
| *** ctennis has joined #openstack | 22:24 | |
| JordanRinke | sirp- a box running glance, what services would it potentially have running to function? | 22:24 |
| JordanRinke | for instance nova have network, api, scheduler, volume, objectstore, compute - what is that in relation to glance? | 22:25 |
| sirp- | right now, we only have the two services listed above— in the long run we'll probably have some monitoring/cleanup scripts but nothing there yet | 22:25 |
| JordanRinke | so there is no longer, teller, parallax etc? | 22:25 |
| sirp- | jaypipes renamed them, avoided some collisions with some other OSS projects out there | 22:26 |
| sirp- | they both still exist, teller -> api, and parallax -> registry | 22:26 |
| JordanRinke | cool, thx for the help, much appreciated | 22:26 |
| sirp- | JordanRinke: np | 22:26 |
| openstackhudson | Project nova build #356: SUCCESS in 1 min 19 sec: http://hudson.openstack.org/job/nova/356/ | 22:29 |
| JordanRinke | dubsquared: Check out the evernote general notes, I updated the naming scheme - look good? | 22:30 |
| creiht | JordanRinke: and don't forget replicators, updaters, auditors, and the reaper! | 22:32 |
| *** lvaughn_ has joined #openstack | 22:32 | |
| JordanRinke | creiht: Are those services that would potentially not be enabled? | 22:33 |
| creiht | they would be enabled | 22:33 |
| JordanRinke | What I am doing is making a based36 naming scheme | 22:33 |
| JordanRinke | so from a 4char unit code I can tell what services it was initially configured to run | 22:33 |
| anotherjesse | soren: stil up? | 22:33 |
| dubsquared | jordanrinke: looks solid…is that too big thou? just thought about it, and that's large | 22:33 |
| anotherjesse | soren: /usr/sbin/ietadmn is missing from nova_sudoers | 22:33 |
| eday | anotherjesse: he went to bed | 22:33 |
| anotherjesse | anyone else have the ability to push to the trunk ppa? | 22:34 |
| JordanRinke | dubsquared: 12 char total unit name, I think that is reasonable-15 is the NetBIOS limit so that is what I was thinking as a max upper bound | 22:34 |
| *** lvaughn has quit IRC | 22:34 | |
| JordanRinke | all these unit names are going to be automatically created based on deployment configs etc | 22:34 |
| termie | sirp-: what is the prupose of giving them the extra silly names (parallax and teller) when they are just components of a single project | 22:34 |
| dubsquared | jordanrinke: makes sense | 22:34 |
| creiht | JordanRinke: Oh, so you would only be interested if the object-server services are enabled, which would include the server, replication, auditor, and reclaimer | 22:35 |
| creiht | for example | 22:35 |
| creiht | not each individual service for the object server | 22:35 |
| termie | sirp-: i would just as soon not have the words parallax or teller anywhere in the projects | 22:35 |
| sirp- | termie: they *were* called that to make it a little easier to discuss the components individually, but they *arent* called that anymore, so ++ on removing any reference to that | 22:36 |
| JordanRinke | ok sweet, so then I am looking at Proxy, Auth, Account, Object, Container as services then right? | 22:36 |
| termie | sirp-: ah, okay | 22:36 |
| JordanRinke | for swift that is | 22:36 |
| creiht | right | 22:36 |
| JordanRinke | excellent | 22:37 |
| JordanRinke | thx | 22:37 |
| openstackhudson | Project nova build #357: SUCCESS in 1 min 18 sec: http://hudson.openstack.org/job/nova/357/ | 22:39 |
| dubs | i got a traceback when running python setup.py build_sphinx - http://paste.openstack.org/show/405/ - anyone know what that's all about? | 22:39 |
| dubs | am I missing some package? | 22:40 |
| JordanRinke | dubsquared: updated the evernote with an example tree too | 22:43 |
| *** littleidea has joined #openstack | 22:45 | |
| *** fabiand_ has joined #openstack | 22:47 | |
| *** allsystemsarego_ has quit IRC | 22:47 | |
| dubs | ah, i guess Maverick has sphinx 0.6.6 and doesn't yet have sphinx.builders.manpage.ManualPageBuilder, which was added in 1.0 | 22:49 |
| dubsquared | jordanrinke: nice! | 22:53 |
| *** matiu has joined #openstack | 22:55 | |
| *** matiu has quit IRC | 22:56 | |
| *** matiu has joined #openstack | 22:56 | |
| *** matiu has joined #openstack | 22:56 | |
| *** fabiand_ has left #openstack | 22:59 | |
| openstackhudson | Project nova build #358: SUCCESS in 1 min 21 sec: http://hudson.openstack.org/job/nova/358/ | 22:59 |
| *** johnpur has quit IRC | 23:02 | |
| vishy | dendrobates: are you here? | 23:03 |
| dirakx | Hi all, I'm working on getting up a project on nova, but nova-manage project zipfile is not working for me, openssl generates the certificates..how can i .zip the file without the .py script in order to source novarc, and run an instance ?. | 23:04 |
| *** ppetraki has quit IRC | 23:05 | |
| dendrobates | vishy: sup | 23:06 |
| *** anotherjesse_ has joined #openstack | 23:07 | |
| sirp- | dirakx: is nova-manage project raising an exception, or is the zip file its generating just not working? | 23:08 |
| vishy | dendrobates: so we have some potential fixes to the ppa | 23:09 |
| vishy | dendrobates: it looks like it is automatically built via ~openstack-ubuntu-packagers branch | 23:09 |
| dendrobates | vishy: it is | 23:09 |
| dirakx | sirp-: nova-manage generates an exception..about not been able to execute an openssl comand.. | 23:10 |
| vishy | dendrobates: you and soren are the only ones with the ability to upload there | 23:10 |
| vishy | dendrobates: and soren is in bed | 23:10 |
| dirakx | sirp-: doing this openssl comand alone is working. but not with nova-manage | 23:10 |
| sirp- | dirakx: could you make a paste using http://paste.openstack.org/ ? | 23:11 |
| dendrobates | vishy: do you have a diff? | 23:11 |
| vishy | dendrobates: is it possible that we can get someone else added so we can do fixes | 23:11 |
| *** anotherjesse has quit IRC | 23:11 | |
| *** anotherjesse_ is now known as anotherjesse | 23:11 | |
| vishy | dendrobates: we're still debugging | 23:11 |
| dirakx | sirp-: im working this porblem in this question https://answers.launchpad.net/nova/+question/139603 ;) | 23:11 |
| dendrobates | vishy: I don;t think I can add people to that team, let me check | 23:12 |
| dendrobates | vishy: ah, I am member because I'm an Ubuntu dev. I can't add to it, but I can push code for you. | 23:13 |
| vishy | dendrobates: ok will you be around for a bit? | 23:14 |
| dendrobates | yep | 23:14 |
| vishy | dendrobates: ok i'll get you a branch/diff asap | 23:14 |
| JordanRinke | anyone want to take a look at this naming scheme and provide some feedback before I tweet it out and try to convince everyone it is awesome? | 23:14 |
| JordanRinke | http://www.jordanrinke.com/2011/01/05/based36-naming-scheme-for-openstack-servers/ | 23:15 |
| dirakx | sirp-: anyway here is a paste http://paste.openstack.org/show/406/ | 23:16 |
| creiht | JordanRinke: so how would these be used? | 23:18 |
| sirp- | dirakx: thx, i haven't run into this yet, so i'm just grep'ing around the source atm | 23:18 |
| dirakx | sirp-: oh thanks..i grep nova-manage without finding a clue. | 23:19 |
| *** gondoi has quit IRC | 23:19 | |
| creiht | I'm still a bit confused as to what problem you are trying to install | 23:19 |
| creiht | solve | 23:19 |
| JordanRinke | creiht: so, when I build out a large implementation I would just select what features what boxes would have on it, and then it would name them and build them accordingly (provided all the backend work was done obviously) | 23:19 |
| JordanRinke | then, if a box ever had a problem in a customer environment I would know what/where it was just from the name | 23:20 |
| *** dendrobates is now known as dendro-afk | 23:20 | |
| creiht | oh a naming scheme for the boxes? | 23:20 |
| JordanRinke | could also easily analyze designs/installs from anyone using my setup config/naming standard | 23:20 |
| JordanRinke | yeah | 23:20 |
| JordanRinke | naming of the physical machines | 23:20 |
| *** dendro-afk is now known as dendrobates | 23:20 | |
| creiht | so if I decide to install another service, I have to rename the machine? | 23:20 |
| JordanRinke | in large scale installs that type of change wouldn't be common and would typically happen with a migration box anyways | 23:21 |
| JordanRinke | but to answer the question specifically, yes | 23:22 |
| creiht | Seems a bit silly to me, but whatever makes you happy | 23:22 |
| creiht | :) | 23:22 |
| dirakx | sirp-: the manager.get_credentials method. | 23:23 |
| creiht | it seems that for most enterprise installs, you are going to have a handful of different types of machines | 23:24 |
| creiht | and that is usually pretty easy to name in context of how they are setting things up, without having to look up some matrix | 23:24 |
| creiht | for example for swift at rs, we basically have proxy nodes and storage nodes | 23:25 |
| creiht | all proxies are set up the same, and all storage nodes are set up the same | 23:25 |
| JordanRinke | creiht: right, I am trying to abstract that one level so a single globalized naming scheme could apply to all configs | 23:25 |
| *** enigma has left #openstack | 23:25 | |
| JordanRinke | I would expect that people would learn that 01j4 boxes were swift boxes etc, but the standard would apply to any install using it | 23:26 |
| *** trin_cz has quit IRC | 23:26 | |
| dubs | JordanRinke: why not just simple namespaces? | 23:27 |
| JordanRinke | thinking large scale with the possibility of moving systems across providers, and the potential for the division between providers to be less clear | 23:27 |
| creiht | so when we run out of bits, and have to add more, then every install has to rename there machines? | 23:27 |
| creiht | their | 23:27 |
| JordanRinke | well, that is why I gave it 20 bits, the assumption (always the weak point) is that it we won't just keep adding service names | 23:28 |
| creiht | Except what happens when we add a new project like block storage which is likely to have as many services as swift? | 23:28 |
| JordanRinke | which is why I said it would work for a number of years also :-D | 23:29 |
| creiht | and what happens when we have more projects that are optional | 23:29 |
| creiht | I doubt it will last until the end of this year :) | 23:29 |
| JordanRinke | that depends how they are implemented, but that is definitely the issue with an coded naming scheme | 23:30 |
| creiht | either way, I don't think it is a good idea as a global strategy | 23:30 |
| sirp- | dirakx: what platform are you running on? | 23:32 |
| dirakx | sirp-: ubuntu 10.10 | 23:33 |
| *** schisamo has joined #openstack | 23:34 | |
| openstackhudson | Project nova build #359: SUCCESS in 1 min 15 sec: http://hudson.openstack.org/job/nova/359/ | 23:34 |
| JordanRinke | creiht: have a better suggestion that allows for globalization of naming standards and any style config? | 23:35 |
| creiht | don't do it | 23:36 |
| creiht | :) | 23:36 |
| JordanRinke | why not? | 23:36 |
| creiht | because, while it may sound great in theory, in practice you are going to cause more problems than you solve | 23:37 |
| sirp- | dirakx: can you run this in your nova-root directory find -name openssl.cnf -print | 23:38 |
| creiht | or we could do it right, and offer globalized naming as a service | 23:38 |
| creiht | ;) | 23:38 |
| creiht | but then again, that is just my opinion | 23:39 |
| dirakx | sirp-: ./bin/openssl.cnf | 23:39 |
| dirakx | ./contrib/nova/CA/openssl.cnf | 23:39 |
| *** dragondm has quit IRC | 23:39 | |
| JordanRinke | creiht: well my initial thought was to just use some standard names but I figured people are going to implement this in lots of different ways | 23:39 |
| dubs | dirakx: thats from the root of your nova directory? | 23:39 |
| JordanRinke | creiht: one thought is to just step it up to a 6char ID which gives 31 service bits | 23:40 |
| dirakx | dubs yep. | 23:40 |
| creiht | IDv6? | 23:40 |
| creiht | :) | 23:40 |
| JordanRinke | ohh buzz word! win! haha | 23:40 |
| dubs | dirakx: then you've files appear to have been moved around. from the root you should see CA/openssl.cnf | 23:40 |
| dirakx | dubs should i copy those from contrib ? . | 23:41 |
| creiht | each service could have a uuid, then name each server as the checksum of the ids of the servers :) | 23:42 |
| dubs | it looks like to me your contrib/nova dir *is* your root nova dir | 23:42 |
| * creiht is not serious btw | 23:42 | |
| dirakx | dubs running from contri/nova i'm getting a database error . | 23:45 |
| dirakx | something like this http://paste.openstack.org/show/408/ | 23:46 |
| *** trin_cz has joined #openstack | 23:46 | |
| sirp- | dirakx: two part question: in contrib/nova is there a file called "HACKING" and b) is there nova.sqlite3 or was that in the other directory? | 23:48 |
| openstackhudson | Project nova build #360: SUCCESS in 1 min 19 sec: http://hudson.openstack.org/job/nova/360/ | 23:49 |
| *** daleolds has joined #openstack | 23:51 | |
| dirakx | sirp-: a) yes | 23:51 |
| *** opengeard has joined #openstack | 23:52 | |
| dirakx | sirp-: and b...i'm trying to find that file :( ..xtrange. | 23:54 |
| sirp- | dirakx: hmm, one option is to start over with a fresh nova checkout and make sure all the commands are run from that nova-root dir; it's possible running cmds from various directories caused some files to get misplaced | 23:55 |
| *** rogue780 has quit IRC | 23:56 | |
| dirakx | sirp-: i should do that..also because i tried with nova.sh and with packages..so is just etter to start over again, | 23:56 |
| dirakx | sirp-: and dubs thanks for your great help. :) | 23:57 |
| sirp- | dirakx: np | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!