| *** lloydde has quit IRC | 00:00 | |
| *** rbasak has quit IRC | 00:06 | |
| *** pixelbeat has quit IRC | 00:07 | |
| *** epim has quit IRC | 00:13 | |
| *** mcclurmc_ has quit IRC | 00:14 | |
| *** torgomatic has quit IRC | 00:15 | |
| *** bsza has joined #openstack-dev | 00:16 | |
| *** jasonz has left #openstack-dev | 00:17 | |
| *** danwent has quit IRC | 00:18 | |
| *** utlemming has joined #openstack-dev | 00:18 | |
| *** danwent has joined #openstack-dev | 00:22 | |
| *** andresambrois has quit IRC | 00:26 | |
| *** spiffxp has quit IRC | 00:26 | |
| *** mgius has quit IRC | 00:26 | |
| *** torgomatic has joined #openstack-dev | 00:28 | |
| *** mnewby_ has joined #openstack-dev | 00:40 | |
| *** mnewby has quit IRC | 00:43 | |
| *** issackelly has joined #openstack-dev | 00:45 | |
| issackelly | It seems weird the way this is presented http://dl.dropbox.com/u/70585674/Screen%20Shot%202012-05-10%20at%204.34.13%20PM.png It says five projects, highlights three, and then lists two as "new projects" | 00:46 |
|---|---|---|
| *** edygarcia has quit IRC | 00:49 | |
| *** maplebed has quit IRC | 00:52 | |
| *** vikram1 has joined #openstack-dev | 00:52 | |
| *** rods has quit IRC | 00:53 | |
| *** ohnoimdead has quit IRC | 00:53 | |
| *** danwent has quit IRC | 00:53 | |
| *** danwent has joined #openstack-dev | 00:54 | |
| *** jemartin has joined #openstack-dev | 00:58 | |
| *** danwent has quit IRC | 01:00 | |
| *** asdfasdf has quit IRC | 01:00 | |
| *** danwent has joined #openstack-dev | 01:02 | |
| *** koolhead17 has quit IRC | 01:03 | |
| *** dprince has joined #openstack-dev | 01:04 | |
| *** danwent_ has joined #openstack-dev | 01:04 | |
| *** danwent has quit IRC | 01:08 | |
| *** danwent_ has quit IRC | 01:09 | |
| *** jdurgin has quit IRC | 01:18 | |
| *** jemartin has quit IRC | 01:21 | |
| *** jgriffith has quit IRC | 01:22 | |
| *** utlemming has quit IRC | 01:22 | |
| *** shang has quit IRC | 01:23 | |
| *** bsza has quit IRC | 01:27 | |
| *** adjohn has joined #openstack-dev | 01:29 | |
| *** littleidea has quit IRC | 01:30 | |
| *** torgomatic has quit IRC | 01:33 | |
| *** shang has joined #openstack-dev | 01:36 | |
| *** danwent has joined #openstack-dev | 01:36 | |
| *** johnpostlethwait has quit IRC | 01:39 | |
| *** Mandell has quit IRC | 01:40 | |
| *** jgriffith has joined #openstack-dev | 01:41 | |
| *** hub-cap has joined #openstack-dev | 01:53 | |
| *** novas0x2a|lapto1 has quit IRC | 01:55 | |
| *** roge has quit IRC | 01:56 | |
| *** hub_cap has quit IRC | 01:57 | |
| *** littleidea has joined #openstack-dev | 01:58 | |
| *** littleidea has left #openstack-dev | 02:00 | |
| *** rgoodwin has quit IRC | 02:00 | |
| *** Ryan_Lane has quit IRC | 02:09 | |
| *** vikram1 has left #openstack-dev | 02:12 | |
| *** jgriffith has quit IRC | 02:13 | |
| *** jgriffith has joined #openstack-dev | 02:28 | |
| *** dtroyer_zzz is now known as dtroyer | 02:28 | |
| *** tryggvil_ has quit IRC | 02:37 | |
| *** issackelly has quit IRC | 02:39 | |
| *** mnewby has joined #openstack-dev | 02:39 | |
| *** dprince has quit IRC | 02:43 | |
| *** johnpostlethwait has joined #openstack-dev | 02:55 | |
| *** harlowja has quit IRC | 02:58 | |
| *** jgriffith has quit IRC | 03:01 | |
| *** lloydde has joined #openstack-dev | 03:06 | |
| *** markvoelker has joined #openstack-dev | 03:08 | |
| *** markvoelker has left #openstack-dev | 03:08 | |
| *** dhellmann has joined #openstack-dev | 03:09 | |
| *** benner has quit IRC | 03:10 | |
| *** adjohn has quit IRC | 03:13 | |
| *** lloydde has quit IRC | 03:18 | |
| *** lloydde has joined #openstack-dev | 03:19 | |
| *** lloydde_ has joined #openstack-dev | 03:22 | |
| *** lloydde has quit IRC | 03:23 | |
| *** markmcclain has joined #openstack-dev | 03:29 | |
| *** benner has joined #openstack-dev | 03:29 | |
| *** mjfork has quit IRC | 03:30 | |
| *** utlemming has joined #openstack-dev | 03:40 | |
| *** Mandell has joined #openstack-dev | 03:42 | |
| *** utlemming has quit IRC | 03:46 | |
| *** littleidea has joined #openstack-dev | 03:57 | |
| *** troytoman-away is now known as troytoman | 04:03 | |
| *** sandywalsh has quit IRC | 04:04 | |
| *** johnpostlethwait has quit IRC | 04:06 | |
| *** koolhead17 has joined #openstack-dev | 04:10 | |
| *** ywu has quit IRC | 04:10 | |
| *** markmcclain has quit IRC | 04:14 | |
| *** anderstj has joined #openstack-dev | 04:20 | |
| *** markmcclain has joined #openstack-dev | 04:24 | |
| *** adjohn has joined #openstack-dev | 04:28 | |
| *** johnpostlethwait has joined #openstack-dev | 04:29 | |
| *** jgriffith has joined #openstack-dev | 04:33 | |
| *** Ryan_Lane has joined #openstack-dev | 04:47 | |
| *** mnewby has quit IRC | 05:02 | |
| *** mnewby has joined #openstack-dev | 05:03 | |
| *** sdague_ has joined #openstack-dev | 05:05 | |
| *** sdague has quit IRC | 05:08 | |
| *** openstack has joined #openstack-dev | 05:11 | |
| *** ahale has joined #openstack-dev | 05:12 | |
| *** mcclurmc_ has joined #openstack-dev | 05:12 | |
| *** seats_ has joined #openstack-dev | 05:12 | |
| *** jgriffith has quit IRC | 05:12 | |
| *** seats has quit IRC | 05:12 | |
| *** seats_ is now known as seats | 05:12 | |
| *** glenc has joined #openstack-dev | 05:13 | |
| *** dtroyer is now known as dtroyer_zzz | 05:17 | |
| *** bcwaldon has joined #openstack-dev | 05:18 | |
| *** johnpostlethwait has quit IRC | 05:21 | |
| *** jaypipes has joined #openstack-dev | 05:22 | |
| *** sacharya has quit IRC | 05:23 | |
| *** vincentricci has left #openstack-dev | 05:29 | |
| *** davidha has quit IRC | 05:30 | |
| *** davidha_who_took has joined #openstack-dev | 05:30 | |
| *** troytoman is now known as troytoman-away | 05:30 | |
| *** mrunge has quit IRC | 05:32 | |
| *** mrunge has joined #openstack-dev | 05:32 | |
| *** davidha_who_took is now known as davidha | 05:33 | |
| *** littleidea has quit IRC | 05:42 | |
| *** hattwick has quit IRC | 06:01 | |
| *** asalkeld has quit IRC | 06:06 | |
| eglynn__ | Mikal: mornin'/evenin' ;) | 06:20 |
| eglynn__ | Mikal: have you any views on the usefullness of https://review.openstack.org/7302 for your image replication tool? | 06:20 |
| *** pmezard has joined #openstack-dev | 06:22 | |
| mikal | eglynn__: heya | 06:23 |
| mikal | I haven't looked yet to be honest, I've been off sick most of this week | 06:24 |
| mikal | I shall look now | 06:24 |
| *** anderstj has quit IRC | 06:24 | |
| mikal | Wow, that's a lot of review comments | 06:24 |
| mikal | So... awsome doesn't do integer ids at all, right? | 06:25 |
| *** lloydde_ has quit IRC | 06:25 | |
| mikal | I wonder if we should just drop this id thing entirely and hand uuids back to ec2 clients | 06:25 |
| mikal | I wonder what would break | 06:25 |
| *** lloydde has joined #openstack-dev | 06:28 | |
| eglynn__ | Mikal: well, the illusion that the clients are using EC2 would break I guess ... | 06:29 |
| mikal | Yeah, I wonder what assumptions tools make | 06:29 |
| mikal | I suspect Amazon handles this stuff by making you upload images once per region? | 06:29 |
| eglynn__ | Mikal: yeah, I don't think the ami-* IDs are region-portable | 06:30 |
| mikal | eglynn__: your code looks reasonable to me... | 06:30 |
| eglynn__ | cool | 06:31 |
| mikal | I think the other reviewers have fair points, but this isn't a simple problem. | 06:32 |
| mikal | Basically the ec2 api makes replication really hard to do properly | 06:32 |
| *** lloydde has quit IRC | 06:32 | |
| mikal | But it was clear from the design summit session that people really want replication | 06:32 |
| mikal | I sure do. | 06:32 |
| mikal | I wonder if the awsome people have experienced any tool problems with just using uuids for image ids. I should ask. | 06:33 |
| *** davidkranz has quit IRC | 06:34 | |
| eglynn__ | mikal: IIRC AWSOME just relies on the ID mapping maintained internally by nova, but it would be good to ask those folks what their thoughts are ... | 06:35 |
| mikal | I thought I saw it return UUIDs in the euca-describe-images output | 06:35 |
| mikal | I might be confused though | 06:35 |
| *** davidkranz has joined #openstack-dev | 06:41 | |
| *** zaitcev has quit IRC | 06:46 | |
| *** mrunge has quit IRC | 06:46 | |
| *** mrunge has joined #openstack-dev | 06:49 | |
| mikal | eglynn__: you still around? | 06:52 |
| *** markvoelker has joined #openstack-dev | 06:53 | |
| *** shang has quit IRC | 06:54 | |
| eglynn__ | mikal: I'm just about to leave ... the ami-* style ID would be used in say the euca-run-instances tool to identify the image to boot off | 06:57 |
| eglynn__ | mikal: not sure this usage could shift to the native glance UUID without breaking lots of scripts etc. | 06:57 |
| *** Shrews has quit IRC | 06:58 | |
| *** Shrews has joined #openstack-dev | 06:58 | |
| *** mrunge has quit IRC | 06:58 | |
| *** mrunge has joined #openstack-dev | 06:58 | |
| eglynn__ | mikal: I gotta scoot off now, we can continue the conversation over email or later on IRC | 07:00 |
| mikal | Okie | 07:01 |
| mikal | I just confirmed, awsome just uses the uuid | 07:01 |
| mikal | Not sure what it breaks though | 07:01 |
| *** heyho has quit IRC | 07:03 | |
| *** eglynn__ has quit IRC | 07:05 | |
| *** Vek has joined #openstack-dev | 07:08 | |
| *** Vek has quit IRC | 07:08 | |
| *** Vek has joined #openstack-dev | 07:08 | |
| *** reidrac has joined #openstack-dev | 07:11 | |
| *** mcclurmc_ has quit IRC | 07:13 | |
| *** mszilagyi has quit IRC | 07:15 | |
| *** mnewby has quit IRC | 07:17 | |
| *** Mandell has quit IRC | 07:20 | |
| *** mnot has joined #openstack-dev | 07:28 | |
| *** adjohn has quit IRC | 07:33 | |
| *** pixelbeat has joined #openstack-dev | 07:33 | |
| *** danwent has quit IRC | 07:38 | |
| *** Stackops-Jorge has joined #openstack-dev | 07:41 | |
| *** clayg has quit IRC | 07:42 | |
| *** clayg_ is now known as clayg | 07:42 | |
| *** eglynn__ has joined #openstack-dev | 07:42 | |
| *** clayg_ has joined #openstack-dev | 07:43 | |
| *** apevec has joined #openstack-dev | 07:46 | |
| *** eglynn__ has quit IRC | 07:47 | |
| *** apevec has quit IRC | 07:48 | |
| *** adjohn has joined #openstack-dev | 07:48 | |
| *** eglynn__ has joined #openstack-dev | 07:50 | |
| *** dachary has joined #openstack-dev | 07:55 | |
| *** hattwick has joined #openstack-dev | 07:58 | |
| *** mancdaz has joined #openstack-dev | 08:00 | |
| *** adjohn has quit IRC | 08:01 | |
| *** eglynn__ has quit IRC | 08:05 | |
| *** vincentricci has joined #openstack-dev | 08:15 | |
| *** adjohn has joined #openstack-dev | 08:24 | |
| *** eglynn__ has joined #openstack-dev | 08:30 | |
| *** apevec has joined #openstack-dev | 08:39 | |
| *** Ryan_Lane has quit IRC | 08:40 | |
| *** gael has joined #openstack-dev | 08:47 | |
| *** danpb has joined #openstack-dev | 08:51 | |
| eglynn__ | mikal: you still there? (sorry, I drop my kids to school) | 08:52 |
| eglynn__ | mikal: about that glance UUID versus ami-* style IDs discussion earlier | 08:52 |
| mikal | eglynn__: np | 08:54 |
| *** derekh has joined #openstack-dev | 08:55 | |
| eglynn__ | mikal: AFAIK DescribeInstances just returns the ami-* style ID and the image name, but not the glance UUID | 08:55 |
| *** darraghb has joined #openstack-dev | 08:55 | |
| mikal | I was using describe images | 08:55 |
| * mikal talks a look at describe instances | 08:55 | |
| eglynn__ | mikal: sorry I meant DescribeImages | 08:55 |
| mikal | Yeah, ok | 08:56 |
| mikal | That's true for the ec2 api built into OS | 08:56 |
| mikal | However, awsome just returns the UUI | 08:56 |
| mikal | D | 08:56 |
| mikal | I've asked them if they have reports of this causing tool breakages or not | 08:56 |
| eglynn__ | mikal: similarly I thought the OS impl of RunInstances allows either the ami-* style ID or image name, but not the glance UUID | 08:57 |
| mikal | There has to be an amazon spec for what's allowed in that filed | 08:57 |
| mikal | field even | 08:57 |
| mikal | I should go find it | 08:57 |
| eglynn__ | mikal: so I'm wondering how AWSOME just passes thru' the glance UUID | 08:58 |
| mikal | Oh, I see your point | 08:58 |
| *** anniec72 has joined #openstack-dev | 08:58 | |
| *** markmc has joined #openstack-dev | 08:58 | |
| * eglynn__ needs to take a closer at the AWSOME code ... | 08:59 | |
| mikal | Oh, the code seems to return image['id'] as the id | 08:59 |
| * mikal goes to find image's birthplace | 08:59 | |
| *** vincentricci has quit IRC | 09:00 | |
| mikal | Ahhh, it calls client.get_images() and then just uses the id for each image | 09:01 |
| mikal | I am being ordered to come to dinner, but will be back in a bit | 09:02 |
| mikal | I suspect that the id being returned by the OS API is a uuid? | 09:03 |
| eglynn__ | mikal: cool, no rush, laters ... | 09:04 |
| eglynn__ | mikal: and yep, the OS API returns the UUID, whereas the ami-* IDs are only relevant to EC2-stylee usage ... | 09:05 |
| hugokuo2 | Hi all | 09:09 |
| hugokuo2 | [Swift]Does any one knows the hashes.pkl come from ? | 09:09 |
| hugokuo2 | After an object been rsync to remote node by replicator | 09:10 |
| hugokuo2 | I found that a hashes.pkl file been created | 09:11 |
| *** davidkranz has quit IRC | 09:11 | |
| hugokuo2 | I want to know the usage of that pkl file . | 09:11 |
| hugokuo2 | I try to open it by pikle . first value is the last 3 characters of object hash | 09:12 |
| hugokuo2 | but what's the meaning of second value ? | 09:12 |
| *** davidkranz has joined #openstack-dev | 09:26 | |
| *** armax has joined #openstack-dev | 09:27 | |
| *** eneabio has joined #openstack-dev | 09:35 | |
| eneabio | I have a problem with scheduler, in particular least_cost.py... | 09:35 |
| mikal | eglynn__: yeah, so I think awsome is just using the glance uuid | 09:39 |
| eglynn__ | mikal: yep, from the looks of the code for RunInstances, it needs a UUID to be passed, as it just passes thru' whatever it gets | 09:42 |
| eglynn__ | mikal: so if it happened to receive an ami-* style ID (which would be normal usage for EC2 clients), the server boot would fail | 09:43 |
| eglynn__ | mikal: (as nearly the first thing nova does is to check if the passed imageRef param is UUID-like and if not rejects it) | 09:44 |
| eglynn__ | mikal: so seems to me AWSOME is broken in that respect | 09:44 |
| eglynn__ | mikal: (well at least it breaks the normal EC2 client expectation about the type of IS to pass...) | 09:45 |
| zykes- | What kind of networking switches is optimal to use for Nova / Swift ? | 09:45 |
| eglynn__ | mikal: backing up a bit ... | 09:46 |
| mikal | eglynn__: yeah, in a perfect world we would discover that uuids magically meet the spec somehow, and then we'd just forget about this integer mapping entirely | 09:47 |
| eglynn__ | mikal: just chatting with markmc about this, his take was the key issue is just that the current sytle of ami-* ID generation is misleading coz it might give the wrong impression that IDs are region-portable | 09:47 |
| eglynn__ | mikal: (as they kinda look and smell non-random, ami-00000001, ami-00000002 ... etc) | 09:47 |
| mikal | Yeah, ok | 09:47 |
| mikal | As in we expect users to have some sort of script which looks up the ami id for that region before launching an instance? | 09:48 |
| mikal | Instead of assuming it is the same in all regions? | 09:48 |
| markmc | right, since that's the way EC2 works | 09:48 |
| eglynn__ | mikal: more that we make it absolutely clear that there should be no expectation of region-portability | 09:48 |
| eglynn__ | mikal: there isn't currently, but maybe folks might assume there is coz the Ids look similar | 09:49 |
| mikal | That would make replication easier, the replicator would just need to ensure an ec2 id is allocated, but not that its nessesarily the same as that of the master region. | 09:49 |
| eglynn__ | mikal: I was comming at this from the direction of establishing that portability in "most" cases (modulo collisions) | 09:49 |
| mikal | It would be _nice_ to have portable ids, it just seems not possible | 09:50 |
| eglynn__ | mikal: yep, perfect portability is not possible without a central authority (122 bits into 32 bits doesn't go unfortunately :( ...) | 09:51 |
| *** armax has quit IRC | 09:51 | |
| mikal | Agreed | 09:51 |
| eglynn__ | mikal: if we could rewind back to when glance changed from integer image IDs to UUIDs (pre-Essex I think) | 09:52 |
| eglynn__ | mikal: we could ensure for the Amazon image formats (ami, ari, aki) a reduced UUID space was alllocated from | 09:53 |
| eglynn__ | mikal: (i.e. instead of the full 122 bit randomness in the generated UUIDs, we could just have 32 bits of randomness for those image types) | 09:53 |
| mikal | I do note that the ec2 api spec just says "string" for that field | 09:54 |
| mikal | http://docs.amazonwebservices.com/AWSEC2/latest/APIReference/ApiReference-query-DescribeImages.html | 09:54 |
| eglynn__ | mikal: de facto String == 'a[mkr]i-[a-f0-9]{8}' | 09:54 |
| mikal | Yeah, it comes down to what people hardcoded into their tools... | 09:55 |
| eglynn__ | => 32 bits of randomness | 09:55 |
| mikal | But a uuid does meet the requirements of the spec as written | 09:55 |
| eglynn__ | mikal: but we could allocate UUIDs from a "reduced" domain (just for the amzn images) | 09:56 |
| eglynn__ | mikal: we being the glance registry in this case | 09:56 |
| mikal | Or we could just give up and assume people already have tools to handle this because of how amazon does it. | 09:56 |
| eglynn__ | mikal: allocating the IDs from a reduced space would enable a perfect mapping going forward | 09:56 |
| mikal | I don't have a good feel either way | 09:56 |
| mikal | eglynn__: true | 09:57 |
| eglynn__ | mikal: "going forward" == only applying to fresh images | 09:57 |
| mikal | Yep | 09:57 |
| eglynn__ | mikal: yeah, I'll run the reduced UUID space idea past bcwaldon and vish later on (or on gerrit) | 09:58 |
| mikal | Okie | 09:58 |
| mikal | Thanks for taking the time to think about this by the way | 09:58 |
| mikal | I really should cleanup my prototype code and put a review up | 09:58 |
| eglynn__ | mikal: np ... otherwise I'll need to look at extending the native OS API to accept ami-* IDs as ImageRefs | 09:59 |
| eglynn__ | mikal: yep, t'would be good to have a look at that replication code | 09:59 |
| mikal | Its nothing special | 10:01 |
| mikal | I would have done it this week except for being off sick | 10:01 |
| mikal | I'll try and make some time over the weekend to get it done | 10:01 |
| *** tryggvil has joined #openstack-dev | 10:08 | |
| *** asalkeld has joined #openstack-dev | 10:12 | |
| *** tryggvil has quit IRC | 10:14 | |
| *** tryggvil has joined #openstack-dev | 10:16 | |
| *** berendt has joined #openstack-dev | 10:39 | |
| *** adjohn has quit IRC | 10:50 | |
| *** rkukura has joined #openstack-dev | 10:54 | |
| *** mjfork has joined #openstack-dev | 11:00 | |
| *** bsza has joined #openstack-dev | 11:01 | |
| *** dachary has quit IRC | 11:09 | |
| zykes- | anyone from hp here ? | 11:09 |
| *** markvoelker1 has joined #openstack-dev | 11:20 | |
| *** hub_cap has joined #openstack-dev | 11:23 | |
| *** hub_cap has quit IRC | 11:24 | |
| *** sandywalsh has joined #openstack-dev | 11:26 | |
| mikal | I am confused about where the database for run_tests live. It is angry about the schema version... | 11:41 |
| *** sandywalsh has quit IRC | 11:43 | |
| *** GheRivero has quit IRC | 11:50 | |
| *** berendt has quit IRC | 11:55 | |
| *** tryggvil has quit IRC | 11:59 | |
| *** mrunge has quit IRC | 12:03 | |
| *** lts has joined #openstack-dev | 12:06 | |
| *** salgado has joined #openstack-dev | 12:07 | |
| *** salgado has joined #openstack-dev | 12:07 | |
| *** dolphm has joined #openstack-dev | 12:15 | |
| *** dachary has joined #openstack-dev | 12:17 | |
| *** alaski has joined #openstack-dev | 12:19 | |
| *** hub_cap has joined #openstack-dev | 12:38 | |
| *** tryggvil_ has joined #openstack-dev | 12:43 | |
| *** dprince has joined #openstack-dev | 12:45 | |
| *** edygarcia has joined #openstack-dev | 12:46 | |
| *** rohitk has joined #openstack-dev | 12:50 | |
| *** sandywalsh has joined #openstack-dev | 12:55 | |
| *** statik has joined #openstack-dev | 12:57 | |
| *** jaypipes has quit IRC | 12:59 | |
| *** sdague_ is now known as sdague | 13:05 | |
| *** sandywalsh_ has joined #openstack-dev | 13:07 | |
| *** sandywalsh has quit IRC | 13:08 | |
| *** roge has joined #openstack-dev | 13:09 | |
| *** jaypipes has joined #openstack-dev | 13:13 | |
| *** eglynn__ has quit IRC | 13:21 | |
| *** eglynn__ has joined #openstack-dev | 13:21 | |
| *** mattray has joined #openstack-dev | 13:22 | |
| *** markmc has quit IRC | 13:22 | |
| *** markmc has joined #openstack-dev | 13:27 | |
| *** eglynn__ has quit IRC | 13:31 | |
| *** dhellmann_ has joined #openstack-dev | 13:31 | |
| *** eglynn has joined #openstack-dev | 13:31 | |
| *** dhellmann has quit IRC | 13:33 | |
| *** dhellmann_ is now known as dhellmann | 13:33 | |
| *** dhellmann has quit IRC | 13:34 | |
| *** AlanClark has joined #openstack-dev | 13:36 | |
| *** markmc has quit IRC | 13:38 | |
| *** derekh has quit IRC | 13:39 | |
| *** dtroyer_zzz is now known as dtroyer | 13:43 | |
| notmyname | mtaylor: ping | 13:46 |
| notmyname | mtaylor: I only seem to get emails from jenkins when I was involved in the review. is it possible to always get merge emails from jenkins? | 13:47 |
| *** markmcclain has quit IRC | 13:52 | |
| eneabio | question: how I can do a rpc.call in least_cost.py to get _get_additional_capabilities from hosts? | 13:58 |
| clarkb | notmyname: under your account settings there are per project email options | 13:58 |
| clarkb | i think one of the options should do what you want but i dont have them in front of me | 13:59 |
| notmyname | clarkb: thanks. I see new changes, all comments, and submitted changes | 13:59 |
| davidkranz | jaypipes: I posted some small changes in reponse to your comments. I have to leave at 3 so if you want anything else let me now a bit before then. I will push it before I leave. | 14:00 |
| notmyname | clarkb: I'd imagine what I'm looking for is in jenkins, not gerrit | 14:00 |
| davidkranz | jaypipes: https://review.openstack.org/#/c/6991/ | 14:00 |
| *** armax has joined #openstack-dev | 14:04 | |
| *** sacharya has joined #openstack-dev | 14:04 | |
| *** armax has quit IRC | 14:04 | |
| *** armax has joined #openstack-dev | 14:05 | |
| *** eglynn has quit IRC | 14:06 | |
| *** belliott_ is now known as belliott | 14:06 | |
| *** eglynn has joined #openstack-dev | 14:06 | |
| *** kbringard has joined #openstack-dev | 14:09 | |
| *** Gordonz has joined #openstack-dev | 14:13 | |
| *** mancdaz has quit IRC | 14:14 | |
| *** mancdaz has joined #openstack-dev | 14:14 | |
| dolphm | is it possible to retrigger smokestack jobs without submitting a new patchset? | 14:16 |
| dolphm | (concerning https://review.openstack.org/#/c/6216/ ) | 14:16 |
| *** alaski has quit IRC | 14:16 | |
| *** alaski has joined #openstack-dev | 14:16 | |
| *** armax has quit IRC | 14:19 | |
| *** armax has joined #openstack-dev | 14:19 | |
| russellb | dolphm: it is if you have an account, i'll try to do it for you | 14:21 |
| *** alaski has quit IRC | 14:21 | |
| russellb | dolphm: done | 14:22 |
| *** markmcclain has joined #openstack-dev | 14:24 | |
| *** hub_cap has quit IRC | 14:24 | |
| dprince | dolphm: Is that something you would fine useful? | 14:29 |
| dprince | dolphm: retriggering SmokeStack? | 14:29 |
| dprince | dolphm: I'm happy to give you an account to the webUI if you'd like to re-trigger things... | 14:30 |
| dprince | russellb: If you do re-trigger to run the tests it won't repost results for the same patchset since results for that Git Hash have already been reported. The reporting is all based on the Git Hash's. | 14:31 |
| dprince | russellb: You can however just re-run the tests again and view them in the SmokeStack UI, etc. | 14:31 |
| dprince | russellb: you know this ;) | 14:31 |
| russellb | ah, ok, thanks for that info | 14:31 |
| dprince | mayby not the Git hash bits... | 14:32 |
| russellb | yep, it's super convenient for smoke testing as i work on a patch set ... | 14:32 |
| russellb | the re-run and view via UI part yes :) | 14:32 |
| *** armax has left #openstack-dev | 14:35 | |
| *** ayoung has joined #openstack-dev | 14:41 | |
| *** sacharya has quit IRC | 14:42 | |
| *** alaski has joined #openstack-dev | 14:43 | |
| *** markmc has joined #openstack-dev | 14:44 | |
| mtaylor | notmyname: hrm. no, actually the email should be coming from gerrit ... it's the one who does the merge | 14:44 |
| mtaylor | notmyname: but I'm not sure if the event you're talking about has an email setting associated with it (although it should) | 14:44 |
| mtaylor | notmyname: let us look in to it | 14:44 |
| notmyname | mtaylor: ok, cool. thanks | 14:44 |
| jeblair | notmyname: is 'submitted changes' checked? | 14:46 |
| jeblair | notmyname: 'submitted' is gerrit-speak for merged | 14:46 |
| notmyname | jeblair: no. ok. I was thinking that submitted was any new patch set for an existing change (to pair with "new changes") | 14:47 |
| notmyname | jeblair: ok, I've set that. I'll see what happens. thanks | 14:48 |
| mtaylor | jeblair: ah, good point. I will potentially close the bug I just filed | 14:49 |
| *** dtroyer is now known as dtroyer_zzz | 14:49 | |
| *** milner has quit IRC | 14:50 | |
| *** dachary has quit IRC | 14:52 | |
| dolphm | russellb: thanks | 14:56 |
| dolphm | dprince: well, this is the first potential false-negative i've seen from smokestack, so maybe maybe not :) | 14:57 |
| *** vincentricci has joined #openstack-dev | 15:00 | |
| *** hub_cap has joined #openstack-dev | 15:00 | |
| hub_cap | vishy: Hey man, ive got a question about adding a feature to novaclient, who should i ping? | 15:01 |
| nvez | os-simple-tenant-usage doesn't give any bandwidth information, how does on go to get that information? | 15:02 |
| hub_cap | vishy: Basically, i want to be able to bypass the auth & service catalog lookup, and provide my own token and compute or volume url. the glanceclient already does this and its very convenient to be able to override this to talk to a local non service catalog'd nova install | 15:02 |
| *** dtroyer_zzz is now known as dtroyer | 15:03 | |
| *** rnirmal has joined #openstack-dev | 15:06 | |
| *** anderstj has joined #openstack-dev | 15:08 | |
| *** reidrac has quit IRC | 15:08 | |
| *** anderstj has quit IRC | 15:08 | |
| *** sacharya has joined #openstack-dev | 15:12 | |
| *** dhellmann has joined #openstack-dev | 15:13 | |
| jaypipes | davidkranz: rock on. getting to the review now. | 15:14 |
| jaypipes | davidkranz: bought a new driver yesterday and the links are calling me this afternoon ;) | 15:15 |
| davidkranz | jaypipes: Excellent. I haven't played since high school but getting back is probably in my future... | 15:15 |
| jaypipes | davidkranz: oh, don't get me wrong. I still totally suck :) But at least now I enjoy the game a bit | 15:16 |
| davidkranz | jaypipes: :) | 15:16 |
| devananda | hub_cap: g'morning. got a few min to chat about kernels? | 15:18 |
| hub_cap | hit me devananda | 15:18 |
| hub_cap | lets mov eit to #openstack-infra | 15:18 |
| *** littleidea has joined #openstack-dev | 15:19 | |
| jaypipes | davidkranz: real quick ... go to https://review.openstack.org/#/c/6991/2/stress/state.py | 15:19 |
| davidkranz | jaypipes: I'm there. | 15:19 |
| jaypipes | davidkranz: is there a reason that server_id isn't passed in the constructor for ServerAssociatedState along with resource_id? | 15:20 |
| davidkranz | jaypipes: Because it is unknown when the object is created. The object will get mapped and remapped to servers in its lifetime. | 15:21 |
| *** maoy has joined #openstack-dev | 15:21 | |
| maoy | jerdfelt: hi are you around? | 15:21 |
| jaypipes | davidkranz: gotcha, k, that's what I thought. just wanted to make sure that was the case. | 15:21 |
| *** Mandell has joined #openstack-dev | 15:24 | |
| nvez | Do I have to subscribe using AMQP to get bandwidth usage information? | 15:25 |
| maoy | jerdfelt: I'd like to chat with you about the tracking of RPC tasks work. | 15:26 |
| hub_cap | jaypipes: can i get the actual image back from python-glanceclient via some command? | 15:34 |
| *** eneabio has quit IRC | 15:36 | |
| jaypipes | hub_cap: soon, yes. | 15:36 |
| jaypipes | hub_cap: bcwaldon is working on that I believe | 15:36 |
| hub_cap | kk so not at present but soon :D | 15:36 |
| jaypipes | hub_cap: yuppers. | 15:37 |
| hub_cap | heart | 15:38 |
| bcwaldon | hub_cap: I would encourage you to implement it! | 15:38 |
| hub_cap | im sure u would! can i use netcat? | 15:38 |
| bcwaldon | hub_cap: can you describe how that would work? | 15:39 |
| hub_cap | i have no clue i was joking so u would say hell no | 15:39 |
| bcwaldon | thats what I thought | 15:39 |
| hub_cap | and i could be off that hookj | 15:39 |
| hub_cap | hehe | 15:39 |
| bcwaldon | I'll look at it today | 15:39 |
| hub_cap | ive got to fix something in novaclient | 15:39 |
| hub_cap | its not a big deal for me | 15:40 |
| bcwaldon | it should be super easy | 15:40 |
| hub_cap | dont rush to it for my sake | 15:40 |
| bcwaldon | 20 minutes | 15:40 |
| bcwaldon | you'll have it | 15:40 |
| hub_cap | i dont really need it i was just wondering trying to diagnose an issue | 15:40 |
| hub_cap | wow udaman bcwaldon | 15:40 |
| nvez | hub_cap: I see you done some work on billing-related openstack stuff? Do you know if I can request a "compute.instance.exists" or I just have to sit and wait for it? If I wait for it, how can I do my testing and force nova to send that event? | 15:42 |
| hub_cap | nvez: i have not actually done work myself on the notifications or billing. i have a team mate cp16net that has. I believe (cp16net can correct me) that u subscribe to a q | 15:43 |
| hub_cap | cp16net: fill in the blanks | 15:43 |
| *** jdurgin has joined #openstack-dev | 15:43 | |
| cp16net | nvez: yeah so in order to get the exists events you have to cron a script called instance-usage-audit | 15:44 |
| cp16net | nvez: i think this defaults to the last month | 15:44 |
| nvez | I see, so if I run instance-usage-audit -- itll calculate and send an instance.exists event? | 15:44 |
| cp16net | nvez: if you set the flag in your nova.conf instance_usage_audit_period you can set it to hour/day/month/year | 15:45 |
| cp16net | but this is always the previous period | 15:45 |
| cp16net | correct | 15:45 |
| hub_cap | ya nvez what cp16net said :P | 15:46 |
| nvez | cp16net: This might sound really silly, but I'm still figuring out AMQP, if the daemon waiting for instance.exists messages is down, will I miss the information? | 15:46 |
| hub_cap | nvez: the messages go to the Q | 15:46 |
| hub_cap | and wait for you to pluck them off | 15:46 |
| nvez | ah i see | 15:46 |
| hub_cap | so they will reside there until your daemon/process/magic broomstick pulls them off | 15:46 |
| cp16net | thats if you are using the rabbit_notifier | 15:46 |
| hub_cap | tru | 15:47 |
| nvez | non-nova related question but when i pull them from the queue do they automatically get removed from teh queue, or do I specifically say "remove", obviously id want to "remove" after I successfully added all the data | 15:47 |
| *** danwent has joined #openstack-dev | 15:47 | |
| cp16net | nvez: i think that depends on the client i believe but i think once you fetch a message it comes off the queue | 15:50 |
| cp16net | i think its because the queue is declared as a topic queue. | 15:50 |
| *** dachary has joined #openstack-dev | 15:51 | |
| *** dachary has joined #openstack-dev | 15:51 | |
| *** dachary has quit IRC | 15:52 | |
| *** dachary1 has joined #openstack-dev | 15:52 | |
| nvez | cp16net: ah i see, trying to call it however it keeps trying to run it for a month, where can I change the audit period if you know? | 15:53 |
| *** dachary has joined #openstack-dev | 15:54 | |
| nvez | source code has this: cfg.StrOpt('instance_usage_audit_period', .. and it's set to month by default, but running instance-usage-audit --instance-usage-audit-period day doesnt work, hmm | 15:55 |
| *** dachary1 has quit IRC | 15:57 | |
| nvez | okay, so its not a cli option, its a nova cfg option, my bad | 15:57 |
| *** nati has joined #openstack-dev | 15:57 | |
| *** markmc has quit IRC | 15:58 | |
| *** jdurgin has quit IRC | 15:58 | |
| *** dachary has quit IRC | 15:58 | |
| *** dhellmann has quit IRC | 15:59 | |
| *** epim has joined #openstack-dev | 16:00 | |
| *** markmcclain has quit IRC | 16:00 | |
| *** dachary has joined #openstack-dev | 16:02 | |
| *** alaski has quit IRC | 16:02 | |
| *** alaski has joined #openstack-dev | 16:03 | |
| *** Mandell has quit IRC | 16:04 | |
| *** mnewby has joined #openstack-dev | 16:05 | |
| *** andresambrois has joined #openstack-dev | 16:07 | |
| *** maplebed has joined #openstack-dev | 16:08 | |
| *** issackelly has joined #openstack-dev | 16:12 | |
| cp16net | nvez: yes its picked up from the nova.conf options | 16:12 |
| nvez | sweet, alright, ill start working a bit on this hopefully ill have some sort of result by today! | 16:13 |
| cp16net | nice | 16:13 |
| nvez | oh small question cp16net -- does the instance exists give info for servers that were terminated in that period? | 16:13 |
| nvez | aka server running for 3 hours only and got terminated after | 16:13 |
| cp16net | no its only what exists | 16:13 |
| *** mnot has quit IRC | 16:14 | |
| nvez | so basically if someone starts a server everyday at 1am and shuts it down at 11pm and i do daily audit reports, i cant see the bandwidth usage for it? | 16:14 |
| cp16net | when you terminate/resize an instance a new exists event is created at that time with the audit-period-beginning and audit-peroid-ending with the times | 16:14 |
| nvez | ah gotcha | 16:15 |
| *** adam_g has joined #openstack-dev | 16:15 | |
| *** eglynn has quit IRC | 16:16 | |
| nvez | cp16net: so resize/terminate/create sends an exists as well, hmm, i would only have to use the terminate to get bandwidth usage however, because if i use resize i think ill have duplicate data | 16:16 |
| *** torgomatic has joined #openstack-dev | 16:16 | |
| cp16net | oh nevermind it looks like terminate instance doesnt send an exists | 16:16 |
| cp16net | only resize and rebuild instance | 16:17 |
| bcwaldon | hub_cap: well I've had it done for a while, it just took forever to get a working dev env up | 16:17 |
| hub_cap | hehe, that makes sense then for 20m | 16:17 |
| cp16net | nvez: i've been trying to figure out when all these events occur as i am working on adding volume notifications that are similar to the instance | 16:18 |
| nvez | cp16net: hmm, okay, well that leaves me with a possibility that i cant have precise bw monitoring | 16:18 |
| nvez | im curious if the delete sends a bandwidth used .. maybe.. | 16:19 |
| hub_cap | bcwaldon: quick Q for you, v1/images/image_id is returning a empty reply.. am i doing something drastically wrong? | 16:19 |
| bcwaldon | hub_cap: look in your headers | 16:19 |
| bcwaldon | hub_cap: and do you have data up there? | 16:19 |
| hub_cap | images/image GET will return the image, rigth? | 16:19 |
| bcwaldon | hub_cap: yes | 16:19 |
| hub_cap | kk, let me verify | 16:19 |
| bcwaldon | hub_cap: assuming you've put it up | 16:20 |
| hub_cap | its possible i didnt put the image properly up | 16:20 |
| hub_cap | ya | 16:20 |
| hub_cap | i have a 63 meg gzip raw/bare image in /var/lib/glance/images/id | 16:20 |
| bcwaldon | hub_cap: and headers dont complain? | 16:20 |
| bcwaldon | hub_cap: does it have a size? | 16:20 |
| hub_cap | sec let me list | 16:21 |
| hub_cap | bcwaldon: http://paste.openstack.org/show/16973/ | 16:21 |
| cp16net | nvez: after looking at the code it looks like the exists event should generate an event for each instance that was deleted in the time perioid and give the bandwidth | 16:21 |
| bcwaldon | hub_cap: yep, something wrong | 16:21 |
| *** issackelly has quit IRC | 16:21 | |
| hub_cap | do i have to provide the tenant header in the images/id call? | 16:21 |
| *** issackelly has joined #openstack-dev | 16:22 | |
| bcwaldon | hub_cap: no, just the auth token | 16:22 |
| hub_cap | ya thats what im putting. did u see something wrong from the listing i displayed? | 16:22 |
| *** Stackops-Jorge has quit IRC | 16:22 | |
| bcwaldon | hub_cap: nope | 16:22 |
| hub_cap | durn, time to grok the logs :D | 16:22 |
| nvez | cp16net: yep, you're right --- in notify_usage_exists: admin_context = nova.context.get_admin_context(read_deleted='yes') | 16:22 |
| *** thingee has quit IRC | 16:23 | |
| cp16net | yes :) | 16:23 |
| hub_cap | i do see an error bcwaldon http://paste.openstack.org/show/16974/ | 16:23 |
| nvez | and it does return a deleted_at as well in the usage | 16:23 |
| hub_cap | ps im not using devstack or anything to set up glance, so its VERY possible i borked something config related | 16:24 |
| nvez | woo, sounds good, time to write up a small client (in the language everyone hates, java :D) :p | 16:24 |
| bcwaldon | hub_cap: hmm, thats pretty odd | 16:24 |
| bcwaldon | hub_cap: can I see your glance-api.conf and glance-api-paste.ini? | 16:24 |
| hub_cap | bcwaldon: im good at screwing things up | 16:24 |
| hub_cap | sure sec bcwaldon | 16:24 |
| hub_cap | bcwaldon: u want registry confs too? | 16:26 |
| bcwaldon | hub_cap: not yet | 16:26 |
| hub_cap | im curling to 9292 not 9191, is that my problem? | 16:26 |
| hub_cap | ok | 16:26 |
| bcwaldon | 9292 is correct | 16:26 |
| *** spiffxp has joined #openstack-dev | 16:29 | |
| *** salgado is now known as salgado-lunch | 16:29 | |
| *** jdg has joined #openstack-dev | 16:31 | |
| *** jemartin has joined #openstack-dev | 16:33 | |
| *** rods has joined #openstack-dev | 16:35 | |
| *** danwent has quit IRC | 16:36 | |
| *** tryggvil_ has quit IRC | 16:37 | |
| *** tryggvil has joined #openstack-dev | 16:38 | |
| *** dprince has left #openstack-dev | 16:43 | |
| *** dprince has quit IRC | 16:43 | |
| *** eglynn has joined #openstack-dev | 16:44 | |
| jdg | rnirmal: ping | 16:46 |
| *** bsza1 has joined #openstack-dev | 16:46 | |
| *** bsza has quit IRC | 16:47 | |
| *** jdg is now known as jgriff | 16:49 | |
| rnirmal | jgriff: | 16:50 |
| jgriff | rnirmal: Just checking to see if you got the notification email? | 16:50 |
| rnirmal | let me check | 16:50 |
| jgriff | rnirmal: Or if things are still hosed | 16:50 |
| rnirmal | jgriff: nope... I even reset everything yesterday | 16:51 |
| *** Mandell_ has joined #openstack-dev | 16:51 | |
| *** nati has quit IRC | 16:51 | |
| jgriff | rnirmal: dang it... :( | 16:51 |
| jgriff | rnirmal: just curious, have you checked your settings in gerrit? | 16:52 |
| rnirmal | yeah I did check... like I said reset everything. | 16:53 |
| jgriff | bummer... that exhausts my helpfulness | 16:54 |
| rnirmal | I'll figure it out someday.. :) | 16:54 |
| jgriff | LOL | 16:54 |
| rnirmal | jgriff: I'll take a look at the latest patch | 16:54 |
| jgriff | rnirmal: thanks :) | 16:55 |
| rnirmal | is there any way to look at all the diffs on a single tab..instead of the multiple tabs | 16:55 |
| jgriff | rnirmal: you mean all files in one? No I don't believe so... the way to do that would be something like pull the review and do a git diff | 16:55 |
| jgriff | rnirmal: but I could be wrong | 16:56 |
| rnirmal | yeah that's what I did the last time around | 16:56 |
| rnirmal | didn't find anything in the ui | 16:56 |
| jgriff | rnirmal: there are only three files that changed | 16:56 |
| rnirmal | I'll look thru them | 16:57 |
| jgriff | rnirmal: openstack/volume/volumes.py, db/api.py and cinder/exceptions.py | 16:57 |
| rnirmal | jgriff: thanks | 16:58 |
| *** anderstj has joined #openstack-dev | 16:58 | |
| *** koolhead17 has quit IRC | 16:58 | |
| *** mszilagyi has joined #openstack-dev | 16:58 | |
| jaypipes | hub_cap: speak of the devil. https://review.openstack.org/#/c/7352/ | 17:02 |
| *** eglynn has quit IRC | 17:04 | |
| *** apevec has quit IRC | 17:06 | |
| *** harlowja has joined #openstack-dev | 17:06 | |
| *** vladimir3p has joined #openstack-dev | 17:07 | |
| harlowja | bcwaldon: in regard to that duplication registering of that opt, i guess moving the create_stores call out of this image api would solve it, although it is odd that the constructor would be called twice... thats really the only way that can happen | 17:08 |
| *** vladimir3p has quit IRC | 17:08 | |
| *** koolhead17 has joined #openstack-dev | 17:11 | |
| bcwaldon | harlowja: or duplicate calls to create_stores can just be ignored | 17:12 |
| bcwaldon | harlowja: not sure what the *best* way is here | 17:12 |
| harlowja | ya, or that, i think i see what the issue is | 17:12 |
| harlowja | ImageDataController in v2/ calls create_stores | 17:12 |
| bcwaldon | harlowja: since we are going to import glance.store separately in the v2 and v1 api | 17:12 |
| harlowja | Controller in v1/ calls the same | 17:12 |
| harlowja | ya, hmm | 17:12 |
| harlowja | bcwaldon: should the create stores happen at a higher level, not in the apis? | 17:15 |
| *** bhuvan_ has joined #openstack-dev | 17:15 | |
| *** markmcclain has joined #openstack-dev | 17:16 | |
| bcwaldon | harlowja: there is no other component, so I don't think so | 17:16 |
| bcwaldon | you could force it to happen in paste then pass it in, but that would be a massive hack | 17:17 |
| *** dhellmann has joined #openstack-dev | 17:18 | |
| harlowja | hmmm, there is a method like setup_logging, in glance/common/config, maybe there should be an equiv called setup_stores, that way it happens even before wsgi/paste | 17:20 |
| harlowja | except for this in the v2, stuff, self.store_api = store_api or glance.store | 17:21 |
| harlowja | self.store_api.create_stores(conf), where it seems like it changes what store_api its constructing it from | 17:21 |
| bcwaldon | harlowja: ok, so we could just call a common setup method in bin/glance-api | 17:24 |
| bcwaldon | harlowja: then never call create_stores down below | 17:24 |
| bcwaldon | that makes more sense | 17:24 |
| *** mnewby has quit IRC | 17:24 | |
| harlowja | right, except for the v2 image_data class | 17:24 |
| bcwaldon | harlowja: you can remove the create_stores call from that | 17:24 |
| harlowja | kk | 17:25 |
| harlowja | let me see what will happen with this | 17:25 |
| bcwaldon | kk | 17:25 |
| bcwaldon | harlowja: I'm going afk for a little bit | 17:25 |
| harlowja | k | 17:25 |
| *** asisin has joined #openstack-dev | 17:25 | |
| *** mnewby has joined #openstack-dev | 17:26 | |
| *** mnewby has joined #openstack-dev | 17:26 | |
| *** salgado-lunch is now known as salgado | 17:26 | |
| *** darraghb has quit IRC | 17:30 | |
| *** armax has joined #openstack-dev | 17:30 | |
| *** armax has left #openstack-dev | 17:30 | |
| *** danpb has quit IRC | 17:36 | |
| *** davidha has quit IRC | 17:37 | |
| *** Ryan_Lane has joined #openstack-dev | 17:42 | |
| *** mcclurmc_ has joined #openstack-dev | 17:45 | |
| *** vincentricci has quit IRC | 17:46 | |
| *** johnpostlethwait has joined #openstack-dev | 17:56 | |
| *** edygarcia has quit IRC | 17:58 | |
| *** edygarcia has joined #openstack-dev | 17:59 | |
| *** edygarcia_ has joined #openstack-dev | 18:00 | |
| *** epim_ has joined #openstack-dev | 18:00 | |
| *** eglynn has joined #openstack-dev | 18:01 | |
| *** epim__ has joined #openstack-dev | 18:01 | |
| *** edygarcia has quit IRC | 18:03 | |
| *** edygarcia_ is now known as edygarcia | 18:03 | |
| *** epim has quit IRC | 18:04 | |
| *** epim__ is now known as epim | 18:04 | |
| *** hub_cap has quit IRC | 18:05 | |
| *** epim_ has quit IRC | 18:05 | |
| *** jemartin has quit IRC | 18:05 | |
| *** tryggvil has quit IRC | 18:13 | |
| *** markmcclain has quit IRC | 18:18 | |
| *** dhellmann has quit IRC | 18:18 | |
| *** markmcclain has joined #openstack-dev | 18:18 | |
| *** dhellmann has joined #openstack-dev | 18:18 | |
| *** lloydde has joined #openstack-dev | 18:22 | |
| *** dachary1 has joined #openstack-dev | 18:23 | |
| *** dachary has quit IRC | 18:24 | |
| *** AlanClark has quit IRC | 18:26 | |
| *** mattray has quit IRC | 18:27 | |
| *** Shrews has quit IRC | 18:31 | |
| *** edygarcia has quit IRC | 18:31 | |
| *** mrunge has joined #openstack-dev | 18:34 | |
| *** bhuvan_ has quit IRC | 18:35 | |
| *** lloydde has quit IRC | 18:38 | |
| cp16net | pvo: i have a question on how to get a service_type into the notifications that are published? One thing i thought about was getting the service from the context or auth information and passing that along or maybe using a flag. The issue is because other services will utilize a nova installation and need to identify the service that is requesting to create a new resource and associate that resource with the service that called l | 18:40 |
| cp16net | reddwarf in my case. | 18:40 |
| *** belliott has quit IRC | 18:42 | |
| jeblair | jaypipes: gate-tempest-devstack-vm is enabled again | 18:43 |
| jeblair | it should start voting on changes after run #602 | 18:43 |
| *** koolhead17 has quit IRC | 18:43 | |
| openstackgerrit | Verification of a change to openstack/nova failed: Create an internal key pair API. https://review.openstack.org/7342 | 18:47 |
| nvez | cp16net: i'm not having good luck right now.. I am binding my client to the "notifications" queue under the "nova" exchange and starting a queueingconsumer (java), everything gets created but im not getting any messages.. | 18:49 |
| nvez | I'm running instance-usage-audit and I don't see much, any idea how I can check and see if the msg is being sent? | 18:50 |
| nvez | There is data being sent cause it does say "found 2 instances" | 18:50 |
| *** kbringard has quit IRC | 18:51 | |
| *** kbringard has joined #openstack-dev | 18:52 | |
| cp16net | nvez: are you using the notification_driver='rabbit_notifier'? | 18:54 |
| cp16net | in your nova.cong | 18:54 |
| cp16net | f* | 18:54 |
| nvez | cp16net: not set, using devstack, let me see what the default is | 18:55 |
| nvez | cp16net: nova.notifier.no_op_notifier is the default, woops. | 18:55 |
| cp16net | nvez: so that notify method does nothing | 18:56 |
| cp16net | :) | 18:57 |
| nvez | THERE WE GO! | 18:57 |
| nvez | two messages for 2 instances, man, thank you so much cp16net -- im new to amqp so yeah | 18:58 |
| cp16net | nice | 18:58 |
| nvez | guess i had everything working, just didnt think of the notification driver, thanks again | 18:58 |
| cp16net | np | 18:58 |
| cp16net | vishy: would you have any input on the question i asked pvo? ^^ | 18:59 |
| *** markmcclain has quit IRC | 18:59 | |
| *** markmcclain has joined #openstack-dev | 18:59 | |
| *** adjohn has joined #openstack-dev | 19:00 | |
| nvez | cp16net: just to confirm as well that i did get 2 events, one for a deleted server and for an active one | 19:00 |
| cp16net | awesome | 19:01 |
| nvez | bandwidth key is empty tho, so hmm | 19:01 |
| cp16net | nvez: there is also a log_notifier that writes the events to the log file | 19:01 |
| *** zaitcev has joined #openstack-dev | 19:02 | |
| nvez | cp16net: ill check out all the other notifiers and also check out why the bandwidth has nothing inside of it, it should have a network assigned to it, maybe if 0 bw used then it wont do anything | 19:02 |
| cp16net | nvez: yeah the logic on how it gets the bandwidth is in the nova/compute/utils.py | 19:04 |
| *** gyee has joined #openstack-dev | 19:04 | |
| cp16net | nvez: looks like its just getting it from the db and i dont know what would write that to the db in the first place unless its some other type of agent | 19:05 |
| nvez | hmm, im gonna check it out and figure it out | 19:06 |
| *** lloydde has joined #openstack-dev | 19:07 | |
| zykes- | what does people use to deploy openstack in places ? | 19:09 |
| *** jakedahn is now known as jakedahn_zz | 19:10 | |
| cp16net | nvez: looks like only the xen driver supports bandwidth | 19:13 |
| *** lorin1 has joined #openstack-dev | 19:15 | |
| nvez | cp16net: xen or xenserver? | 19:17 |
| cp16net | i found it in the xenapi.py i dont really know the difference | 19:18 |
| cp16net | i've been using kvm and openvz | 19:18 |
| nvez | i think xenapi is xenserver/xcp | 19:18 |
| *** lifeless has quit IRC | 19:18 | |
| *** fc__ has quit IRC | 19:23 | |
| *** andrewbogott_ has quit IRC | 19:24 | |
| *** andrewbogott_ has joined #openstack-dev | 19:25 | |
| nvez | you are correct cp16net -- looks like only xenapi has get_all_bw_usage implemented (which is called to get the bw usage) | 19:25 |
| nvez | and that is actually the xenserver/xcp version | 19:25 |
| *** mattray has joined #openstack-dev | 19:29 | |
| nvez | is it acceptable to say that as of right now xenserver is the best implemented openstack driver? | 19:30 |
| russellb | the libvirt (kvm) driver is used heavily, as well. | 19:31 |
| epim | Anyone aware of any blueprints to make openstack try to fping an IP and make sure it's alive before allocating it? | 19:32 |
| epim | make sure it's not in use, I should say | 19:32 |
| zykes- | epim: isn' | 19:35 |
| zykes- | isn't that what the db is for ? | 19:35 |
| epim | Well, if my vlan is used for more than just VMs, or if someone swipes an IP from a dedicated vlan, nova will try to blindly allocate that IP | 19:36 |
| *** hub_cap has joined #openstack-dev | 19:37 | |
| hub_cap | jaypipes: ya bcwaldon showed me that :D | 19:37 |
| *** mcclurmc_ has quit IRC | 19:41 | |
| lorin1 | Are the "[no]foo" style config options still valid? Or do we just do "foo=false" now? | 19:42 |
| *** lloydde has quit IRC | 19:43 | |
| *** lloydde has joined #openstack-dev | 19:44 | |
| *** andrewbogott_ has quit IRC | 19:44 | |
| *** andrewbogott_ has joined #openstack-dev | 19:44 | |
| nvez | So uh, why pick XenServer or XCP, I have a hard time deciding and I'm not sure which one would be better/worse, there doesnt seem to be an openstack comparision | 19:44 |
| nvez | I understand they're supposed to be "the same thing" but anyone had to make that choice and found a reason to pick one over the other for an OpenStack deployment? | 19:45 |
| epim | nvez: in theory XCP is just Xen with fancy management tools. Most, if not all of which are redundant to openstack | 19:47 |
| epim | actually, sorry, I was thinking Xenserver. XCP is mostly just xen-in-a-box. | 19:47 |
| nvez | epim: I'm aware of that, my conern is XenServer vs XCP (open source version of Citrix XenServer).. they seem to share a similar featureset | 19:47 |
| *** lloydde has quit IRC | 19:48 | |
| *** jgriff has quit IRC | 19:55 | |
| *** dolphm has quit IRC | 19:58 | |
| *** davidha has joined #openstack-dev | 19:59 | |
| *** jakedahn_zz is now known as jakedahn | 19:59 | |
| *** dachary1 is now known as dachary | 20:04 | |
| *** anderstj has quit IRC | 20:08 | |
| *** anderstj_ has joined #openstack-dev | 20:08 | |
| nvez | cp16net: sweet code review ;) hopefully it makes it! | 20:15 |
| cp16net | yeah that should add create/delete/exists events for volumes | 20:16 |
| *** rnirmal has quit IRC | 20:18 | |
| *** markvoelker1 has quit IRC | 20:18 | |
| *** anniec72 has quit IRC | 20:20 | |
| *** maoy has quit IRC | 20:20 | |
| *** lifeless has joined #openstack-dev | 20:22 | |
| *** littleidea has quit IRC | 20:22 | |
| *** littleidea has joined #openstack-dev | 20:23 | |
| *** jgriff has joined #openstack-dev | 20:25 | |
| *** anniec72 has joined #openstack-dev | 20:26 | |
| *** alaski has quit IRC | 20:30 | |
| *** vincentricci has joined #openstack-dev | 20:30 | |
| *** novas0x2a|laptop has joined #openstack-dev | 20:32 | |
| *** anniec72 has quit IRC | 20:33 | |
| *** jemartin has joined #openstack-dev | 20:38 | |
| *** Gordonz has quit IRC | 20:40 | |
| *** johnpostlethwa-1 has joined #openstack-dev | 20:41 | |
| *** johnpostlethwait has quit IRC | 20:41 | |
| *** johnpostlethwa-1 has quit IRC | 20:43 | |
| *** johnpostlethwait has joined #openstack-dev | 20:43 | |
| *** ayoung_ has joined #openstack-dev | 20:47 | |
| bcwaldon | jeblair: around? | 20:51 |
| jeblair | bcwaldon: t | 20:51 |
| bcwaldon | jeblair: why do we automagically set bugs to fixed by a committer rather than author? | 20:52 |
| bcwaldon | jeblair: its frustrating when I fix up someones patch, become the committer, then it merges | 20:52 |
| bcwaldon | jeblair: I didnt really fix that bug | 20:52 |
| jeblair | bcwaldon: because we didn't write the script right? :) | 20:52 |
| bcwaldon | jeblair: we didnt? | 20:53 |
| jeblair | bcwaldon: sorry, my way of agreeing with you. :) | 20:53 |
| bcwaldon | jeblair: ok, which script is it? | 20:53 |
| jeblair | https://github.com/openstack/openstack-ci-puppet/blob/master/modules/gerrit/files/scripts/update_bug.py | 20:54 |
| *** dhellmann has quit IRC | 20:55 | |
| jeblair | bcwaldon: it may be that only the uploader, not the author of the commit, is a hook argument; it may need a bit more code to go grab that from the git log | 20:56 |
| *** markmcclain has quit IRC | 20:57 | |
| bcwaldon | jeblair: ok, I might look into fixing that little guy | 20:57 |
| jeblair | bcwaldon: awesome! | 20:57 |
| *** mcclurmc_ has joined #openstack-dev | 21:03 | |
| hub_cap | pvo: u around? | 21:04 |
| *** lts has quit IRC | 21:13 | |
| *** mcclurmc_ has quit IRC | 21:16 | |
| *** ayoung_ is now known as ayoung_afk | 21:16 | |
| *** lloydde has joined #openstack-dev | 21:21 | |
| *** dachary has quit IRC | 21:25 | |
| *** lloydde has quit IRC | 21:25 | |
| *** dachary has joined #openstack-dev | 21:25 | |
| *** lloydde has joined #openstack-dev | 21:26 | |
| *** hub_cap has quit IRC | 21:29 | |
| *** lorin1 has quit IRC | 21:31 | |
| *** jakedahn is now known as jakedahn_zz | 21:31 | |
| *** mcclurmc_ has joined #openstack-dev | 21:38 | |
| *** anderstj has joined #openstack-dev | 21:51 | |
| *** anderstj_ has quit IRC | 21:51 | |
| *** hattwick has quit IRC | 21:57 | |
| *** hattwick has joined #openstack-dev | 22:00 | |
| *** kbringard has quit IRC | 22:01 | |
| *** mrunge has quit IRC | 22:08 | |
| *** andrewsben is now known as andrewsben_zz | 22:10 | |
| mikal | vishy: you around? | 22:11 |
| *** sacharya has quit IRC | 22:14 | |
| *** koolhead17 has joined #openstack-dev | 22:14 | |
| *** gyee has quit IRC | 22:15 | |
| *** pixelbeat has quit IRC | 22:17 | |
| *** dachary has quit IRC | 22:29 | |
| *** dachary has joined #openstack-dev | 22:29 | |
| *** bsza1 has quit IRC | 22:32 | |
| *** jakedahn_zz is now known as jakedahn | 22:35 | |
| *** spiffxp has quit IRC | 22:38 | |
| *** mattray has quit IRC | 22:46 | |
| notmyname | mtaylor: what's the best way to see the differences in two patch sets of a large merge if there have been commits between the patch sets? | 22:46 |
| harlowja | bcwaldon: whenever u get a chance, fixed the glance stuff.... | 22:49 |
| bcwaldon | harlowja: kk, sometime today for sure | 22:49 |
| zykes- | notmyname: what alternatives are there for servers with 12+ drives ? | 22:50 |
| notmyname | zykes-: how do you mean? are you looking for hardware recommendations? | 22:52 |
| *** spiffxp has joined #openstack-dev | 22:52 | |
| zykes- | yes | 22:52 |
| zykes- | like enterprise ish | 22:53 |
| *** anniec72 has joined #openstack-dev | 22:54 | |
| zykes- | I think that the C* node listed in the reference arch is not so dense as it should be ? | 22:54 |
| notmyname | zykes-: I don't have any recommendations for brands. but for example, there is http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100008044+600029570+600010490&QksAutoSuggestion=&ShowDeactivatedMark=False&Configurator=&IsNodeId=1&Subcategory=412&description=&hisInDesc=&Ntk=&CFG=&SpeTabStoreType=&AdvancedSearch=1&srchInDesc= | 22:54 |
| notmyname | change the number of drives you are looking for to find different stuff | 22:55 |
| notmyname | zykes-: one possible config is a smaller head unit and a JBOD or two. | 22:55 |
| zykes- | Ah, it sucks that Dell / HP isn't delivering stuff for it that's more dense | 22:55 |
| notmyname | wow. 72 drives in one 4U http://www.newegg.com/Product/Product.aspx?Item=N82E16811152212 | 22:56 |
| zykes- | Can't HP / Dell make such a box | 22:57 |
| notmyname | no idea. you'd have to ask them :-) | 22:57 |
| zykes- | grrrr | 22:57 |
| zykes- | the crappy with with SM is that big customers here don't want to rely on anything that isn't HP, IBM / Dell | 22:58 |
| notmyname | ya, I feel your pain on that | 22:58 |
| *** dtroyer is now known as dtroyer_zzz | 22:58 | |
| jgriff | zykes: I believe they have versions on part with what you're looking at but typically recommend going to an array or jbod as notmyname suggested | 22:58 |
| notmyname | you may try looking at the opencompute reference platforms. HP is big into that (not sure about dell), and they have dense storage stuff | 22:58 |
| jgriff | s/on part/on par/ | 22:58 |
| zykes- | notmyname: but no public products as of yet I get ? | 22:59 |
| notmyname | http://opencompute.org/projects/open-vault-storage/ | 22:59 |
| zykes- | jgriff: versions on part ? | 22:59 |
| notmyname | I saw some things last week at the opencompute summit that was help at the rackspace office | 23:00 |
| jgriff | zykes: I don't work at HP any more and not a sales guy.. you'd have to ask them :) | 23:00 |
| zykes- | jgriff: you worked at HPCS ? | 23:00 |
| jgriff | zykes: Nope, storage group... unified storage | 23:01 |
| notmyname | zykes-: you can probably get the cheapest $/GB storage with a head unit and JBODs. on the other hand, I've heard of large deployments doing all 12 dirve configs in a 1U. it comes down a lot to what your use case is and what kind of deals you can get from vendors | 23:01 |
| notmyname | zykes-: but I've definitely seen problems convincing traditional storage and infrastructure people that consumer drives are ok (instead of enterprise drives) | 23:02 |
| zykes- | notmyname: yeah especially enterprise customers | 23:03 |
| zykes- | :p | 23:03 |
| * jgriff big grin | 23:03 | |
| notmyname | reliable software allows for unreliable (ie cheap) hardware | 23:03 |
| zykes- | jgriff: I wonder what HP uses for their swift stuff | 23:04 |
| zykes- | would like to get ahold of that :p | 23:04 |
| *** kbringard has joined #openstack-dev | 23:04 | |
| jgriff | zykes: I think notmyname pretty much indicated what they're doing | 23:04 |
| *** jemartin has quit IRC | 23:04 | |
| *** kbringard has quit IRC | 23:05 | |
| notmyname | zykes-: if you ever see johnpur in IRC, ask him. I don't know if he can share, but he would know (or know who to talk to) | 23:05 |
| notmyname | jgriff: I have no idea what HP is using | 23:05 |
| zykes- | is he ever online ? | 23:05 |
| zykes- | I feel that noone from HP is ever here :p | 23:05 |
| *** dolphm has joined #openstack-dev | 23:05 | |
| notmyname | zykes-: he's on during the ppb meetings normally | 23:06 |
| jgriff | notmyname: Understand, I'm suggesting that you're strategy of scaling smaller (ie 12 drive units) or using jbods is "likely" accurate | 23:06 |
| jgriff | notmyname: I'm not claiming to have that information either... | 23:06 |
| notmyname | jgriff: oh. perhaps. I'm not convinced that's the best way to go for general use case storage (ie public cloud provider). public cloud is generally more interested in denser storage than that | 23:07 |
| jgriff | notmyname: Yeah, it's interesting the different use cases... all comes down to what they're targeting | 23:08 |
| notmyname | but really, it's up to each provider to evaluate their use case (that's my default non-answer answer) | 23:08 |
| jgriff | notmyname: :) | 23:08 |
| notmyname | zykes-: you may ask the wikipedia guys. they have a relatively small cluster, but they've been pretty open about all of the details | 23:08 |
| mtaylor | notmyname: hrm... | 23:08 |
| harlowja | someone from y! might be able to tell u also | 23:09 |
| harlowja | i can ask... | 23:09 |
| harlowja | we like da storage, lol | 23:09 |
| harlowja | epim: might have more details... | 23:09 |
| zykes- | harlowja: could you check ? | 23:09 |
| *** dhellmann has joined #openstack-dev | 23:09 | |
| * epim perks up | 23:09 | |
| harlowja | zykes-: ask epim whatever u want :-p | 23:10 |
| zykes- | are you allowed to share ? ;p | 23:10 |
| harlowja | epim: how much hardware info can we share ;) | 23:10 |
| epim | perhaps, what am I sharing? I'm backscroll deficient :D | 23:10 |
| zykes- | Swift specifications | 23:10 |
| harlowja | ah, nm that, haha | 23:10 |
| epim | mrmm, a bit I suppose. I was thinking about releasing the results of my hypervisor performance tests to the world. But I don't want to deal with the hate mail | 23:10 |
| harlowja | we aren't using swift ... (duplicate techs...) | 23:10 |
| epim | Yeah, what Josh said | 23:11 |
| zykes- | what's yahoo using openstack for + | 23:11 |
| epim | Stuff. | 23:11 |
| epim | heehehe | 23:11 |
| harlowja | lol | 23:11 |
| epim | jk | 23:11 |
| notmyname | maplebed: are you able to share wikipedia's swift hardware specs? | 23:11 |
| maplebed | yeah. | 23:11 |
| epim | We operate some of the largest and most power efficient datacenters int he world, Openstack will be another great way for us to leverage our awesome computing stack and become even more efficient | 23:12 |
| maplebed | hmm. I was hoping they'd be here: http://wikitech.wikimedia.org/view/Swift/Dev_Notes#Hardware | 23:12 |
| mtaylor | notmyname: that's a good question | 23:12 |
| epim | my god I swear that wasn't a practiced PR piece. | 23:12 |
| harlowja | lol | 23:13 |
| notmyname | epim: too late. you said "leverage" ;-) | 23:13 |
| *** koolhead17 has quit IRC | 23:13 | |
| epim | Ugh, I know.. too many meetings :( | 23:13 |
| harlowja | synergy | 23:13 |
| harlowja | ha | 23:13 |
| epim | at least I didn't say.. yes that. | 23:13 |
| maplebed | notmyname: one sec; I'll find the specifics. | 23:13 |
| notmyname | maplebed: zykes- is looking for info | 23:13 |
| zykes- | harlowja: you based in the US ? | 23:13 |
| harlowja | san jose, ca | 23:13 |
| zykes- | Ok | 23:14 |
| notmyname | mtaylor: I always manage to give you the best ones :-) | 23:14 |
| zykes- | I wonder when European companies will catch up with US ones :p | 23:14 |
| maplebed | cool. I'm not sure our specs are the best... our disks are pegged at 100% utilization pretty often. | 23:14 |
| maplebed | I wouldn't be surprised if, after doing some more analysis, we change stuff around. | 23:14 |
| notmyname | mtaylor: it may not be possible since the commit is amended and rebased. that info may simply be lost in the process | 23:14 |
| *** dtroyer_zzz is now known as dtroyer | 23:14 | |
| notmyname | zykes-: need more meetings and "synergies" in eastern europe? | 23:15 |
| zykes- | northern europe really | 23:16 |
| harlowja | ** do u need to leverage more synergistic meetings in northern europe? | 23:16 |
| *** lloydde has quit IRC | 23:16 | |
| *** ywu has joined #openstack-dev | 23:16 | |
| notmyname | zykes-: not sure where I got eastern from.... | 23:17 |
| *** dtroyer is now known as dtroyer_zzz | 23:17 | |
| *** pmezard has quit IRC | 23:18 | |
| zykes- | I don't think there's been like any even in Scandinavia yet | 23:18 |
| *** dolphm has quit IRC | 23:18 | |
| maplebed | notmyname, zykes-: storage: dell poweredge c2100 with 2 xeon E5645 2.4GHz chips, 48G ram and 12 2TB 7.2kRPM SAS disks. | 23:18 |
| *** rbasak has joined #openstack-dev | 23:18 | |
| zykes- | do you need that much ram for a storage node ?! | 23:19 |
| notmyname | do you need SAS drives? | 23:19 |
| zykes- | refernce architecture says like 12 gb | 23:19 |
| zykes- | on a 24 tb node | 23:20 |
| maplebed | our boxes don't use all of the ram, so we could probably cut back a bit. many of the disks peg 100% utilization (as reported by iostat). | 23:20 |
| zykes- | well, well, I need to go :) | 23:20 |
| zykes- | nn all | 23:20 |
| maplebed | pending data from some more stats around async pendings, I'm considering putting in 2 SSDs per storage node for containers, but I need evidence for that first. | 23:20 |
| notmyname | me too :-) | 23:20 |
| zykes- | Would be cool if there came some events soon though to say CopenHagen or StockHolm ;p | 23:21 |
| zykes- | Copenhagen / Stockholm sorry | 23:21 |
| notmyname | maplebed: yes, that is a wise choice. RAX dramatically improved performance doing something like that | 23:21 |
| maplebed | zykes-: stats on our utilization: http://ganglia.wikimedia.org/latest/?c=Swift%20pmtpa&m=load_one&r=day&s=by%20name&hc=4&mc=2 | 23:21 |
| *** anderstj has quit IRC | 23:40 | |
| nvez | zykes-: bit late but I've spoken with Dell DCS and you'll likely rarely see a really high density storage server for now, but the C6100's are freaking sexy for how much power they give in it | 23:44 |
| *** roge has quit IRC | 23:45 | |
| *** lloydde has joined #openstack-dev | 23:46 | |
| *** jdurgin has joined #openstack-dev | 23:47 | |
| *** epim has quit IRC | 23:53 | |
| *** Ryan_Lane has quit IRC | 23:54 | |
| *** Ryan_Lane has joined #openstack-dev | 23:54 | |
| *** novas0x2a|laptop has quit IRC | 23:56 | |
| *** jgriff has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
