Thursday, 2011-06-02

*** yamahata_lt has quit IRC00:00
*** heckj has quit IRC00:01
*** ohnoimdead has quit IRC00:02
*** RobertLaptop has joined #openstack-dev00:16
*** rnirmal has joined #openstack-dev00:29
*** mgius has quit IRC00:29
*** vladimir3p has quit IRC00:30
*** mgius has joined #openstack-dev00:31
*** yamahata__ has joined #openstack-dev00:31
*** rnirmal has quit IRC00:31
*** jtran has quit IRC00:45
*** dragondm has quit IRC00:47
*** deshantm_laptop has joined #openstack-dev01:11
*** Binbin has joined #openstack-dev01:20
*** mattray has joined #openstack-dev01:32
*** mattray has quit IRC02:53
*** Zangetsue has joined #openstack-dev03:00
*** Zangetsue has quit IRC03:02
*** Zangetsue has joined #openstack-dev03:02
*** deshantm_laptop has quit IRC03:07
*** zaitcev has quit IRC03:08
*** zaitcev has joined #openstack-dev03:09
*** sandywalsh has quit IRC03:24
*** vladimir3p has joined #openstack-dev03:56
*** Binbin is now known as Binbin_afk03:57
*** zaitcev has quit IRC04:44
*** zaitcev has joined #openstack-dev04:44
*** vladimir3p has quit IRC04:45
*** Binbin_afk is now known as Binbin04:54
*** Zangetsue has quit IRC04:57
*** Zangetsue has joined #openstack-dev04:58
*** zaitcev has quit IRC05:05
*** Arminder-Office has quit IRC05:40
*** Binbin is now known as Binbin_afk05:57
*** antonyy has joined #openstack-dev05:59
*** openpercept_ has joined #openstack-dev06:13
*** openpercept_1 has joined #openstack-dev06:22
*** openpercept_ has quit IRC06:22
*** openpercept_1 is now known as openpercept06:23
*** openpercept has joined #openstack-dev06:23
*** vladimir3p has joined #openstack-dev06:42
*** antonyy has quit IRC06:44
*** vladimir3p has quit IRC06:47
ttxmorning06:48
ttxvishy, jaypipes: without news from you guys I assume current milestone-proposed is good to go.06:52
*** reidrac has joined #openstack-dev07:04
ttxok then, let's release those things.07:18
*** Binbin_afk has quit IRC07:20
ttxGlance and Nova "diablo-1" milestones are out !07:54
*** RobertLaptop has quit IRC08:20
*** RobertLaptop has joined #openstack-dev08:26
*** BK_man has joined #openstack-dev09:07
BK_mansoren: hi. around?09:12
BK_manCan't run nova-compute on RHEL6.1 anymore: http://paste.openstack.org/show/1523/09:13
reidracthere are reports in the mailing list related to problems with RHEL 6.1 packaged python09:27
BK_manreidrac: thanks, will investigate09:44
* BK_man .oO( please, please fork() me.... )09:45
*** markvoelker has joined #openstack-dev11:15
*** adiantum_ has joined #openstack-dev11:42
jaypipesttx: yep.12:01
jaypipesttx: had to delay S3 bug because of bug in webob: https://bitbucket.org/ianb/webob/issue/12/fix-for-issue-6-broke-chunked-transfer12:05
jaypipescreiht: the above bug may also affect Swift. Much of the code in Glance for doing chunked transfer is identical to Swift's code in the swift.common.client module...12:06
*** bcwaldon has joined #openstack-dev12:20
*** interrante has joined #openstack-dev12:28
*** openpercept has quit IRC12:48
*** deshantm_laptop has joined #openstack-dev12:54
*** adiantum_ has quit IRC12:58
gholtannegentle: https://answers.launchpad.net/swift/+question/16000612:59
gholtttx: soren: Are the ppa links in the description at https://launchpad.net/swift accurate? If so, can we get the 1.4.0 release put in the release ppas area?13:02
*** mattray has joined #openstack-dev13:24
*** foxtrotgulf has joined #openstack-dev13:28
*** troytoman-away is now known as troytoman13:41
*** bcwaldon has quit IRC13:44
annegentlegholt: ah it's ammo  you're after eh? :)13:54
*** dprince has joined #openstack-dev13:56
*** bcwaldon has joined #openstack-dev14:07
jaypipesBK_man: awesome blog post on your experiences getting 200 (then 400) vms running on just 3 physical nodes. nice :)14:21
jaypipesdprince: around?14:30
*** foxtrotdelta has joined #openstack-dev14:31
dprincejaypipes: yo14:32
*** foxtrotgulf has quit IRC14:32
jaypipesdprince: hey bro, think you have a minute to re-check the s3 branch?14:32
*** jkoelker has joined #openstack-dev14:32
jaypipesdprince: got workaround for https://bitbucket.org/ianb/webob/issue/12/fix-for-issue-6-broke-chunked-transfer#comment-504514...14:32
jaypipesor https://bitbucket.org/ianb/webob/issue/12/fix-for-issue-6-broke-chunked-transfer rather14:32
dprincejaypipes: Sure. Are you setting Content-Length as the workaround?14:32
jaypipesdprince: nope.14:33
jaypipesdprince: req.is_body_readable = True14:33
dprincegood!14:33
jaypipesdprince: for webob versions 1.0.2-1.0.7 that is necessary, and a noop for other versions.14:33
dprincejaypipes: Cool. I'll check it out. Fun stuff.14:33
jaypipesdprince: you can read all the gory details on that issue if you feel like it ;)14:33
* jaypipes goes back to working on keystone...14:35
*** vladimir3p has joined #openstack-dev14:47
*** vladimir3p has quit IRC14:48
BK_manjaypipes: thanks, appreciated14:49
gholtannegentle: :)14:51
*** lorin1 has joined #openstack-dev15:01
*** dragondm has joined #openstack-dev15:03
*** reidrac has quit IRC15:11
*** troytoman is now known as troytoman-away15:21
*** vladimir3p has joined #openstack-dev15:24
*** heckj has joined #openstack-dev15:46
jaypipesblamar, bcwaldon: I see you two are getting into a nice grammar nit battle :P15:59
bcwaldonthat's how we roll15:59
*** dprince has quit IRC15:59
jaypipesbcwaldon: :)15:59
blamarThat's how we roll. **15:59
bcwaldonits not blamar that big of a deal16:00
* blamar explodes16:00
jaypipespoof.16:00
blamarWhat did you call me?16:00
jaypipeshehe16:00
blamar:)16:00
heckjjaypipes: I've been disengaged for a while on the CI stuff. Are most folks using vagrant for smoke testing these days?16:02
* heckj (trying to catch up)16:03
jaypipesheckj: mtaylor took the reins of that. AFAIK, vagrant is not being used for the hudson CI stuff, no. Check here: http://jenkins.openstack.org/view/Nova/job/nova-smoketests/16:04
jaypipesheckj: he's using cloudservers to deploy on I believe.16:04
mtaylorthat is correct16:04
mtayloralthough I think there are certainly devs using vagrant for their personal local testing16:05
heckjI saw a mention of it yesterday and read through some of the bits at http://cloudbuilders.github.com/deploy.sh/16:05
mtaylorbut vagrant is not in any of my plans for moving forward (why use virtualbox when you have a datacenter full of other things?)16:06
heckjI'd like to set up a secondary (not official) build and smoke test setup for my own branches - and want to copy the official setup as closely as possible16:06
mtaylorheckj: why?16:06
mtaylorheckj: why not just use the main one?16:06
heckjmtaylor: makes perfect sense, I just don't yet have my own machines to drop it onto...16:06
mtayloryou don't need to16:07
heckjmtaylor: maybe lack of knowledge - don't know how to use the main one other than watch it16:07
mtaylorwould it make your life better if I told you you could just submit your branch locatoin to the main jenkins and have it run through everything?16:07
mtaylorheckj: yes- I'm about to be dropping some emails on folks...16:07
* heckj wonders if he's being absymally stupid16:07
* mtaylor hasn't been great at the communicating about the capabilities here16:07
mtaylorheckj: nope16:07
heckjmtaylor: would be wonderfu!16:07
mtaylorheckj: my bad - I should talk more16:07
mtaylorheckj: but I just wanted to make sure there wasn't a specific reason you wanted something local that might need to be solved or something16:08
heckjI wouldn't have seen it anyway - just getting re-engaged after a few months of other-ness16:08
mtaylorcool16:08
mtaylorwell, my goal is to drop some knowledge and some links out there today or tomorrow to get this party started16:09
heckjmtaylor: sweet! If I wanted to extend the testing setup (say for running automated tests against dashboard), is that something we could hook in there too?16:10
* heckj liked the ideas that were passing around about performance tests and tracking that over time... want to help put that in place16:12
mtayloryup16:12
mtaylorheckj: I'm going to be sending out a thing real soon about setting up a testing group so that we can all talk/coordinate16:12
heckjmtaylor: excellent - looking forward to it!16:13
*** bcwaldon has quit IRC16:19
*** bcwaldon has joined #openstack-dev16:21
*** antonyy has joined #openstack-dev16:30
s1rpjaypipes: hey if you get a second, could i get some feedback on http://etherpad.openstack.org/GlanceZones16:31
jaypipess1rp: surely.16:32
*** Zangetsue has quit IRC16:34
s1rpjaypipes: thanks!16:34
*** bcwaldon has quit IRC17:03
*** bcwaldon has joined #openstack-dev17:06
jk0vishy: you have a chance to look at Scott Moser's feedback on this? https://code.launchpad.net/~vishvananda/nova/fix-novarc-mac/+merge/6318317:14
openstackjenkinsProject nova build #964: SUCCESS in 2 min 46 sec: http://jenkins.openstack.org/job/nova/964/17:24
openstackjenkins* Tarmac: Sudo chown the vbd device to the nova user before streaming data to it. This17:24
openstackjenkinsresolves an issue where nova-compute required 'root' privs to successfully17:24
openstackjenkinscreate nodes with connection_type=xenapi.17:24
openstackjenkins* Tarmac: Bugfix #780784. KeyError when creating custom image.17:24
*** deshantm_laptop_ has joined #openstack-dev17:32
*** dprince has joined #openstack-dev17:33
*** deshantm_laptop has quit IRC17:34
*** deshantm_laptop_ is now known as deshantm_laptop17:34
vishyjk0: done17:37
jk0thankya17:37
openstackjenkinsProject nova build #965: SUCCESS in 2 min 45 sec: http://jenkins.openstack.org/job/nova/965/17:39
openstackjenkinsTarmac: This adds the ability to publish nova errors to an error queue.17:39
*** bcwaldon has quit IRC17:40
*** bcwaldon has joined #openstack-dev17:41
*** antonyy has quit IRC17:49
openstackjenkinsProject nova build #966: SUCCESS in 2 min 48 sec: http://jenkins.openstack.org/job/nova/966/17:55
openstackjenkinsTarmac: Makes novarc work properly on a mac and also for zsh in addition to bash. Other shells are not guaranteed to work.17:55
jk0who handles jenkins, soren?17:56
dprincejk0: I think monty and soren do.17:59
jk0great, thanks17:59
*** mgius has quit IRC18:06
*** mgius has joined #openstack-dev18:10
dprincejk0: Heads up.18:20
dprinceThe latest revision of trunk fails on SmokeStack.18:21
dprinceLooks like the 'nova-manage project zipfile' is causing an error now.18:21
dprinceA regression.18:21
jk0hm18:23
*** avati has left #openstack-dev18:23
* jk0 points to vishy :)18:23
dprinceYeah. My guess is Vish's branch to fix novarc on Mac is the issue.18:23
dprinceYep.18:23
vishygah18:23
vishybleh18:27
vishyit is reading the % as a replacement string18:27
vishyhttps://code.launchpad.net/~vishvananda/nova/fix-novarc-mac/+merge/6327418:29
dprinceShall I smoke it?18:30
vishysure18:30
westmaasdprince: phrasing18:30
vishyi just ran it manually and it works, but couldn't hurt18:30
dprincewestmaas: I though you were cool with it.18:31
westmaashaha18:32
dprincejaypipes: req.is_body_readable doesn't seem to do the trick for me with glance S3.18:46
dprincejaypipes: You wanna connect to that VPC group again and try it?18:47
dprincejk0/vish: lp:~vishvananda/nova/fix-novarc-mac runs clean.18:48
jk0ok18:48
*** jbryce has joined #openstack-dev18:52
jaypipesdprince: ok18:56
dprincejaypipes: So question for you. Have we considered wrapping curl via Glance client?18:57
dprincejk0: We'll address those conflicts here shortly.18:58
jaypipesdprince: no18:58
dprincejaypipes: Is that too heavy handed. Sorry. When I see us write our own chunking code I start to think this already has to exist somewhere.19:00
dprincejaypipes: Good to stay light too (dependency wise).19:00
dprincejaypipes: just throwing it out there.19:00
jaypipesdprince: this actually isn't about the chunking code... it's about webob not handling the HTTP protocol correctly...19:02
jaypipesdprince: it's a server-side bug, not a client-side (where the chunking occurs) bug19:02
dprincejaypipes: Right. But some clients work correctly. Sending a curl command from the command line works right?19:05
jaypipesdprince: yes. but I think it constructs the Content-Length header... haven't investigated it too much further.19:06
dprincejaypipes: Right. Same solution we talked about a couple days ago. Sorry man.19:07
dprincejaypipes: another one for you. So glance-registry.19:07
dprincejaypipes: it automatically runs the SQL migrations on startup. Is that intentional?19:08
dprincejaypipes: I'd actually prefer to just run the 'db sync' manually on the command line. I think the PPA packages actually run it as well in the setup.19:09
jaypipesdprince: gah, we just went over this in this bug: https://bugs.launchpad.net/glance/+bug/77931119:10
uvirtbotLaunchpad bug 779311 in glance "Glance update for Ubuntu failing" [Medium,Fix committed]19:10
jaypipesdprince: bottom line is if we don't run sync, we end up with errors like that bug. If we do run it, we end up with automatic behaviour that you don't like...19:11
dprincejaypipes: Hmm. This just doesn't seem right. I feel like we should always intentially run the migrations. I'll read the thread a bit more closely.19:12
openstackjenkinsProject nova build #967: SUCCESS in 2 min 47 sec: http://jenkins.openstack.org/job/nova/967/19:14
openstackjenkinsTarmac: Fix error with % as replacement string.19:14
jaypipesdprince: that bug fix, frankly, fixed a user error, IMHO. I wouldn't be opposed to reverting the patch.19:15
jaypipesdprince: but it's tough to argue with users ;)19:15
dprincejaypipes: We'll consider me a user then.19:16
dprincejaypipes: :)19:16
sorenjaypipes: Which one? The "Fix error with % as replacement string." one?19:26
sorenjaypipes: Oh, that one further up. gotchan.19:28
sorengotcha, evne.19:28
sorengah!19:28
sorenjaypipes: Why on Earth would you consider reverting that?19:29
dprincesoren: I personally think auto migrating is a slightly evil business.19:31
sorendprince: You'd rather have stuff fail after upgrades by default?19:31
sorendprince: Or what is the third option?19:31
dprinceDon't we already run the 'db sync' in the packages?19:32
sorenfor sqlite, yes.19:32
sorenI'm confused now.19:32
sorenIsn't that what you're saying is slightly evil?19:32
dprinceI guess my thought is that it is plausible someone would create the DB with one account (which can create tables) and then run glance-registry with a SQL account that can only read and write stuff.19:33
dprinceIt just seems bad to auto-migrate when glance-registry starts.19:33
dprinceI'm saying that is evil.19:34
sorenOh, who's suggesting that?19:34
dprinceI think that is what we are doing now.19:34
dprinceSee revision 136 in glance.19:34
sorenOh.19:34
sorenOh, dear.19:34
sorenYeah, that's bad.19:34
jaypipessoren, dprince: sorry, multitasking fail.19:35
sorenI specifically only run "db sync" in the postinst script iff sql_connection points to an sqlite db.19:35
dprincesoren: Correct. I'm in agreement with that.19:36
sorendprince: Otherwise you need to ensure that every host that uses the db is upgraded at the same time. there's no way to do that.19:37
sorenjaypipes: I just wonder why you set it to "fix committed".19:37
jaypipessoren: because s1rp's bug fix went into trunk?19:38
sorenjaypipes: As I understand Rick's comment on that bug, people have to manually create tables and shove data into it for this to work.19:38
jaypipessoren: you can read the merge proposal for a discussion of why we did what we did.19:38
sorenThat's hardly solving the problem in my book.19:38
* soren runs a test19:40
gholtttx: soren: [just a repeat to be sure it was received] Are the ppa links in the description at https://launchpad.net/swift accurate? If so, can we get the 1.4.0 release put in the release ppas area?19:41
sorengholt: Sure.19:41
gholtsoren: Coolio. I know Swift's different in this case as our milestone releases are stable releases.19:42
sorenttx: ping?19:45
*** jbryce has quit IRC19:54
*** dprince has quit IRC19:54
*** dprince has joined #openstack-dev20:13
*** blamar__ has joined #openstack-dev20:13
sorenjaypipes: Ok, it took me a while, but I've verified it now. If you install glance as found in Ubuntu Natty, and upgrade to the trunk version, it will fail.20:13
*** dprince has quit IRC20:14
sorenjaypipes: ...because the sqlite db found after installing glance from ubuntu natty wasn't created using migrate, but rather by sqlalchemy, so there's not migrate_version table.20:14
sorenjaypipes: I don't consider it a solution to tell people to go and create that table themselves.20:15
*** troytoman-away is now known as troytoman20:24
*** heckj_ has joined #openstack-dev20:35
*** heckj has quit IRC20:38
*** heckj_ has quit IRC20:41
jaypipessoren: neither do I...20:49
sorenjaypipes: Phew :)20:52
sorenjaypipes: Ok, glad we agree on that :)20:52
vishysandywalsh: ping20:56
sorenjaypipes: Heh: https://bugs.launchpad.net/glance/+bug/79206120:59
uvirtbotLaunchpad bug 792061 in glance "Table 'images' already exists (Migration Error)" [Undecided,New]20:59
*** bcwaldon has quit IRC21:01
vishydabo: ping21:03
sorengholt: Sorry, I'd really, really like to check with ttx if he's really expecting this particular release to go into the release ppa.21:04
sorengholt: As soon as I have the go-ahead from him, it'll only take a couple of minutes.21:05
vishys1rp: ping21:09
gholtsoren: No problem. Just wanted to get in the hopper.21:10
vishyeveryone is on vacation this afternoon :)21:12
sorenjaypipes: Ok, so this feeds back into my question earlier: If you agree that telling people to create that table themselves doesn't count as fixing it, why do you think that bug should be set to "fix committed"?21:14
*** blamar__ has quit IRC21:17
*** lorin1 has quit IRC21:18
*** troytoman is now known as troytoman-away21:26
*** larzy has joined #openstack-dev21:28
larzyGuys, I'm looking for a document somewhere that details the data that is metered in nova and how that data is made available.  Can someone give me a link to some documentation with that information?21:29
s1rpvishy: ack21:34
sorenlarzy: We're not really doing any metering in Nova yet.21:36
vishys1rp: are you familiar with zone code enough to know why it is using datetime.now()21:37
larzyOK, just to make sure I"m not using the wrong terminology, what I'm looking for is a way to basically enable charging for compute hours used / month, etc.  Are there any plans to include that, or is anyone working on that to your knowledge?21:39
vishylarzy: pretty easy to extract it from database21:40
s1rpvishy: not off-hand; looking now21:40
s1rpvishy: is this in trunk or one of the branches?21:40
vishytrunk21:40
vishys1rp: here is why i'm asking https://code.launchpad.net/~vishvananda/nova/consistent-utcnow/+merge/6330221:40
larzyOK, so the data is captured somewhere, it's just a matter of grabbing it and utilizing to enable billing.21:41
*** markvoelker has quit IRC21:44
vishylarzy: correct.21:45
vishylaunched_at and terminated_at21:45
vishyare columns in the db for every instance21:45
vishyso a query with some multiplications by cores should be all you need21:46
s1rpvishy: i'd guess it was just a mistake not to use utcnow21:46
dabovishy: just got back21:47
vishythat is what i'm thinking but i also notice it uses datetime.datetime.min, so I'm not sure i can change them all21:47
vishydabo: same question as s1rp21:47
dabothe values used there are ephemeral, if that matters21:47
dabothey're used to track polling intervals21:48
dabothey can switch to utcnow() without a problem21:48
vishydabo: ok I'll change them then thx21:48
*** foxtrotdelta has quit IRC21:49
vishydabo, s1rp: thanks, updated21:52
*** interrante has quit IRC21:59
*** larzy has quit IRC22:03
*** mattray has quit IRC22:18
*** dysinger has joined #openstack-dev22:34
vishydprince: ping (in the proper channel this time)22:38
vishyah nm looks like you aren't here22:39
*** jkoelker has quit IRC22:52
jaypipessoren: sorry, was picking up my folks at airport... bout to have dinner. will respond later.23:01
*** jaypipes is now known as jaypipes-afk23:01
*** dragondm has quit IRC23:47
*** deshantm_laptop has quit IRC23:48

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!