Thursday, 2010-12-16

*** kevnfx has joined #openstack00:00
*** joearnol_ has joined #openstack00:01
*** krish has quit IRC00:05
*** Ryan_Lane has joined #openstack00:08
*** reldan has joined #openstack00:13
jbryceKyleM1: - bexar freeze dates00:15
KyleM1Thanks, jbryce00:15
*** dubsquared has quit IRC00:17
*** krish has joined #openstack00:18
jbrycejk0: that xenserver wiki page looks awesome. thanks for putting it up00:19
jk0no problem00:19
*** normanrichards has quit IRC00:20
*** hggdh has quit IRC00:22
*** hadrian has quit IRC00:25
*** reldan has quit IRC00:25
*** sebastianstadil has quit IRC00:28
*** hggdh has joined #openstack00:29
*** dfg_ has quit IRC00:30
*** hadrian has joined #openstack00:33
*** kashyapc has quit IRC00:33
*** sebastianstadil has joined #openstack00:34
*** iammartian has quit IRC00:35
*** joearnol_ has quit IRC00:37
*** jimbaker has quit IRC00:42
*** daleolds has quit IRC00:45
*** rlucio has quit IRC00:48
*** hggdh has quit IRC00:50
*** kashyapc has joined #openstack00:50
*** normanrichards has joined #openstack00:50
*** daleolds has joined #openstack00:51
*** hggdh has joined #openstack00:52
edaytermie: the eventlet branch ready for merge now?01:07
*** KyleM1 has left #openstack01:07
*** dragondm has quit IRC01:09
*** ar1_ has joined #openstack01:13
*** kashyapc has quit IRC01:23
*** kashyapc has joined #openstack01:26
*** sophiap has joined #openstack01:28
*** daleolds has quit IRC01:29
*** abecc has quit IRC01:30
*** jc_smith has quit IRC01:33
*** burris has joined #openstack01:49
*** sophiap_ has joined #openstack01:50
*** sophiap has quit IRC01:50
*** sophiap_ is now known as sophiap01:50
sirp1ant: just noticed that XENAPI_PLUGIN_FAILURE exceptions aren't working the same between XS 5.5 and 5.601:53
sirp1ant: in 5.5 we get the traceback, for some reason in 5.6 we don't…01:53
uvirtbotsirp1: Error: "^^" is not a valid command.01:53
*** jdurgin has quit IRC02:01
*** zul has joined #openstack02:01
*** arcane has quit IRC02:03
sirp1spy: nm, got it sorted out, needed to chmod +x objectstore02:03
*** arcane has joined #openstack02:08
*** krish has quit IRC02:22
*** miclorb_ has quit IRC02:32
*** miclorb_ has joined #openstack02:38
*** miclorb_ has quit IRC02:45
*** rchavik has joined #openstack02:47
*** miclorb_ has joined #openstack02:52
*** miclorb_ has quit IRC02:55
*** sophiap has quit IRC02:56
*** krish has joined #openstack03:04
*** deshantm_ has joined #openstack03:08
*** deshantm_cosi_ has joined #openstack03:08
*** deshantm has quit IRC03:09
*** deshantm_cosi has quit IRC03:09
*** maplebed has quit IRC03:17
*** deshantm_cosi_ has quit IRC03:19
*** deshantm_ has quit IRC03:19
*** deshantm_cosi has joined #openstack03:19
*** deshantm_ has joined #openstack03:19
*** miclorb has joined #openstack03:21
*** ar1_ has quit IRC03:24
*** BK_man has quit IRC03:24
*** ar1_ has joined #openstack03:24
*** burris has quit IRC03:30
*** burris has joined #openstack03:38
*** Ryan_Lane is now known as Ryan_Lane|away03:40
*** localhost has joined #openstack03:53
*** filler_ is now known as filler03:53
*** MarkAtwood has quit IRC03:57
*** MarkAtwood has joined #openstack03:58
*** sirp1 has quit IRC04:03
*** zaitcev has quit IRC04:04
*** deshantm_cosi has quit IRC04:05
*** deshantm_ has quit IRC04:05
*** deshantm_cosi has joined #openstack04:05
*** deshantm_ has joined #openstack04:05
*** Ryan_Lane|away has quit IRC04:07
*** Ryan_Lane has joined #openstack04:09
*** ccustine has joined #openstack04:10
*** rchavik has quit IRC04:10
*** Ryan_Lane has quit IRC04:13
*** Ryan_Lane has joined #openstack04:14
*** rchavik has joined #openstack04:23
*** kashyapc has quit IRC04:28
*** miclorb has quit IRC04:31
*** sirp1 has joined #openstack04:36
*** joearnold has joined #openstack04:37
*** miclorb_ has joined #openstack04:37
*** sirp2 has joined #openstack04:45
*** sirp1 has quit IRC04:46
*** ccustine has quit IRC04:46
*** kashyapc has joined #openstack05:00
*** joearnol_ has joined #openstack05:05
*** joearnold has quit IRC05:05
*** hadrian has quit IRC05:19
*** rchavik has quit IRC05:25
*** BK_man has joined #openstack05:35
*** joearnol_ has quit IRC05:35
*** joearnold has joined #openstack05:36
*** rchavik has joined #openstack05:39
*** sebastianstadil has quit IRC05:46
*** sebastianstadil has joined #openstack05:53
*** dubsquared has joined #openstack05:53
*** f4m8_ is now known as f4m805:57
*** joearnold has quit IRC06:02
*** apanasenko has joined #openstack06:04
*** sirp2 has quit IRC06:06
*** miclorb_ has quit IRC06:22
*** guigui has joined #openstack07:00
*** krish has quit IRC07:01
*** dubsquared has quit IRC07:10
*** krish has joined #openstack07:16
*** arthurc_ has joined #openstack07:35
*** ramkrsna has joined #openstack07:36
*** ramkrsna has joined #openstack07:36
*** almaisan-away is now known as al-maisan07:36
*** gaveen has joined #openstack07:47
*** apanasenko has quit IRC07:48
*** allsystemsarego has joined #openstack07:48
*** ArdRigh has joined #openstack07:59
*** brd_from_italy has joined #openstack08:04
*** rcc has joined #openstack08:15
*** bobl has quit IRC08:19
*** calavera has joined #openstack08:23
*** larstobi has joined #openstack08:23
*** sebastianstadil has quit IRC08:31
*** miclorb_ has joined #openstack08:41
*** krish has quit IRC08:47
*** ibarrera has joined #openstack08:50
*** GasbaKid has joined #openstack08:52
*** ar1_ has quit IRC08:53
*** gaveen has quit IRC08:55
*** Ryan_Lane has quit IRC08:59
*** gaveen has joined #openstack09:07
*** miclorb_ has quit IRC09:15
*** JordanRinke_ has quit IRC09:20
*** irahgel1 has joined #openstack09:24
*** apanasenko has joined #openstack09:32
*** colinnich has joined #openstack09:32
*** guigui has quit IRC09:36
*** reldan has joined #openstack09:39
*** rcc has quit IRC09:41
*** MarkAtwood has quit IRC09:43
*** rcc has joined #openstack09:50
*** threedot has joined #openstack10:12
*** reldan has quit IRC10:29
*** threedot has left #openstack10:29
*** GasbaKid has quit IRC10:59
*** rds__ has joined #openstack10:59
*** GasbaKid has joined #openstack11:00
*** icarus902 has quit IRC11:06
*** icarus902 has joined #openstack11:07
uvirtbotNew bug: #691044 in nova "having LC_ALL and LANG configured, some scripts are not working right, including one unittest" [Undecided,New]
*** colinnich has quit IRC11:23
*** colinnich has joined #openstack11:25
*** ArdRigh has quit IRC11:31
*** reldan has joined #openstack11:38
*** kashyapc has quit IRC11:45
*** Daviey has quit IRC11:50
*** Daviey has joined #openstack11:52
*** rds__ has quit IRC11:53
*** sophiap has joined #openstack12:12
*** ramkrsna has quit IRC12:21
*** yunomc has left #openstack12:23
*** joearnold has joined #openstack12:44
*** sophiap has quit IRC12:48
*** ctennis has quit IRC12:54
*** rchavik has quit IRC13:00
*** reldan has quit IRC13:02
*** apanasenko has quit IRC13:02
*** ctennis has joined #openstack13:11
*** _skrusty has quit IRC13:15
*** gaveen has quit IRC13:21
*** joearnol_ has joined #openstack13:22
*** joearnold has quit IRC13:22
*** bobl has joined #openstack13:23
*** _skrusty has joined #openstack13:28
*** sophiap has joined #openstack13:29
*** mdomsch has quit IRC13:33
*** sophiap_ has joined #openstack13:34
*** sophiap has quit IRC13:34
*** sophiap_ is now known as sophiap13:34
*** GasbaKid has quit IRC13:35
*** blpiatt has joined #openstack13:37
*** bobl has quit IRC13:44
*** normanrichards has quit IRC13:45
*** reldan has joined #openstack13:48
*** hazmat has joined #openstack13:50
*** Adiantum has quit IRC13:50
*** reldan has quit IRC13:52
*** joearnol_ has quit IRC13:52
*** joearnold has joined #openstack13:52
*** joearnold has quit IRC13:54
notmynameannegentle: ^ docs, or something, in japanese13:55
*** sophiap has quit IRC14:01
*** kashyapc has joined #openstack14:04
*** jdarcy has joined #openstack14:05
*** rds__ has joined #openstack14:09
*** aliguori has joined #openstack14:11
*** cloudmeat has joined #openstack14:12
blpiattnotmyname, it looks like a more comphrensive guide based on the presentation he gave at the JOSUG about swift, this is based off of the english readable words and diagrams I recognize ;)14:14
notmynameblpiatt: ah. based on the English I could read, I was guessing it is similar to our docs on the wiki14:15
*** reldan has joined #openstack14:21
*** gondoi has joined #openstack14:21
annegentlenotmyname: yeah I'd like to get that into the wiki, I'll have to contact the author14:22
annegentleblpiatt: can you give me the contact information?14:23
*** Podilarius has joined #openstack14:23
*** kashyapc has quit IRC14:24
*** joearnol_ has joined #openstack14:24
*** allsystemsarego has quit IRC14:29
*** hadrian has joined #openstack14:29
*** allsystemsarego has joined #openstack14:31
*** dkocher has joined #openstack14:39
*** cloudmeat has quit IRC14:44
*** jdarcy has quit IRC14:45
dabosandywalsh: did you get the eventlet-merge code running on your system?14:48
*** cloudmeat has joined #openstack14:52
sandywalshdabo, yup, merging pause now14:59
dabosandywalsh: how did you get objectstore to run? The code I got in that branch still uses twisted14:59
sandywalshdabo, yes, you still run objectstore the old way. It's going away with glance soon-ish15:00
sandywalshdabo, there's a discussion in the merge-prop about it15:00
dabosandywalsh: ok, thanks. I didn't install twisted so that I could test the eventlet stuff15:00
sandywalshdabo, np ... if you run into problems let me know. There's a couple of gotcha's.15:01
dabosandywalsh: thx15:02
sandywalshdabo, also, make sure your using the push from yesterday afternoon. There were important fixes for xen15:02
dabosandywalsh: yeah, I saw those.15:02
*** mdomsch has joined #openstack15:03
*** Podilarius has quit IRC15:04
*** Adiantum has joined #openstack15:04
*** BK_man has quit IRC15:04
sandywalshthis conversion to eventlet is frickin' sweet.15:08
sandywalsh(let's see how much more code I can delete)15:09
*** hggdh has quit IRC15:11
*** brd_from_italy has quit IRC15:12
*** dendro-afk is now known as dendrobates15:14
*** brd_from_italy has joined #openstack15:14
*** allsystemsarego has quit IRC15:14
*** hggdh has joined #openstack15:15
sandywalshin case you missed it yesterday, jk0 posed his xen server/nova install notes:
*** dkocher has quit IRC15:17
annegentleyay jk0, appreciate it15:17
*** sirp1 has joined #openstack15:18
*** MarkAtwood has joined #openstack15:22
*** Podilarius has joined #openstack15:25
*** dubsquared has joined #openstack15:26
*** f4m8 is now known as f4m8_15:29
*** kashyapc has joined #openstack15:36
sandywalshdabo, does ./ work for you in that branch? fails horribly for me. nosetests work though.15:37
sandywalshdabo, looks like it's trying to do a redis connection15:37
edaytermie, vishy: anything else you guys want to do with eventlet branch, or should we hit merge?15:37
* ttx cheers15:38
sandywalsheday, does runtests run clean for you on that branch? I'm getting redis connection complaints15:41
sandywalsheday, checking to ensure it's not my env15:42
edayredis connection, huh?15:43
*** ppetraki has joined #openstack15:43
edayoh yeah, fake ldap still uses it15:43
*** johnpur has joined #openstack15:44
*** ChanServ sets mode: +v johnpur15:44
sandywalsheday, lots of these:
*** joearnol_ has quit IRC15:44
sandywalsheday, checking against raw branch (no changes)15:44
*** aliguori has quit IRC15:45
*** dragondm has joined #openstack15:45
*** sophiap has joined #openstack15:48
sandywalsheday, confirmed ... tests fail from the raw branch (no changes) ... for me at least.15:48
edaysandywalsh: hm, all tests pass for me15:48
edaybut I have redis running15:49
sandywalsheday, try turning it off ... unit tests shouldn't need it, correct?15:49
*** kevnfx has quit IRC15:49
*** sophiap has quit IRC15:50
edaysandywalsh: fake ldap uses it, so I thin it does15:50
edayyeah, trunk uses it still15:51
sandywalshsorta violates the idea of a unittest, no?15:52
*** sophiap has joined #openstack15:53
edaywell, at least it doesn't require ldap :)15:53
edaythey also use sqlite15:54
*** larstobi has quit IRC15:54
*** sophiap_ has joined #openstack15:54
*** blpiatt has quit IRC15:55
sandywalshat least it doesn't require windows :)15:55
*** sophiap has quit IRC15:58
*** sophiap_ is now known as sophiap15:58
jbrycesandywalsh, eday: i'm seeing the same thing as sandywalsh. i think it's because in trunk was skipping ldapmanager tests (and not requiring redis) before15:59
*** daleolds has joined #openstack15:59
creihtpolvi: congrats!!!! :)15:59
creihtand welcome16:00
*** dirakx has joined #openstack16:00
sandywalshjbryce, you're correct, but I get other tests failing too, not just the ldap ones16:00
jbrycesandywalsh: yeah...i do as well16:01
*** MarkAtwood has quit IRC16:01
edayjbryce: if you test trunk and watch redis, it gets used during the run16:01
sandywalshhmm, even with redis running it fails16:01
*** joearnold has joined #openstack16:01
sandywalshdo I need to init redis somehow? I see the connections hitting it, but no joy16:02
edaysandywalsh: nope, should just work16:02
jk0sandywalsh: is python-redis installed?16:03
jk0the unit tests call it16:03
sandywalshjk0, yes, it's part of the virtenv16:03
sandywalshI'm thinking a version mismatch16:03
edaysandywalsh: so trunk runs fine for you?16:03
sandywalshserver is 1.0216:03
sandywalsheday, it did ... lemme try again16:03
jk0ah, I've not been using venv16:04
edayi have redis 2.016:04
sandywalshah ... that'll do it16:04
jk0packages are much better IMHO :)16:04
jbrycea clean trunk runs fine me and after applying just the merge start getting failures16:05
*** sophiap has quit IRC16:06
*** sophiap has joined #openstack16:06
*** colinnich has quit IRC16:11
edayjbryce: trunk and trunk+eventlet merge run fine for me.. must be a dependency issue16:12
*** hggdh has quit IRC16:14
jk0yeah, I had 0 issues firing up the new eventlet branch16:14
jk0but I also didn't use the new launcher...16:15
edayjk0: what about unittests?16:15
*** vvuksan has joined #openstack16:15
jk0haven't tried, let me check16:15
sandywalshback from standup ... upgrading to redis 216:16
jk0oh my, yeah, definitely redis failures16:17
jbryceeday: i agree, but it's a changed dependency since it was working before. i'm just trying to answer what it is since it's affected 2 of us already and will probably affect others.16:17
jbrycesandywalsh: i've got redis-2.0.0-rc416:18
edayjbryce: yup, definitely need to find out what, hard on my end since I can't reproduce. any stack traces?16:18
*** guigui1 has joined #openstack16:18
sandywalshrun clean with redis 216:19
sandywalshjbryce, I'm using 2.0416:20
jbrycesandywalsh: stable 2? like 2.0.4?16:20
jbrycesandywalsh: thx...i'll try that too16:20
edayI have 2.0.0 and works fine16:20
jbryceeday: - all variations of that same Error 6116:21
sandywalsheday, perhaps you have a compatible python client ... looked like marshaling errors16:21
sandywalsheday, I'm using virtenv redis client16:21
edayIt's python-redis 2.0.016:21
*** brd_from_italy has quit IRC16:21
sandywalsheday, me too16:22
edayit's a simple text based protocol, can't imagine what might have changed at that level on a minor version bump :)16:22
edaybut you went from 1.x to 2, which was probably the issue16:23
edayDidn't think it ever worked with 1.x16:23
sandywalshwell, mine works now, but the larger issue is this dependency (my $0.02)16:23
sandywalshfunny also that the ldap tests aren't being skipped as before16:24
edaysandywalsh: so the *only* think you changed was redis-server from 1.x to 2?16:24
sandywalshand running the server at all16:24
sandywalshwhich I never did before16:25
jk0I'll try the same thing16:25
edayhmm, nothing changed in auth (module or unittests). weird16:27
*** seshu has joined #openstack16:27
jk0sandywalsh: so you installed the actual server and not just the python module?16:27
*** kevnfx has joined #openstack16:28
*** calavera has quit IRC16:28
sandywalshjk0, the python module (client) is part of virtenv install16:28
jk0yeah, I don't use venv16:28
jk0I just installed python-redis in the past16:28
jk0but not redis-server16:28
sandywalshjk0, but I had to install & run redis-server 2.0.4 to get tests to pass16:29
jk0ok, that's what I was wondering16:29
sandywalshjk0, likewise ... that's why this is smelly16:29
edaywere you seeing a bunch of [SKIPPED] before?16:29
sandywalshno longer16:29
edayin output?16:29
jk0same here16:29
edayahh, ok. just shutdown my redis server and I get that16:29
sandywalshI remember distinctly seeing .skip = True in the tests previously16:30
edayand now they are errors16:30
sandywalshbut skip was only for ldap ... these errors are all over the place16:30
edayso redis-server was always needed, they were just skipped before16:30
jbryceeday: run against a clean trunk with redis stopped16:31
edaywell, fake-ldap == redis16:31
* jk0 has never installed redis-server16:31
edayjbryce: I did16:31
sandywalsheday, for example, my pause test fails and it definitely never needed redis server running before16:31
edayok, sorry, not needed, but needed if you wanted everything run16:31
jk0only the pythos-redis module, not the actual server16:32
*** iammartian has joined #openstack16:32
sandywalsh./                self.skip = True16:33
sandywalshsomethings up ... if this is being ignored16:34
edayjk0: you need redis-server if you didn't want auth tests skipped16:34 trunk did not need redis before because it was apparently skipping things that required redis. now requires redis running because it is running more tests.16:34
*** kashyapc has quit IRC16:34
edayor it doesn't know how to skip, the test base classes were tweaked, trying to find the difference16:35
sandywalshjbryce, actually that's a good point ... my pause test shouldn't run with ... only with nosetest16:35
*** sophiap has quit IRC16:35
*** sophiap has joined #openstack16:36
ttxdevcamcar: would you consider bug 671197 fixed with your recent branch merge ? If yes, please set bug to FixCommitted16:36
uvirtbotLaunchpad bug 671197 in nova "Smoke tests are broken" [Medium,In progress]
jbryceeday: do you know if the switch from test.TrialTestCase to test.TestCase would have had any impact on that?16:40
*** jdarcy has joined #openstack16:40
*** joearnold has quit IRC16:41
jk0jbryce: I bet that's it16:42
jk0just did a grep for that and it's all over the place in nova16:42
jk0I'll bet they need to me moved to TestCase16:43
edayjbryce: possibly, but trial should be a simple wrapper over unittest, just allowing defers16:43
*** joearnold has joined #openstack16:43
*** jfluhmann has quit IRC16:44
edayI was looking for other changes in the base test classes16:45
*** larstobi has joined #openstack16:45
edayit appears standard unittest has no skip functionality, so trial does add this16:49
*** deshantm_ has quit IRC16:49
*** deshantm has joined #openstack16:49
creihteday: nosetests allows you to skip tests16:50
edaycreiht: I wish we were using that everywhere :/16:50
creihteday: make it so! :)16:50
*** kashyapc has joined #openstack16:51
edaycreiht: getting closer, most of twisted will be out of the way which will make unittest hacking easier16:51
*** schisamo has joined #openstack16:53
*** schisamo_ has joined #openstack16:55
*** kashyapc has quit IRC16:56
*** ChanServ sets mode: +v pandemicsyn16:56
*** rcc has quit IRC16:56
uvirtbotNew bug: #691174 in nova "On Ubuntu 10.10, upgrading a nova install stalls" [Undecided,New]
*** jfluhmann has joined #openstack16:56
*** schisamo has quit IRC16:58
*** schisamo_ is now known as schisamo16:58
*** mdomsch has quit IRC16:58
*** BK_man has joined #openstack17:02
*** al-maisan is now known as almaisan-away17:04
*** schisamo has quit IRC17:05
*** schisamo has joined #openstack17:05
*** aliguori has joined #openstack17:12
*** kashyapc has joined #openstack17:12
*** abecc has joined #openstack17:13
*** hggdh has joined #openstack17:13
dubsquaredannegentle:  in today?17:16
*** daleolds has quit IRC17:16
spectorclandubsquared: I saw anne on earlier this morning but her iChat shows idle so perhaps she stepped away17:22
*** jfluhmann has quit IRC17:22
dubsquaredspectorclan:  thanks, ill just shoot her a PM then17:22
jbrycedubsquared: she'll probably be on and off today17:23
jbrycedubsquared: or as we in the industry like to say "she'll have intermittent availability"17:24
dubsquaredjbryce:  did she send an e-mail stating she'll be, "out of pocket"17:25
jk0looks like termie removed flags from vm_utils, that's not going to work :)17:25
jk0I'll fix17:25
dubsquaredI still think that term is so odd...17:25
*** Ryan_Lane has joined #openstack17:26
*** guigui1 has quit IRC17:26
jbrycedubsquared: true... brett favre always did some of his most productive work out of the pocket17:27
mtaylorttx: so, I need to patch tarmac to use your plugin?17:34
*** Ryan_Lane has quit IRC17:36
*** schisamo_ has joined #openstack17:38
*** schisamo has quit IRC17:38
*** schisamo_ is now known as schisamo17:38
jk0termie: I proposed a small merge to your banch:
jk0termie: it just puts nova flags back in vm_utils17:40
*** Ryan_Lane has joined #openstack17:42
*** maplebed has joined #openstack17:43
*** joearnold has quit IRC17:44
*** joearnold has joined #openstack17:45
*** schisamo has quit IRC17:45
*** schisamo has joined #openstack17:46
jk0sandywalsh: did you put together any docs on getting getting the cloudservers API working, or does it "just work" in your fork?17:46
*** ccustine has joined #openstack17:46
*** jdurgin has joined #openstack17:49
*** kashyapc has quit IRC17:49
*** abecc_ has joined #openstack17:51
*** kashyapc has joined #openstack17:51
*** abecc has quit IRC17:53
*** abecc_ is now known as abecc17:53
*** kashyapc has quit IRC17:59
xtoddxjaypipes: around?18:00
*** henrichrubin has quit IRC18:00
xtoddxor does anybody else have a recipe for letting optparse ignore unknown flags?18:00
*** schisamo has quit IRC18:00
*** schisamo has joined #openstack18:01
*** MarkAtwood has joined #openstack18:01
*** henrichrubin has joined #openstack18:01
*** rlucio has joined #openstack18:02
*** kashyapc has joined #openstack18:03
*** schisamo has quit IRC18:10
*** schisamo has joined #openstack18:10
*** ibarrera has quit IRC18:16
*** rlucio has quit IRC18:16
*** spy is now known as antonym18:17
*** joearnold has quit IRC18:17
dabois there anything in the xenserver stuff for reading/writing to xenstore? I didn't find anything with grep18:18
*** joearnold has joined #openstack18:19
*** daleolds has joined #openstack18:22
*** BK_man has quit IRC18:23
vishyafaik if you have python-redis installed and redis is not running you will get a lot of failures18:27
vishyeday: I recall termie saying something about there being no skip in unittest18:27
jk0could someone explain to me what actually fires off the db creations when, say, creating a new user on a new nova install?18:29
vishyjk0: sqlalchemy18:29
edayjk0: nova/db.sqlalchemy.models.register_models18:29
jk0is there something specific that gets called though?18:29
vishyjk0: called from in sqlalchemy18:30
* jk0 completely overlooked that18:30
*** smoser is now known as smoser`18:30
sandywalshjk0, you have to do a few things:18:30
*** smoser` is now known as smoser18:30
sandywalshjk0, 1. --allow_admin_api=true to flagfile18:31
sandywalsh2. set CLOUD_SERVERS_USERNAME = your nova username18:31
* jk0 is taking notes and will update wikis accordingly18:31
edayvishy: so, I've been trying to start breaking out the data into different bits for distributed updates, and I keep getting stuck down various tangents.. to do things in a distributed fashion, we're really going to need to blockbox different workers I think, but right now we're a bit off18:32
*** arthurc_ has quit IRC18:32
sandywalsh3. set CLOUD_SERVERS_API_KEY=EC2_ACCESS_Key (with the :project stripped off at the end)18:32
sandywalsh4. set CLOUD_SERVERS_URL=   (or whatever)18:32
vishyeday: how can we get to where we need to be?18:32
edayerr, blackbox18:32
edayvishy: the idea of a central data store is very much at the heart of, well, everything18:33
vishyeday: afk a sec / scrum / but keep going18:33
sandywalshjk0, you may need to use instead of ip address18:33
sandywalshjk0, or remove the subdomain part from api/__init__18:33
sandywalshjk0, we need a switch for that in nova18:33
sandywalshjk0, "don't use subdomains"18:34
sandywalshbut that should do it18:34
sandywalshany news on the eventlet branch tests?18:34
jk0thanks sandywalsh18:34
edayvishy: it's hard to explain, but the only way I've made sense of it so far is to radically overhaul the way workers integrate with other parts of the system18:34
sandywalshvishy, eday correct, there is no skip in unittest, only nosetest18:36
sandywalshuhh, sorry, that's twisted18:36
edayvishy: basically, make the workers completely self containted, ie, create their own dbs/tables/..., and hide behind a simple API. This also means having a generic schema/api for each service, along with the implementation18:37
sandywalshre: skip ...
edayie, the api server/scheduler would have it's own 'volume' table, but wouldn't have any details about blades/targets/...18:37
*** cloudmeat is now known as bobl18:38
sandywalshjk0, any progress on tests?18:39
jk0basically comes down to needing redis-server running18:40
jk0I know :(18:41
sandywalshI smell an evening hack coming up18:41
edaywe should just get fake_ldap off of redis18:42
sandywalshreplace it with a dict :)18:42
xtoddxsomewhere i have a fakeldap that is totally in-memory18:42
edaywell, it may need to be multi-process18:42
termiejk0: sounds like something needs tests ;)18:42
sandywalsheday, well then tests shouldn't use fake-ldap18:43
sandywalshwe need a faker-ldap18:43
jk0termie: citrix guys are working on that18:43
termieeday: i say merge it and start ammending it, having this large patch outstanding is horrible18:43
edaytermie: it wont merge until at least pep8 is fixed18:43
termiesure, there only appear to be a couple issues there to fix though, i mostly just mean there are a variety of things we still want to do18:45
termiebut the patch is big enough as it is18:45
termieeday: for example, objectstore, i agree we should get it running again18:47
termieeday: but in a subsequent patch18:47
edaysure, I'm fine with that18:48
*** schisamo has quit IRC18:48
*** schisamo has joined #openstack18:48
edayI want to get it in asap too, but we at least need pep8 fixed or tarmac will reject it18:48
edaynote I've already hit the approve button :)18:48
*** rlucio has joined #openstack18:50
jk0sandywalsh: we're supposed to pull your cloudservers fork, right?18:52
termieeday: i am guessing the tools directory is not being checked by pep8?18:53
edaytermie: nope, bin/* and nova18:53
sandywalshjk0, yup18:54
edayalthough we probably shoudl check everything18:54
*** kevnfx has quit IRC18:54
vishytermie, eday: someone is going to have to fix the initscripts too right?18:55
termieeday: agreed, i just don't want to go fix the whole tools directory right nwo :p18:55
edaytermie: nope, of course not :)18:55
termievishy: aye18:55
edaytermie: I left the command in the merge req comment with output18:55
* vishy starts jamming on the approve button18:55
termievishy: where are those?18:55
termiedifferent branch somewhere?18:56
vishytermie: they are all in soren's packaging branch18:56
termieis there a reason that isn't in the main branch?18:56
vishytermie: all the debian specific stuff is separate18:56
termievishy: why?18:57
vishyapparently that is the standard way for package maintainers, not to clutter up the main branch18:57
*** AimanA has quit IRC18:57
termievishy: hmm18:57
* vishy stays out of that one18:57
jk0sandywalsh: seen this before?
*** AimanA has joined #openstack18:58
*** rds__ has quit IRC18:59
reldanHi guys :) Please help me - are there any blueprint about availability zones?18:59
sandywalshjk0, hmm, assuming of course you've replaced <domain> :)19:01
jk0tried the IP as well19:01
sandywalshjk0, can you change the subdomain to just os.19:01
jk0sure thing19:02
sandywalshor have you set FLAGS.osapi_subdomain19:02
jk0nah, I haven't19:02
sandywalshtry the latter19:02
sandywalshoh, I think tr3buchet was running into a problem with image-list19:03
sandywalshtry cloudservers list19:03
sandywalshjust to see if auth works generally19:03
jk0same thing19:03
sandywalshok, try that subdomain thing19:03
sandywalshbeyond that see what api log is saying19:04
sandywalshusername is the name you set in nova-manage user admin <name> correct?19:05
jk0hang on a sec19:05
edayopenstack api is 8774 by default19:05
jk0mines running on 8773 and 419:06
eday8773 == ec2, 8774 == openstack/rs19:06
sandywalshyeah I use 877319:06
sandywalsheday, I don't think it cares19:07
edaywith the exception of requests to /, either should work19:07
sandywalshit routes by /v1.0/19:07
edaysandywalsh: yeah, but there is a conflict on /19:07
edayif the APIs are checking that first19:07
jk0ok, changing it to 8774 worked19:07
jk0thanks guys19:07
sandywalshso they both auth to the same url, huh19:08
sandywalshjust on different ports19:08
* sandywalsh make mental note19:08
edaysandywalsh: you can setup subdomain checks too on the same port, but there are no subdomains either19:08
edayerr, no subdomains by default19:09
sandywalshgotcha ... that's why I was running into subdomain problems earlier19:09
sandywalshwell. almost, even setting subdomains={} causes it to use subdomain checks I think19:09
sandywalshyou have to remove the subdomains flag completely19:10
termieis there a reason that isn't in the main branch?11:08 < sandywalsh> so they both auth to the same url, huh19:10
edaysandywalsh: yeah, it chooses one by default19:10
termieweird copypaste19:10
termieanyway, that branch doesn't exist and i am not sure what else it would be19:10
termieyou haev swift/packaging19:10
termieoh, i missed a whole line19:10
edaytermie: soren can sort that out later19:10
*** BK_man has joined #openstack19:11
termiesoren: ^^ was wondering where the packaging branch was19:11
*** kevnfx has joined #openstack19:11
*** apanasenko has joined #openstack19:13
jk0I added the CS API instructions to if anyone is interested19:14
sandywalshthanks jk0!19:14
jk0my pleasure :)19:14
*** kevnfx has quit IRC19:15
KnightHackerjk0: This is great. Thanks a lot.19:18
jk0KnightHacker: no problem19:22
edayvishy: so, any comments on my thoughts earlier?19:22
vishyeday: can we limit it to tables for the first version19:23
edayvishy: hmm?19:23
jaypipesxtoddx: back now... sorry, was doing some last minute holiday shopping..19:24
xtoddxjaypipes: no worries, i was just hoping you'd have some insight on how to handle unexpected arguments.19:25
vishyeday: It would be nice if workers could still talk to the same db19:25
vishyeday: but basically you're talking about getting rid of all of the foreign keys19:25
xtoddxjaypipes: right now we try to keep them in their own arg list (returned from FLAGS()) but optparse errors out.  Even if you subclass the parser and override error() it still stops processing later args, and doesn't add that one back into the arg list19:26
jaypipesxtoddx: I believe you will need to wrap parse_options() with a try block and ignore InvalidOption (or something like that...) I did that to ignore the DuplicateOption errors.)19:26
edayvishy: and more.. basically getting rid of workers even knowing/having access to other tables19:26
jaypipesxtoddx: sorry, parse_args(), not parse_options()19:26
xtoddxjaypipes: will i need to add in some sort of iteration to keep parsing args after the one that threw the error?19:27
jaypipesxtoddx: hmm... scratch that... I did a try/excpt around add_option(), not parse_args...19:28
jaypipesxtoddx: unfortunately, I believe the only way would be to call parse_args() in a loop, catching the exception for unknown option, and when catching that error, issue an add_option() to the parser for it. :(19:29
jaypipesxtoddx: what is the use case you are trying to solve?19:29
vishyeday: probably going to have to go one at a time there...19:29
edayvishy: so, for a given service (lets take volume), it would implement an a volume API. the api server/workers can use this, but there is nothing saying it will use the DB/queues/...   Data being returned from this API needs to be in a common format, and without details (ie, blades, iscsi targets details ,...), UNLESS we want to use a schemaless DB, but I won't go there yet19:29
xtoddxchanging how we do flags in nova to the options class in openstack-common19:30
edayvishy: sure, I'm just trying to find the target to make sure we're moving towards the same place19:30
xtoddxjaypipes: we want to have unknown flags be ignored, as they are now19:30
vishyeday: how does one request other data?19:30
edayvishy: use the volume API19:30
vishyeday: iscsi target, etc.19:30
xtoddxjaypipes: and also accessible to whatever fires the parse option, so that it can get a list of args from the command line ( uses the list to determine what tests to run)19:31
edayit can have components that run in different places19:31
jaypipesxtoddx: I fail to see the point of that, but like I said, the only way I can think of to do that is the above strategy..19:31
xtoddxjaypipes: cool,  i'll start experimenting with that then19:31
vishyeday: it will still need to get that data through the api somehow19:31
edaybut it owns the data underneath, and only exposes common bits19:31
jaypipesxtoddx: I advise to try and keep the hacks to a minimum :) simple and straightforward is better...19:32
xtoddxjaypipes: thanks for clearing that up for me ;)19:32
jaypipesxtoddx: :P19:32
edayvishy: if compute needs to directly know about iscsi targets, how can we ever implement something different? we need to create an abstraction layer that hides the details19:32
jaypipesxtoddx: /me sees a long painful path in your future...19:32
*** Abd4llA has joined #openstack19:33
vishyeday: the part of volume running on compute will need to make a remote request to get info.  So is there a hidden private api for extra info?19:33
edayvishy: when compute makes a request to attach a volume, it enters into the volume-implementation code there, which may also coordinate externally19:36
edayvishy: it's similar to how it works today...19:36
edayvishy: the main issue I'm trying to address though is who owns data, and that there are public and private versions for a given service...19:36
edayvishy: for example, when a volume is created and returned to the scheduler/api server, it doesn't need to know all the details. it just needs to know the name, who owns it, and where to find it19:37
edayvishy: so scheduler should not be poking directly at the implemntation table (which does have the details)19:38
vishyeday: right.  I was considering a simple api extension like get_extended_data that could return an implementation specific blob19:38
*** apanasenko has quit IRC19:38
vishyeday: so every driver doesn't have to invent a secondary api19:38
edayvishy: well, you should never need to get the extra data, it can stay behind the implemented API, no?19:38
*** zaitcev has joined #openstack19:39
*** BK_man has quit IRC19:40
vishyeday: it seems like biting off this whole thing at once is going to be hard19:40
edayvishy: it will be, and I'm not suggesting we do it all at once. I've thrown away multiple branches over the past 3 days trying to start with a small bite :)19:41
vishyeday: passing dictionaries with relevent info first19:41
vishyeday: instead of db ids19:41
*** dfg_ has joined #openstack19:41
edayvishy: so, I was thinking of splitting up nova.db, where each worker defines it's own models (doing one at a time first)19:42
edayand trying to build an API on top of it, passing only dicts with common values19:42
vishyeday: i think that makes sense, I would probably do it in the other order19:43
edayvishy: one thing with nova.db, do we really need those abstractions over sqlalchemy? I know we discussed this before during the orm-branch merge, but punted to the summit (but forgot)19:44
edayfor example, what if nova.auth.dbdriver used sqlalchemy directly? it seems just as easy to write a nova.auth driver using, say, cassandra, rather than put the nova.db shim in there19:44
edayand if so, we should make nova.db provide some more common things, but rip all the specific calls out (those need to move to driver implementations since they vary between implementation)19:47
jk0is there a reason we haven't used Floats yet in the db?19:50
jk0(other than just not needed them yet)19:50
edayprecision varies by implementation? what do you need them for?19:50
eday(probably just not needing them)19:51
jk0diags come back from xapi as floating points19:51
*** ArdRigh has joined #openstack19:51
edayjk0: you can always multiply by precision factor and store as bigints for most stuff19:52
jaypipesdoes anyone have any objections to me merging in i18n-strings now?19:53
jk0BigInts aren't used yet either19:53
edayjaypipes: I'm guessing it will conflict with eventlet, so I guess the q is who wants to fix conflicts? :)19:54
jaypipestermie: If you want to go first, I can do the conflict fixes in my branch...19:54
jaypipeseday: termie's eventlet_merge is ready to go?  or pep8 fixes not done yet>?19:54
edayjk0: well, ints if it's within range... but bigints are safe to use too. I'm not sure if there is a good reason not to use floats though19:55
jaypipesjk0: better to use a fixed precision decimal field than float, imho.19:55
edayjaypipes: looks like they're in there, lets try a merge with eventlet first :)19:55
jk0I'll go that route19:55
jaypipeseday: you got it.  you do it.19:56
* jaypipes waits for eday to mark Approved with baited breath.19:56
edayvishy: did I lose you? :)19:56
*** reldan has quit IRC19:58
vishyeday: we put the db abstraction in because we had changed backends twice already19:59
vishylets approve19:59
sorentermie: It's at lp:~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu20:00
vishyunless you are putting a nighcrawler in your lungs20:00
edayvishy: so, the issue with nova.db is it's a reimplementation of many of the specific APIs rolled back into one20:01
vishyyes, it should be split up20:01
edayvishy: and we need to create api and implementation versions for each service, if we move forward with plan above20:02
vishybut I think the api changes are more important20:02
edaybut I think the split needs to happen before we can do the API changes :) (i tried it already and tossed the branch)20:02
edayvishy: I think we may need a separate lookup db for routing at the api/scheduler levels, and then the implementation specific dbs (which may have more details)20:06
uvirtbotNew bug: #691254 in swift "Auditors Duplicate Tests and Miss Entire Devices" [Undecided,New]
vishytermie: any input on this?20:06
*** GasbaKid has joined #openstack20:06
vishy^^ the db / api separation discussion with eday that is20:07
uvirtbotvishy: Error: "^" is not a valid command.20:07
edayunless we wanted to switch to something schemaless, and have a well defined key set for the public APIs to poke at20:07
vishyeday: I must eat, but it sounds resonable, I just want to avoid complicating the system until necessary20:08
*** mdomsch has joined #openstack20:08
ttxmtaylor: either patch it to deploy the patch at the usual location, or just add the plugin inside tarmac code20:09
edayvishy: what are we waiting for? :)20:09
edaytermie: eventlet still has pep8 issues, see output20:09
*** GasbaKid has quit IRC20:11
ttxmtaylor: the bug in tarmac being it doesn't support plugin outside the tarmac/plugins directory.20:11
*** irahgel1 has left #openstack20:12
*** GasbaKid has joined #openstack20:13
*** pothos has quit IRC20:15
*** pothos has joined #openstack20:15
*** JuanPerez has joined #openstack20:17
*** k0p has quit IRC20:24
sandywalshhas eventlet been merged into trunk?20:26
sorensandywalsh: They tried half an hour ago, but it got rejected due to a couple of pep8 things.20:28
sorensandywalsh: So any minute now, I suppose.20:28
* soren isn't here, by the way20:29
sandywalsh:) thanks soren20:29
dabotrying to remove redis from the fake-ldap. Looking good, but running the tests I get a twisted plugin error:
sorenI have no clue what /usr/lib/python2.6/dist-packages/twisted/plugins/ is.20:33
dabosoren: you and me both!20:33
*** EdwinGrubbs is now known as Edwin-afk220:34
edaywhy would it be trying to create files inside of /usr/lib!?20:34
edaythats what /var is for20:35
soreneday: Probably because that's where sys.path happens to be.20:36
daboi killed a bunch of processes that were still running from previous sessions. that seemed to get rid of the issue20:37
dabonow I run the tests, and get no output20:37
creihtcourtesy of littleidea20:38
sorendabo: Which OS is this?20:38
*** reldan has joined #openstack20:38
daboubu 10.10, with xenserver20:38
sorenHm.. Ok.20:40
* soren asked in #twisted. They said to blame the OS vendor.20:40
dabogee thanks, #twisted!20:41
sorenTrouble is... that's me.20:41
vishyself-blame = shame20:41
edaysoren: it used to be you, now you can punt :)20:41
vishyare you ashamed?20:41
termievishy: weird i just ran it, will check again20:42
soreneday: /me is still an Ubuntu core dev20:42
termievishy: i am not getting any errors from pep820:42
sorenvishy: Ever so slightly.20:42
dabosoren: :)20:43
sorentermie: Which version of pep8 are you using?20:43
vishydifferent version of pep8 perhaps?20:43
termiewhichever one installs normally20:43
dabosoren: it was a stale objectstore process that I had to kill to get rid of the error20:43
vishydid you do pep8 bin/*20:44
termiei am in the directory running it right now20:44
vishyit doesn't pick up the bin files because they don't end in .py20:44
*** apanasenko has joined #openstack20:44
sorendabo: That fixed it?!?20:44
sorenShouldn't have.20:45
dabosoren: the twisted plugin thing, yeah20:45
* soren calls Heisenbug on it20:45
dabosoren: prolly20:45
sorenDid you run anything as root?20:46
sorenThat would "fix" it.20:46
sorenAs in make the problem go away, but in the process destroy the evidence needed to understand why it broke to begin with.20:46
* vishy waits for lp to catch up so he can jam approve20:47
dabosoren: I didn't change anything except killing the old process, and then re-launching the services20:47
sorendabo: You running them as root?20:47
dabonot right now20:47
* soren wanders off, has stuff to do20:48
vishysoren: have fun, we're about to break all of the packaging for nova20:48
*** mdomsch has quit IRC20:49
edayprobably more than just packaging :)20:49
*** reldan has quit IRC20:49
uvirtbotvishy: Error: "W^W^W^Weverything" is not a valid command.20:49
vishyeday: you reapproved my approve20:50
sorenvishy: Oh, how so?20:52
*** notmyname has quit IRC20:53
vishysoren: the eventlet merge, no more logging / deamonizing stuff20:54
sorenvishy: No.. more.. logging?20:54
vishysoren: as in flags20:54
sorenThat's going to suck, isn't it?20:54
sorenCan you elaborate a bit? Where do logs go now?20:54
vishysoren: upstart or runit or ... needs to put stuff in logfiles20:55
sorenOh, freck.20:55
*** notmyname has joined #openstack20:55
*** ChanServ sets mode: +v notmyname20:55
vishyyeah, that's why we were looking for the packaging branch20:55
vishythe initscripts need hacking20:55
sorenYou saw the link, right?20:55
*** KnightHacker has quit IRC20:56
sorenThis is a good opportunity to upstartify it, though. I should get moving on that.20:56
*** KnightHacker has joined #openstack20:56
sorenThere. Merged.20:56
* soren really wanders off now20:57
openstackhudsonProject nova build #306: SUCCESS in 1 min 14 sec:
jbrycecreiht: now we need the exabyte version of the picture20:59
vishyit meged21:00
vishynow we get to remake all of the other branches!!!!21:01
*** JuanPerez has quit IRC21:02
edayjaypipes: want to merge your i8n branchc with trunk now?21:04
edayvishy: ahh, I saw it as unmarked when I hit approve21:05
jaypipeseday: yup, gimme 1521:05
*** arcane has quit IRC21:08
jk0would anyone mind reviewing this quick? it's just a tiny little thing:
jk0a tiny thing but yet really annoying :)21:09
sandywalshjk0, why is that needed?21:11
jk0get_rrd() uses flags21:12
jk0flags were removed on accident in the new eventlet branch21:12
*** hggdh has quit IRC21:12
sandywalshodd it never caused an issue running21:12
sandywalshor in the tests21:13
jk0only get_diagnostics() calls get_rrd() (my stuff)21:13
jk0there are no tests for it yet21:13
dabojk0: "16:09:40 jk0: a tiny thing but yet really annoying :)" -> TWSS21:13
jk0uh huh21:13
*** hggdh has joined #openstack21:14
sandywalsheday, is that a pep8 violation?21:14
sandywalsh(your comment on jk0's piece)21:15
edaysandywalsh: nope, it's a HACKING file violation (nova specific)21:15
sandywalsheday, oh ... didn't know about that file ... reading21:15
jk0interesting, I thought HACKING said those lines SHOULD be there21:15
jk0unless I'm getting it confused with PEP821:15
edayjk0 line 10 should be there, the others no21:16
sandywalshjk0, I think your time cone is violated as well21:17
*** rlucio has quit IRC21:17
edayjk0: pep8 says the same thing, really21:17
sandywalshlast 'thou shalt' in HACKING21:17
jk0I think they look better separated :P21:18
edayjk0: the HACKING file just repeats it (although omits 3rd party dependency note)21:18
* jk0 fixes21:18
*** arcane has joined #openstack21:19
edayjk0: you can propose a change to the convention, but then you get to convert everything else :)21:19
jk0honestly I thought that's how it was being done the whole time21:19
*** BK_man has joined #openstack21:22
jk0eday: fixed pushed, LP is caught up21:22
*** ppetraki has quit IRC21:25
sandywalshWell, while we're at it. xs-pause is now merged with trunk and could use another review ... pulleaze21:28
*** hadrian has quit IRC21:29
*** miclorb_ has joined #openstack21:30
*** ppetraki has joined #openstack21:31
*** sophiap has quit IRC21:37
*** ctennis has quit IRC21:38
openstackhudsonProject nova build #307: SUCCESS in 1 min 12 sec:
*** hadrian has joined #openstack21:39
*** hazmat has quit IRC21:41
*** hazmat has joined #openstack21:42
sandywalshthat cloudinit thing looks interesting. Keen to read more.21:43
*** perestrelka has quit IRC21:47
*** EdwinGrubbs has joined #openstack21:50
*** lool has quit IRC21:51
*** aliguori has quit IRC21:51
*** jeevan_ullas has quit IRC21:52
*** jeevan_ullas has joined #openstack21:52
*** rlucio has joined #openstack21:54
*** kashyapc has quit IRC21:57
*** Ryan_Lane has quit IRC21:57
*** maplebed has quit IRC21:57
*** dragondm has quit IRC21:57
*** fabiand has quit IRC21:57
*** _0x44 has quit IRC21:57
*** dendrobates has quit IRC21:57
*** infernix has quit IRC21:57
*** uvirtbot has quit IRC21:57
*** larstobi has quit IRC21:57
*** localhost has quit IRC21:57
*** icarus902 has quit IRC21:57
*** zul has quit IRC21:57
*** termie has quit IRC21:57
*** stewart has quit IRC21:57
*** iRTermite has quit IRC21:57
*** cw has quit IRC21:57
*** letterj_ has quit IRC21:57
*** f4m8_ has quit IRC21:57
*** devcamcar has quit IRC21:57
*** sunech has quit IRC21:57
*** KnightHacker has quit IRC21:57
*** notmyname has quit IRC21:57
*** henrichrubin has quit IRC21:57
*** ccustine has quit IRC21:57
*** dirakx has quit IRC21:57
*** cory_ has quit IRC21:57
*** filler has quit IRC21:57
*** rossij has quit IRC21:57
*** sleepsonthefloor has quit IRC21:57
*** klumpie_ has quit IRC21:57
*** _cerberus_ has quit IRC21:57
*** redbo has quit IRC21:57
*** czajkowski has quit IRC21:57
*** tomo_bot_____ has quit IRC21:57
*** hazmat has quit IRC21:57
*** BK_man has quit IRC21:57
*** apanasenko has quit IRC21:57
*** joearnold has quit IRC21:57
*** lvaughn has quit IRC21:57
*** sandywalsh has quit IRC21:57
*** dagger has quit IRC21:57
*** michaeldreamhost has quit IRC21:57
*** ivan has quit IRC21:57
*** zykes- has quit IRC21:57
*** creiht has quit IRC21:57
*** pandemicsyn has quit IRC21:57
*** Xenith has quit IRC21:57
*** Adri2000 has quit IRC21:57
*** patri0t has quit IRC21:57
*** xtoddx has quit IRC21:57
*** ChrisAM has quit IRC21:57
*** morfeas has quit IRC21:57
*** brainproxy has quit IRC21:57
*** cclien has quit IRC21:57
*** schisamo has quit IRC21:57
*** jdurgin has quit IRC21:57
*** seshu has quit IRC21:57
*** Adiantum has quit IRC21:57
*** dubsquared has quit IRC21:57
*** gondoi has quit IRC21:57
*** karmabot` has quit IRC21:57
*** ioso has quit IRC21:57
*** bburhans has quit IRC21:57
*** miclorb_ has quit IRC21:57
*** hggdh has quit IRC21:57
*** ArdRigh has quit IRC21:57
*** daleolds has quit IRC21:57
*** deshantm has quit IRC21:57
*** jdarcy has quit IRC21:57
*** iammartian has quit IRC21:57
*** johnpur has quit IRC21:57
*** Podilarius has quit IRC21:57
*** bobl has quit IRC21:57
*** _skrusty has quit IRC21:57
*** deshantm_cosi has quit IRC21:57
*** burris has quit IRC21:57
*** Edwin-afk2 has quit IRC21:57
*** joshuamckenty has quit IRC21:57
*** davidX- has quit IRC21:57
*** annegentle has quit IRC21:57
*** almaisan-away has quit IRC21:57
*** cloud0_ has quit IRC21:57
*** taihen has quit IRC21:57
*** mtaylor has quit IRC21:57
*** [ack] has quit IRC21:57
*** hadrian has quit IRC21:57
*** pothos has quit IRC21:57
*** Abd4llA has quit IRC21:57
*** abecc has quit IRC21:57
*** odyi has quit IRC21:57
*** jbryce has quit IRC21:57
*** anticw has quit IRC21:57
*** nijaba has quit IRC21:57
*** openstackhudson has quit IRC21:57
*** timrc has quit IRC21:57
*** olivier_ has quit IRC21:57
*** jero has quit IRC21:57
*** PiotrSikora has quit IRC21:57
*** localhost has joined #openstack21:58
*** kevnfx has joined #openstack21:58
*** BK_man has joined #openstack21:58
*** lool has joined #openstack21:58
*** perestre1ka has joined #openstack21:58
*** hazmat has joined #openstack21:58
*** hadrian has joined #openstack21:58
*** miclorb_ has joined #openstack21:58
*** KnightHacker has joined #openstack21:58
*** notmyname has joined #openstack21:58
*** apanasenko has joined #openstack21:58
*** pothos has joined #openstack21:58
*** Abd4llA has joined #openstack21:58
*** schisamo has joined #openstack21:58
*** daleolds has joined #openstack21:58
*** joearnold has joined #openstack21:58
*** kashyapc has joined #openstack21:58
*** henrichrubin has joined #openstack21:58
*** jdurgin has joined #openstack21:58
*** ccustine has joined #openstack21:58
*** maplebed has joined #openstack21:58
*** Ryan_Lane has joined #openstack21:58
*** deshantm has joined #openstack21:58
*** larstobi has joined #openstack21:58
*** jdarcy has joined #openstack21:58
*** seshu has joined #openstack21:58
*** dirakx has joined #openstack21:58
*** dragondm has joined #openstack21:58
*** johnpur has joined #openstack21:58
*** dubsquared has joined #openstack21:58
*** Adiantum has joined #openstack21:58
*** bobl has joined #openstack21:58
*** gondoi has joined #openstack21:58
*** _skrusty has joined #openstack21:58
*** icarus902 has joined #openstack21:58
*** deshantm_cosi has joined #openstack21:58
*** burris has joined #openstack21:58
*** zul has joined #openstack21:58
*** termie has joined #openstack21:58
*** lvaughn has joined #openstack21:58
*** odyi has joined #openstack21:58
*** jbryce has joined #openstack21:58
*** sandywalsh has joined #openstack21:58
*** dagger has joined #openstack21:58
*** stewart has joined #openstack21:58
*** olivier_ has joined #openstack21:58
*** sets mode: +vv notmyname johnpur21:58
*** fabiand has joined #openstack21:58
*** anticw has joined #openstack21:58
*** karmabot` has joined #openstack21:58
*** joshuamckenty has joined #openstack21:58
*** cory_ has joined #openstack21:58
*** _0x44 has joined #openstack21:58
*** ioso has joined #openstack21:58
*** bburhans has joined #openstack21:58
*** nijaba has joined #openstack21:58
*** openstackhudson has joined #openstack21:58
*** timrc has joined #openstack21:58
*** jero has joined #openstack21:58
*** PiotrSikora has joined #openstack21:58
*** taihen has joined #openstack21:58
*** mtaylor has joined #openstack21:58
*** [ack] has joined #openstack21:58
*** annegentle has joined #openstack21:58
*** almaisan-away has joined #openstack21:58
*** cloud0_ has joined #openstack21:58
*** michaeldreamhost has joined #openstack21:58
*** ivan has joined #openstack21:58
*** Adri2000 has joined #openstack21:58
*** zykes- has joined #openstack21:58
*** creiht has joined #openstack21:58
*** pandemicsyn has joined #openstack21:58
*** sets mode: +vvvv _0x44 mtaylor creiht pandemicsyn21:58
*** Xenith has joined #openstack21:58
*** patri0t has joined #openstack21:58
*** xtoddx has joined #openstack21:58
*** ChrisAM has joined #openstack21:58
*** brainproxy has joined #openstack21:58
*** morfeas has joined #openstack21:58
*** cclien has joined #openstack21:58
*** filler has joined #openstack21:58
*** rossij has joined #openstack21:58
*** sleepsonthefloor has joined #openstack21:58
*** klumpie_ has joined #openstack21:58
*** _cerberus_ has joined #openstack21:58
*** redbo has joined #openstack21:58
*** czajkowski has joined #openstack21:58
*** tomo_bot_____ has joined #openstack21:58
*** iRTermite has joined #openstack21:58
*** cw has joined #openstack21:58
*** letterj_ has joined #openstack21:58
*** f4m8_ has joined #openstack21:58
*** devcamcar has joined #openstack21:58
*** sunech has joined #openstack21:58
*** sets mode: +vv _cerberus_ redbo21:58
*** dendrobates has joined #openstack21:58
*** infernix has joined #openstack21:58
*** uvirtbot has joined #openstack21:58
*** davidX- has joined #openstack21:58
*** sets mode: +v dendrobates21:58
*** ctennis has joined #openstack21:58
annegentleBK_man: yah, that's right, you're a CentOSer?21:59
annegentleIs that even a word?22:00
*** schisamo has quit IRC22:00
dubsquaredim working on a OS agnostic scrip, BK_man22:01
dubsquaredbut yeah, for now its 10.04/10.1022:01
termievishy: re input on the db/api stuff, yeah i have a bunch but i don't really know how to respond to what was said here22:01
BK_manannegentle: I'm creating my own port for RHEL6 and I'm almost done22:01
termievishy, eday: the statements were pretty vague "split stuff up before doing api stuff"22:01
termieso i don't know how you intend to split things or why it was deemed necessary22:02
BK_manannegentle: this is some specs from me that might be helpful for comminity22:02
annegentleman, Colloquy hates me today - no URLs opening for you, have to copy/paste... :)22:03
BK_manannegentle: I'm working for Grid Dynamics and we are going to announce RHEL6 build of Openstack Nova till the end of the year :)22:03
*** GasbaKid has quit IRC22:03
termiethe definition of eday's api plans is still a little foggy to me, and i don't like the continued usage of the word api to describe every part of the system22:03
*** GasbaKid has joined #openstack22:03
termieso i am hesitant to agree to changing some other part of the system in preparation for the plans that i don't see clearly22:04
*** aliguori has joined #openstack22:04
*** hggdh has joined #openstack22:04
termieall descriptions i've heard of them imply that it is a lot of work yet all imagination or experience i have with them says there isn't a lot to do and that what there is to do is relatively simple22:04
annegentleBK_man: shall I link to your script from the wiki page now, or wait?22:05
termieso there is a mismatch between what i feel people are talking about and what i can see happening22:05
termiehence the fogginess22:05
dubsquaredbk_man:  "we are going to announce RHEL6 build of Openstack Nova till the end of the year" what do you mean?22:05
BK_manannegentle: this is only home-made scripts and specfiles. I need some more time.22:06
edaytermie: when I mention API above, I mean APIs to compute/volume/auth/network, not the REST interface22:06
termieeday: i know, however some terminology needs to be decided upon that is not "api"22:06
termieeday: because we have at least 3 competing definitions22:07
edayvolume api seems pretty clear to me, but I'm open for something else22:07
BK_mandubsquared: My job is test Openstack. We have a Cisco UCS HW and they only support RHEL and not Ubuntu (as IBM, HP and other enterprise guys do the same). So I created a RHEL6 build of Openstack Nova and we are ready to put this job back to community22:07
termieannegentle: how is managed? there is a persistent bug related to it updating the "Latest" column22:07
eday(assuming we have the context we are not talking about public rest interfaces)22:07
termieannegentle: wondering where i can go to fix it or who i can call22:07
BK_mandubsquared: actually, not "test" Openstack, but evaluate that based on requirements for one of our customers22:08
*** dendrobates is now known as dendro-afk22:08
edaytermie: what I was trying to propose above is that we should have a well-defined interface into each of the internal services, ie, now nova-api talks to compute workers, and how compute workers talk to volume and network workers. I started some of this in nova.compute.api22:09
annegentletermie: I keep bugging our webmaster Todd about it :) He has a new website rollout planned so I think it will go away soon. I do appreciate you paying attention to it though, that sort of detail is important.22:09
dubsquaredcool, ill check out your stuff…i am working on a rhel CC as well22:09
dubsquaredid like to get some deb/rpm OSes CC/nodes playing with each other22:09
termieeday: aye, we've discussed it a bit so i am familiar with the components involved22:09
edaytermie: and we need a way for implementations of that interface to be able to use it's own data store, and not funnel everything back into nova.db22:09
edaytermie: because, in some ways, nova.db is just taking all the apis that have been split and shoving them all back together again. for example, the iscsi and ataoe should not need to be in nova.db, imo22:11
*** aliguori has quit IRC22:12
termieeday: the goal of nova.db was to cut out the many implementations of the same things we had at the time22:12
termieeday: that worked mostly alright, so its job is largely done in my opinion22:13
termieeday: but there is still a difference between the data layer interface and the intra-service interface22:14
edaytermie: yeah, and it makes sense for a more centralized system22:14
BK_mando I need nova-objectstore to run on compute node?22:15
*** jdarcy has quit IRC22:16
termieeday: it isn't really in any way dependent on centralzed vs dsitributed22:16
edaytermie:I'd like to black-box the services better. for example, I could start up a nova-volume service and it can manage it's own details (ataoe, iscsi, san, ...), and it only reports back the minimal info needed for routing and answering REST api requests. That data also needs to be stored within the scheduler/nova-api servers22:16
termieeday: we are only talking about pluggability22:16
termieeday: that is a separate issue22:16
edaytermie: so we end up with tables there, assuming we are in sql. one with generics, and one with implementation22:16
termieeday: this can be done in smaller steps22:16
edaytermie: well, it's not, since right now nova-api looks directly at volume tables that are implementation specific :)22:17
termieeday: first step is creating the inter-service interfaces22:17
jaypipesgod, I'm an idiot...22:17
edaytermie: I know it can be done in smaller steps, and thats what I'm doing, I just want to make sure the end goal is the same22:17
*** pothos_ has joined #openstack22:18
termieeday: the end goal will become more apparent as the system grows22:18
termieeday: and changes22:18
*** schisamo has joined #openstack22:18
termieeday: i just think you are getting a bit ahead of things22:18
termieeday: the goal is to black box things, yes, a different goal is to provide different data backends22:18
edaytermie: not really, we have specific goals to meet in the next couple months, and this is how I see we'll get there. I don't want to be working towards too much of a moving target, and so far we've not really had any consensus on where that is22:19
termieeday: that's because it is still too abstract22:19
edaythats why I've been thinking about this the past month so we have something concrete to work towards22:20
termieeday: nobody is going to (usefully) agree on something that can't be seen yet22:20
*** pothos has quit IRC22:20
*** pothos_ is now known as pothos22:20
termieeday: i think we're agreed on the direction22:20
termieeday: i don't know that you and i have agreed on ordero f operations22:21
edaytermie: if I start proposing branches that start to refactor nova.db, folks probably won't agree without seeing the bigger picture of where it's going. thats what 'm trying to get out of the way now :)22:21
termieeday: i won't agree at all because there is no need to refactor it yet22:21
termieeday: that comes after a system to take advantage of it is in place22:22
edayand to see if it is indeed the best way to go. I have my solution, but I want other's feedback22:22
termiewell you got a bunch of feedback from us last friday and we were relatively in agreement as i understood it22:22
edayI did this at the summit, but it seems a number of folks were not there and now disagree with how I've described it22:22
termiei think you are describing too much, you really need to define this as smaller problems22:23
*** vvuksan has quit IRC22:23
termiei think everything is fine and is on track22:23
*** BK_man_ has joined #openstack22:23
termiemake the inter-service interfaces22:23
*** BK_man has quit IRC22:24
*** BK_man_ is now known as BK_man22:24
termieonce people see how that all works they will ahve a much easier time agreeing or giving feedback on db stuff22:24
edayI'm trying to do this because I've wasted 3 days throwing away branches :)22:25
termiethe compute.api branch seems to have made it in22:26
edaythere are some chicken-egg problems if we want to design the API properly22:26
termienot really22:26
termiewe can change things22:26
*** dendro-afk is now known as dendrobates22:26
termiethings don't have to be correct the first time22:26
edayI know22:26
termiecode can be thrown away when it is no longer useful22:27
termieetc etc22:27
termiei don't know who is making you throw away branches, but surely anybody involved has a better understanding of things after those branches22:27
edayI was hoping to get a bit more agreement on the details of where we are going, because I don't feel there has been any agreement on details22:27
termieeday: do you have a blueprint et al?22:28
edayU'm throwing them away, because of dead-ends. they are useful because thigns are more clear, but before I went ahead with the plan I had I just wanted feedback, is all22:28
edayyeah, distributed-scheduler22:29
edayalthough at this point I probably need to update it22:29
termieeday: that doesn't seem to apply to inter-service interfaces22:29
edaybut before drilling into the details I was hoping to get a first pass agreement22:29
termieyou have it, but don't think that means people won't decide you were wrong later22:30
edaytermie:  first bullet under Design22:30
edayobviously this isn't working out how I had hoped, so I'll jsut start doing things the way I think it needs to be done and we can discuss in the merge reqs22:31
BK_manquick question about instancemonitor - which kind of rrdtool python module it requires?22:32
termieeday: your spec feels a bit hand-wavey :/22:32
edaytermie: it is, because of lack of agreement on details22:33
edayhence this discussion22:33
termieeday: well a spec is a good way to express details22:33
BK_manthere are two modules - one which comes rrdtool package and there is also other one:
termieeday: nobody is going to read a discussion to decide whether they should agree to something22:33
termieeday: but things like "Introduce primitives that will allow for data aggregation from workers up through scheduling layers and API nodes into the core Nova classes." aren't really detailed enough for me to have an opinion one way or the other22:34
termieeday: sort of like, it's not dis-provable22:34
termieeday: the statement is by definition correct because it doesn't make any assertions22:34
edaywell, we discussed thigns in more detail at the summit22:34
*** blpiatt has joined #openstack22:35
edayand I know thats not a replacement for a spec, but if you had that context it would make more sense22:35
*** sophiap has joined #openstack22:35
termieeday: sure but honestly a tight design doc will catch more flies than honey or whatever22:35
termieeday: i have entered metaphor world22:35
termieor analogy22:35
*** reldan has joined #openstack22:36
edayI suppose I should expand the spec a bit more, I'm tired of having the same old conversation everytime someone new enters the discussion :)22:37
termieeday: that's almost the only reason i spend so much time on docs22:37
edaythe assumed context keeps varying too much22:37
termieeday: but it is almost always worth it22:37
*** hadrian has quit IRC22:37
termieeday: and you can shut people up by telling them to read the doc22:37
edayafter the summit discussion, it was pretty clear I thought, and every there was in agreement, but then some folks missed and were were back to square one22:37
termieeday: and things changed too, as 3 failed branches would likely demonstrate22:38
termieeday: i wasn't there so i don't know what exactly changed, but i don't trust things that are right the first time22:38
edaytermie: the end goal is still the same really, the approach has changed some22:39
termieeday: i for one feel like everybody has had the same end goal from day one, its only the approach and order of events that needed to be figured out22:40
edayI would disagree with that now22:40
edayI think the end goal is still different for folks22:40
termieeday: they just aren't thinking big enough ;)22:41
termieeday: i guess i'd say beef up your design doc, i think you'll get a lot of mileage out of it22:42
termieand i promise to read it22:42
termieand argue about specifics as they are described therein22:43
*** dragondm has quit IRC22:43
edaythats what I'm doing, I was just hoping to avoid having to have the in-person summit discussion again over email/irc/whatever :)22:44
*** dragondm has joined #openstack22:44
*** aliguori has joined #openstack22:46
termieeday: btw, it is way easier to discuss complex thigns with you know that i can imagine your demeanor when you say things22:46
*** gondoi has quit IRC22:49
*** gondoi has joined #openstack22:50
edaytermie: heh, good thing you didn't see me cursing your name for the past 30 minutes :)22:51
termieeday: i'm used to it22:51
edaytermie: btw, why did you miss the summit?22:52
termieeday: i was in japan, for one, and financials fell through22:52
*** BK_man_ has joined #openstack22:54
*** BK_man has quit IRC22:54
*** BK_man_ is now known as BK_man22:54
*** hadrian has joined #openstack22:57
*** ivan has quit IRC23:05
*** ivan has joined #openstack23:05
*** ppetraki has quit IRC23:06
* vishy runs off to port a couple things from deploy23:11
*** hggdh has quit IRC23:14
*** reldan has joined #openstack23:16
*** hggdh has joined #openstack23:16
*** DigitalFlux has joined #openstack23:18
edayvishy: where is the counter ever reset in your ec2 lockout branch?23:20
edayahh, I see, setting a ttl on the memcache key23:21
*** larstobi has quit IRC23:21
vishyeday: correct23:22
*** BK_man_ has joined #openstack23:25
*** BK_man has quit IRC23:25
*** BK_man_ is now known as BK_man23:25
*** lool has quit IRC23:26
*** lool has joined #openstack23:26
*** GasbaKid has quit IRC23:26
*** rds__ has joined #openstack23:27
*** seshu has left #openstack23:28
*** schisamo has quit IRC23:30
*** schisamo has joined #openstack23:32
*** johnpur has quit IRC23:35
*** apanasenko has quit IRC23:36
*** Abd4llA has quit IRC23:47
*** reldan has quit IRC23:53
*** BK_man_ has joined #openstack23:55
*** BK_man has quit IRC23:55
*** BK_man_ is now known as BK_man23:55
*** joearnold has quit IRC23:56
*** joearnold has joined #openstack23:56
vishygflags annoyance23:59

Generated by 2.14.0 by Marius Gedminas - find it at!