Monday, 2010-09-13

*** rackerhacker_ is now known as rackerhacker00:07
*** rackerhacker has joined #openstack00:07
*** gazumps has quit IRC00:10
*** ptremblett has quit IRC00:14
*** zooko has quit IRC00:16
*** jfluhmann_ has joined #openstack00:20
*** sirp1 has quit IRC00:21
*** gazumps has joined #openstack00:37
*** Abd4llA has quit IRC00:42
*** e1mer has joined #openstack00:54
*** zheng_li has quit IRC00:56
*** zooko has joined #openstack01:24
*** amscanne_ has quit IRC01:24
*** sirp1 has joined #openstack01:26
*** kirkland has joined #openstack01:26
*** zooko has quit IRC01:30
*** DubLo7 has joined #openstack01:34
*** kirkland has quit IRC01:53
*** kirkland has joined #openstack01:54
*** sophiap has quit IRC01:57
*** silassewell has joined #openstack02:01
*** sophiap has joined #openstack02:08
*** Glacee has quit IRC02:12
*** Glacee has joined #openstack02:14
*** e1mer has quit IRC02:23
*** ArdRigh has joined #openstack02:26
*** Cybo has joined #openstack02:28
*** sirp1 has quit IRC02:39
*** e1mer has joined #openstack02:48
*** sophiap has quit IRC03:01
*** sophiap has joined #openstack03:09
*** Cybodog has quit IRC03:19
*** sophiap has quit IRC03:20
*** jfluhmann_ has quit IRC03:20
*** gazumps has quit IRC03:21
*** silassewell1 has joined #openstack03:29
*** silassewell has quit IRC03:30
*** jfluhmann_ has joined #openstack03:30
*** kashyapc has joined #openstack03:32
*** johndoy_ has joined #openstack03:33
*** zooko has joined #openstack03:36
*** johndoy__ has quit IRC03:36
*** fmontezuma has joined #openstack03:37
*** jfluhmann_ has quit IRC03:37
*** herki_ has joined #openstack03:38
*** sirp1 has joined #openstack03:39
*** herki has quit IRC03:41
*** PiotrSikora has quit IRC03:41
*** PiotrSikora has joined #openstack03:45
*** zooko has quit IRC03:59
*** fmontezuma has quit IRC04:03
*** miclorb_ has quit IRC04:04
*** ArdRigh has quit IRC04:09
*** gazumps has joined #openstack04:18
*** gazumps has quit IRC04:23
*** omidhdl has joined #openstack04:25
*** gazumps has joined #openstack04:27
*** gazumps has quit IRC04:37
*** Mr_T has joined #openstack04:50
*** littleidea has joined #openstack04:53
littleideavishy: ping04:55
littleideavishy: when you can, jump in the 'Intermittent test failures' email thread04:58
*** f4m8_ is now known as f4m805:02
*** silassewell1 has quit IRC05:05
*** kashyapc has quit IRC05:09
*** kashyapc has joined #openstack05:10
*** miclorb_ has joined #openstack05:36
*** sirp1 has quit IRC05:40
*** coredump|br has quit IRC05:52
*** zooko has joined #openstack05:56
*** kashyapc has quit IRC06:05
*** ibarrera has joined #openstack06:05
*** kashyapc has joined #openstack06:08
*** omidhdl has left #openstack06:16
*** zooko has quit IRC06:21
*** omidhdl has joined #openstack06:31
*** debauer__ has joined #openstack06:46
*** debauer_ has quit IRC06:49
sorenvishy: Still here, by any chance?07:00
*** allsystemsarego has joined #openstack07:09
*** allsystemsarego has joined #openstack07:09
*** Cybo has quit IRC07:11
*** zooko has joined #openstack07:12
*** zooko has quit IRC07:18
*** calavera has joined #openstack07:35
*** zheng_li has joined #openstack07:43
*** dele_ted has joined #openstack07:46
*** calavera has quit IRC07:59
*** brd_from_italy has joined #openstack07:59
*** omidhdl has quit IRC07:59
*** calavera has joined #openstack07:59
*** miclorb_ has quit IRC08:08
*** omidhdl has joined #openstack08:19
*** dele_ted has quit IRC08:25
*** DogWater has quit IRC08:26
*** omidhdl has quit IRC08:27
*** omidhdl has joined #openstack08:30
*** DogWater has joined #openstack08:30
*** debauer__ has quit IRC08:32
*** debauer__ has joined #openstack08:32
*** littleidea has quit IRC08:47
*** michalis has joined #openstack09:03
*** miclorb has joined #openstack09:12
*** joshuamckenty has quit IRC10:07
*** sophiap has joined #openstack10:14
*** omidhdl has left #openstack10:48
*** gustavomzw has joined #openstack10:56
*** joshuamckenty has joined #openstack10:57
*** coredump|br has joined #openstack10:57
*** coredump|br has quit IRC10:57
*** michalis has quit IRC11:10
*** DubLo7 has quit IRC11:20
*** ctennis has quit IRC11:20
*** joshuamckenty has quit IRC11:22
*** DubLo7 has joined #openstack11:27
*** DubLo7 has quit IRC11:37
*** joshuamckenty has joined #openstack11:48
*** miclorb has quit IRC11:50
*** ctennis has joined #openstack11:53
*** KanGouLya has joined #openstack11:55
*** michalis has joined #openstack12:05
*** Cybodog has joined #openstack12:36
* joshuamckenty hears crickets...12:40
creihtgood morning12:40
joshuamckentyhow goes it?12:40
creihtit goes12:40
joshuamckentyWhen the tests pass, it's a good day, right?12:41
*** hisak has joined #openstack12:44
*** kashyapc has quit IRC13:00
*** gundlach has joined #openstack13:08
*** Cybodog has quit IRC13:19
*** Cybodog has joined #openstack13:20
*** abecc has joined #openstack13:21
*** Cybodog has quit IRC13:21
*** jaypipes has joined #openstack13:22
*** aliguori has joined #openstack13:22
*** Cybodog has joined #openstack13:23
*** kashyapc has joined #openstack13:26
*** pcrews has joined #openstack13:26
*** pcrews has quit IRC13:26
*** Cybodog has quit IRC13:30
*** Cybodog has joined #openstack13:31
*** e1mer has quit IRC13:35
*** kashyapc has quit IRC13:35
gundlachsirp: i assume you know that in addition to your reply to my merge prop for , I'm waiting for you to push more changes to the branch so I can sign off on it.13:37
gundlach(last thing i see is a comment from you on 9/10, but no code.)13:37
*** jkelly has joined #openstack13:39
*** jkelly is now known as jonkelly13:39
*** kashyapc has joined #openstack13:52
*** jdarcy has joined #openstack13:59
*** f4m8 is now known as f4m8_14:07
*** mdomsch has joined #openstack14:21
*** brd_from_italy1 has joined #openstack14:24
*** brd_from_italy has quit IRC14:25
*** zooko has joined #openstack14:29
*** pharkmillups has joined #openstack14:41
*** Podilarius has joined #openstack14:43
*** zooko has quit IRC14:45
*** jkakar has joined #openstack14:48
*** sirp1 has joined #openstack14:55
burrisare the openstack docs, particularly swift, licensed under the same apache license as the code?14:55
*** Cybodog has quit IRC14:59
*** Cybodog has joined #openstack15:00
burrisoh they are in the same repo so the answer is yes, I think I was getting them confused with the Rackspace cloudfiles docs which are in a repo on github with no license information15:01
*** vvuksan has joined #openstack15:04
*** hisak has left #openstack15:08
creihtburris: the specific licensing for the docs has yet to be determined15:08
creihtdendrobates: Did we ever make any progress on figuring out what we wanted to license the docs as?15:09
creihtburris: and I think the cloudfiles docs were a cc license15:09
burriswell, they're published in a public repo with the apache license15:10
*** johnpur has joined #openstack15:11
*** ChanServ sets mode: +v johnpur15:11
creihtthat part says they are CC licensed15:12
dendrobatescreiht: I talked to legal at RS and they thought apache was fine, but are looking to make sure.15:12
creihtburris: thanks for pointing that out15:12
dendrobatesI think the sphinx stuff has to be apache, since it is pulled from source that is apache licensed15:13
creihtdendrobates: it has always seemed weird to me that documentation would be licensed with a code license15:14
burristhat would be the .rst files in the swift repo, since they are pulled from the docstrings?15:14
creihtthe .rst files are created by hand, the code docs are autogenerated15:14
*** pharkmillups has quit IRC15:15
*** pharkmillups has joined #openstack15:15
creihtwell I guess apache licenses their docs with the apacke license15:20
*** calavera has quit IRC15:23
*** jonkelly has quit IRC15:30
burrishmm, the python cloudfiles lib lets you supply an authurl keyword arg to specify your own auth server but the ruby lib is hardcoded to :-(15:34
*** davidg has joined #openstack15:35
creihtburris: that is unfortunate15:39
creihtburris: if you don't mind posting a bug for that, and I will pass it along to the folks that are responsible for the bindings15:40
burrisI opened a ticket on github I guess I could try to make a patch but I'm not much of a rubyist15:40
burrisif I need it I'll make a patch, right now I'm just investigating15:40
creihtburris: do you have a link for the ticket?15:41
vvuksanspeaking of hardcoding15:41
creihtburris: cool... I'll forward that on15:41
vvuksannova/ has this15:42
vvuksancsock.connect(('', 80))15:42
vvuksanwhen it tries to determine it's own IP15:42
vvuksani caught it when I didn't have network connectivity while working on a train :-/15:43
vvuksanalthough I suspect in some places that may be firewalled off15:43
burrisi don't think there is any reliable way to determine your external ip without using a trusted host on the outside or getting a consensus from a bunch of connected peers15:43
vvuksanone of the daemons has to talk to either redis and rabbitmq15:44
vvuksana daemon15:44
vvuksancause all that thing does is open up a TCP socket than determines what IP it connected as15:44
vvuksanalthough in general you should just look at the iface that is specified as default route15:45
vvuksanalso accessing google is prone with problems15:46
*** jkakar has quit IRC15:47
vvuksanas in security problems15:47
vvuksansince if I had a security policy and noticed that bunch of my hosts were connecting to Google15:47
burrisphp-cloudfiles has an auth_host paramater but it is marked "deprecated"15:47
vvuksanI'd be slightly suspicious15:47
burrisvvuksan, not saying its a good idea in this context or anything, also it's the first time I've seen someone use google for this purpose15:48
creihtburris: Someone should be fixing the ruby bindings in the near future, for at least some measure of "near"15:55
*** jrussell has joined #openstack16:01
*** dendrobates is now known as dendro-afk16:02
*** jbryce has joined #openstack16:08
burrisI created an issue for php-cloudfiles to urge them to un-deprecate auth_host paramater...16:09
burrisjava uses a .properties file for the auth host so it should be OK16:09
creihtburris: cool, I passed that along as well16:10
burrisoops, missed c#16:10
*** dendro-afk is now known as dendrobates16:12
*** brd_from_italy1 has quit IRC16:13
*** littleidea has joined #openstack16:21
*** ibarrera has quit IRC16:25
*** littleidea has quit IRC16:28
*** jkelly has joined #openstack16:33
*** kashyapc has quit IRC16:34
*** xtoddx has joined #openstack16:38
*** vvuksan has quit IRC16:38
*** burris has quit IRC16:39
*** vvuksan has joined #openstack16:43
*** mdomsch has quit IRC16:43
*** dele_ted has joined #openstack16:48
*** vvuksan has quit IRC16:50
*** kashyapc has joined #openstack16:51
*** stewart has joined #openstack16:54
*** dele_ted has quit IRC16:56
*** joschi has quit IRC16:59
*** joschi has joined #openstack16:59
*** brd_from_italy has joined #openstack17:01
*** zheng_li has quit IRC17:04
*** ctennis has quit IRC17:05
*** maplebed has joined #openstack17:05
*** michalis has quit IRC17:09
*** maplebed has quit IRC17:18
*** burris has joined #openstack17:19
*** rnewson has joined #openstack17:24
*** maplebed has joined #openstack17:27
*** mdomsch has joined #openstack17:34
*** zheng_li has joined #openstack17:34
*** rnewson has quit IRC17:34
*** amscanne has joined #openstack17:42
*** mdomsch_ has joined #openstack17:43
*** joearnold has joined #openstack17:45
*** mdomsch has quit IRC17:47
*** amscanne_ has joined #openstack17:49
*** amscanne has quit IRC17:52
*** jakedahn_ has joined #openstack17:54
*** dendrobates has quit IRC17:57
*** dendrobates has joined #openstack17:57
*** ChanServ sets mode: +v dendrobates17:57
*** jakedahn has quit IRC17:57
*** jakedahn_ is now known as jakedahn17:57
*** gasbakid has joined #openstack17:59
*** rnewson has joined #openstack18:03
*** pharkmillups has quit IRC18:06
*** rlucio has joined #openstack18:12
*** sophiap has quit IRC18:22
rluciohey guys, i was wondering, is there any one place to get more info on setting up nova networking ?  Specifically the non-flat networking18:23
*** zooko has joined #openstack18:39
*** kashyapc has quit IRC18:41
dendrobatesrlucio: all we have right now is which is from the docstrings in the source18:41
dendrobatesI'm afraid it may not be what you are looking for.18:41
*** dendrobates has quit IRC18:41
gundlachwhen a bug has its fix merged into trunk (e.g. ), do i just mark it as 'Fix Committed' and that means 'Closed'?  Or is there some other way to mark it as closed?18:41
uvirtbotLaunchpad bug 613175 in nova "Support Flavors operations in API" [Medium,Fix committed]18:41
gundlachthanks uvirtbot18:42
gundlachit's still listed as "related to " but I don't see a way to filter that list only to open bugs18:42
*** dendrobates has joined #openstack18:43
*** ChanServ sets mode: +v dendrobates18:43
rluciodendrobates: unfortunately that doesn't really help atm, its ok I've been looking at the network source code to extract the configuration details18:43
rluciodendrobates: thanks though18:44
edaygundlach: it should be changed to fix released when it hits trunk18:49
edaygundlach: fix comitted is when it's fixed and waiting to merge18:49
gundlacheday: ty -- 'released' sounded like perhaps a milestone release18:49
gundlachbtw i'm about to take a look at rate limiting in the RS API -- lemme know if you are already working on that :)18:50
edaygundlach: and there may be some debate on those states, ie, committed to trunk, and then released is for when it makes it to a package, but I used the above flow :)18:50
*** pharkmillups has joined #openstack18:51
*** Podilarius has left #openstack18:53
*** Podilarius has joined #openstack18:56
*** sophiap has joined #openstack18:56
*** joearnold has quit IRC18:58
*** amscanne_ has quit IRC18:59
*** lool- has joined #openstack19:07
*** lool- is now known as lool-webchat19:07
*** brd_from_italy has quit IRC19:10
*** jkelly has quit IRC19:12
*** amscanne has joined #openstack19:13
*** zooko has quit IRC19:14
*** zooko has joined #openstack19:14
vishyrlucio: the vlan networking isn't too hard to set up19:17
vishybut there are a few gotchas19:17
vishylet me know if you need some help19:17
*** amscanne has quit IRC19:18
*** amscanne has joined #openstack19:18
*** bitmonk has joined #openstack19:21
rluciovishy: cool, thanks.  im still looking over the source, but i'll ping you if things get too bizarre :)19:21
vishyrlucio: the networking code in orm_deux is cleaned up a lot and probably easier to understand19:24
*** littleidea has joined #openstack19:53
*** tobym has joined #openstack19:55
*** sophiap has quit IRC19:56
*** sophiap has joined #openstack19:58
*** joearnold has joined #openstack20:11
*** burris has quit IRC20:15
*** joearnold has quit IRC20:16
*** joearnold has joined #openstack20:17
*** burris has joined #openstack20:19
*** jaypipes has quit IRC20:26
*** miclorb_ has joined #openstack20:31
*** jaypipes has joined #openstack20:33
*** miclorb_ has quit IRC20:35
*** miclorb has joined #openstack20:39
*** jrussell has quit IRC20:44
*** rnewson has quit IRC20:55
*** DubLo7 has joined #openstack21:01
*** sophiap has quit IRC21:03
*** joearnold has quit IRC21:04
*** mdomsch_ has quit IRC21:19
*** amscanne_ has joined #openstack21:19
*** amscanne has quit IRC21:23
*** joearnold has joined #openstack21:24
*** jbryce has quit IRC21:26
*** jbryce has joined #openstack21:28
*** anm_ has joined #openstack21:29
*** ctennis has joined #openstack21:39
*** metcalfc has joined #openstack21:46
*** sophiap has joined #openstack21:46
*** joearnol_ has joined #openstack21:50
*** allsystemsarego has quit IRC21:52
*** joearnold has quit IRC21:53
*** sophiap has quit IRC21:57
*** miclorb has quit IRC21:57
*** kindjal has joined #openstack22:05
*** amscanne_ has quit IRC22:14
sorenvishy: I'm wondering why a lot of the relationships are commented out in the SQLAlchemy models.22:16
*** johnpur has quit IRC22:18
sorenvishy: I'm also wondering whether there's a reason we're not using scoped_session.22:18
sorenvishy: I'm having annoying problems when grabbing objects out of the db layer and using them for input to the db layer again. They're attached to different sessions, and that makes SQLAlchemy, and by extension me, very sad.22:20
sorenvishy: Wrapping our call to sessionmaker in scoped_session fixes that, but I'm wondering there are problems with it that I just haven't spotted.22:20
uvirtbotNew bug: #637621 in swift "invalidate hashes on quarantine" [High,New]
sorenOh, it's swift.22:21
*** joearnold has joined #openstack22:22
*** joearnol_ has quit IRC22:22
*** kindjal has quit IRC22:23
*** blpiatt has joined #openstack22:31
vishysoren: I tried that, but I was worried about it interacting badly with multiple threads22:33
*** ArdRigh has joined #openstack22:33
vishysoren: i think it is a bad idea to pass datalayer objects back22:33
vishysoren: the idea is that the data layer returns dictionaries22:33
vishysoren: but i purposely wrapped in get_session in case we want to switch to using scoped session22:34
*** zheng_li has quit IRC22:47
vishysoren: the idea of the _update family of commands is that you always pass a dictionary for update rather than editing the object and passing it back.  I just added an iterator to NovaBase so we can wrap objects in dict(), and I think ideally all of the sqlalchemy/ calls should be returning actual dictionaries instead of dictionary-like objects if it isn't too much of a performace hit22:50
*** Podilarius has left #openstack22:51
edayvishy, soren: we should be careful not to reinvent half of sqlalchemy in the abstraction layer, might be just fine to eventually sqlalchemy directly. If we want relational data models, we'll never get their full use if we want to keep a non-relational abstraction on top22:53
*** joearnold has quit IRC22:56
vishyeday: that is a good point.22:56
*** silassewell has joined #openstack22:56
PythonPupSwift is reinventing what is done in some advanced filesystems.  I don't see that as a bad thing.  Sometimes, the reinvention lets you get what you really need.22:57
vishyeday, soren: there is an issue that we need to deal with which is quickly getting linked objects back from the datalayer23:01
vishyi think termie's original idea is that you have commands to retrieve linked objects, and they can be cached at the datalayer if you need speed.  I'm ok with nested dictionaries and/or dictionaries with members that are lists.23:02
vishyfor example instances have zero or one fixed ips23:02
vishyand fixed ips can have 0 or more floating ips associated with them23:02
vishyin the relational world it makes sense to grab the fixed ip and all the floating ips in one query23:03
vishybut how do we get that info out of the data layer?  This is where having business objects is an advantage, because you can define the linked objects clearly.23:03
edayvishy: yeah, and sqlalchemy already does that quite nicely with optimizations as needed, which we would just be rewriting23:05
*** joearnold has joined #openstack23:06
*** miclorb has joined #openstack23:08
vishyeday: i suppose, but that means any other backend needs to implement the entire (very generalized) sqlalchemy code instead of very specific datalayer code23:08
edayvishy: yeah, but I think we need to make some decisions around what backends we want to support. pg/mysql/sqlite may be sufficient. I'm honestly not sure what else we would really want to use if we want relational-like data models. possibly cassandra, but that may be overkill since we don't really have big data needs for this piece, even for large installations23:12
edayI'm all for doing the extra work to support extra data stores (ie, ones not supported by sql alchemy) if there is a sound technical reason, I just don't think we're really dug into that yet23:14
vishyeday: the mailing list is big on keeping redis support.  I think implementing a redis backend is easy for our limited set of commands.23:15
vishyeday: much more difficult for all of SQLAlchemy23:15
edayI'm really not sure why people want redis, it doesn't seem like a good choice for this application :)23:16
edayagreed would be much easier with our own layer of sqlA23:16
edayerr, over SQLAlchemy23:16
vishyeday: yes, I think our data is pretty ideal for relational databases, but redis is cool man :)23:17
edayit is cool, but cool is not reason enough for the extra work IMO :)23:18
vishyeday: agreed, but coolness has its value as well, especially when you're trying to attract developers/ gain adoption.23:19
edayvishy: I think there is enough other substance to attract devs/adoption that supporting other intersting technologies.23:20
vishyeday: will you be around for the meeting tomorrow?23:20
vishyeday: orm_deux is the main topic of discussion I think.23:21
edayalso, when it comes to tech minded people, they're going to questions these choices.  I was confused at the choice of Redis over some data store built for persistence first. Persistence in redis is more of an afterthough with snapshots to disk (ie, save your cache so it's warm on boot)23:22
vishyeday: personally I'm happy with a clean db abstraction layer, where we get back linked objects or dictionaries.  I'm used to a bit more OO interface, but I think simple data transfer objects are a good idea.23:22
edayyup, I'll be around for the meeting23:23
vishyeday: cool23:23
edayit's sometimes a tricky balance when trying to find a healthy level of abstraction, just like other tradeoffs you need to make23:24
*** gasbakid has quit IRC23:26
*** tobym has quit IRC23:28
*** pharkmillups has quit IRC23:31
*** hisak has joined #openstack23:32
*** tobym has joined #openstack23:32
*** gundlach has quit IRC23:42
*** burris has quit IRC23:49
*** burris has joined #openstack23:52
*** rlucio has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!