Thursday, 2010-10-21

*** jc_smith has quit IRC00:01
*** jc_smith has joined #openstack00:01
*** jc_smith has quit IRC00:02
uvirtbotNew bug: #664236 in swift "st tool will not error if you try to add ACLs to an object" [Undecided,New]
*** joearnold has joined #openstack00:09
*** pvo has quit IRC00:11
*** joearnold has quit IRC00:14
*** sirp1 has joined #openstack00:15
*** dendro-afk is now known as dendrobates00:16
*** sirp has quit IRC00:17
*** hazmat has joined #openstack00:20
jaypipeseday: yes, agreed.00:27
jaypipeseday: (sorry, had dinner :) )00:28
dendrobatesjaypipes eday: do you guys have any critical bugs outstanding?00:32
dendrobatesI am about to declare trunk frozen00:33
dendrobatesmtaylor around?00:33
dendrobatesjaypipes:  can you look at annegentle's merge again, she made the fixes you asked for.00:34
jaypipesdendrobates: no, not on my plate.00:38
jaypipesdendrobates: looking at annegentle's branch right now.00:38
*** hazmat has quit IRC00:39
jaypipesdendrobates: I've approved annegentle's merge prop.00:40
*** schisamo has quit IRC00:40
*** sirp1 has quit IRC00:46
*** joearnold has joined #openstack00:48
*** daleolds has quit IRC00:54
*** joearnold has quit IRC00:56
*** pharkmillups has joined #openstack00:56
*** blpiatt has joined #openstack01:00
*** ar1 has joined #openstack01:02
*** jakedahn has joined #openstack01:07
*** jakedahn has joined #openstack01:07
*** dysinger has joined #openstack01:15
*** silassewell has joined #openstack01:26
*** dysinger has quit IRC01:26
*** DubLo7 has joined #openstack01:36
*** sophiap has quit IRC01:37
*** ar1 has quit IRC01:43
*** ar1 has joined #openstack01:43
*** kevnfx has joined #openstack01:44
*** kevnfx has left #openstack01:44
*** rnirmal has joined #openstack01:51
*** pvo has joined #openstack01:52
*** ChanServ sets mode: +v pvo01:52
*** daragh has joined #openstack01:58
*** daragh has quit IRC02:00
*** blpiatt has quit IRC02:04
*** rnirmal has quit IRC02:05
*** Ryan_Lane|away is now known as Ryan_Lane02:09
*** rnirmal has joined #openstack02:30
*** khussein has joined #openstack02:37
dendrobatesI really need someone to test trunk.02:39
dendrobatesI am seeing instances stuck in "scheduling", but it might be my environment, I'm testing in the cloud02:40
*** metcalfc has quit IRC02:42
*** dendrobates is now known as dendro-afk02:45
*** jbryce has joined #openstack02:46
*** dendro-afk is now known as dendrobates02:49
*** sirp has joined #openstack02:53
*** DubLo7 has quit IRC02:57
*** DubLo7 has joined #openstack03:02
*** jbryce has quit IRC03:03
dendrobatesany devs around?03:10
dendrobatesbug  #66042803:11
uvirtbotLaunchpad bug 660428 in nova "nova-objectstore start errors after install" [Medium,Triaged]
*** DubLo71 has joined #openstack03:19
*** DubLo7 has quit IRC03:19
*** kashyapc has joined #openstack03:52
*** DubLo71 has quit IRC03:54
*** DubLo7 has joined #openstack04:07
*** Dweezahr has quit IRC04:07
*** rnirmal has quit IRC04:07
*** silassewell has quit IRC04:08
*** Dweezahr has joined #openstack04:08
vishydendrobates: Are all instances stuck in scheduling?  anything in the logs?04:11
*** kashyapc has quit IRC04:12
*** DubLo7 has quit IRC04:13
dendrobatesI am traveling and don't have my usual test env, testing in the cloud has it's issues04:13
vishybut it should get to launching04:13
vishyeven if the launch fails04:13
vishyunless compute is crashing on load04:14
vishydo the scheduling instances have a host assigned?04:14
vishydendro i'lll try an install tonight in a vm, and we'll try some on hardware tomorrow04:18
vishywhere do I get the rc packages?04:18
dendrobates click the download link04:19
dendrobatesbut that does not include todays merges.04:19
dendrobatesI deleted my vm and am trying to reproduce my problem04:20
*** DubLo7 has joined #openstack04:20
*** pvo has quit IRC04:28
*** jarrod_ has joined #openstack04:29
*** jarrod has quit IRC04:31
*** kashyapc has joined #openstack04:39
*** blamar has quit IRC04:54
*** blamar has joined #openstack04:55
*** suchitp has joined #openstack04:56
*** suchitp has left #openstack04:56
*** Ryan_Lane is now known as Ryan_Lane|away05:09
*** gaveen has quit IRC05:14
*** BK_man has joined #openstack05:14
*** gaveen has joined #openstack05:15
*** ArdRigh has quit IRC05:15
*** almaisan-away is now known as al-maisan05:21
*** sirp has quit IRC05:23
*** gaveen has quit IRC05:35
*** ibarrera has joined #openstack05:55
*** arthurc has joined #openstack06:28
*** perestrelka has quit IRC06:37
*** perestrelka has joined #openstack06:50
*** guigui has joined #openstack06:59
*** BK_man has quit IRC07:02
*** miclorb has quit IRC07:18
*** calavera has joined #openstack07:34
*** befreax has joined #openstack07:35
*** befreax has quit IRC07:50
*** guigui has quit IRC07:51
*** kashyapc has quit IRC07:53
*** ar1 has quit IRC07:53
*** eday has quit IRC07:53
*** tomo_bot has quit IRC07:53
*** coredump|br has quit IRC07:53
*** PiotrSikora has quit IRC07:53
*** jakedahn has quit IRC07:53
*** karmabot has quit IRC07:53
*** exlt has quit IRC07:53
*** letterj has quit IRC07:53
*** _cerberus_ has quit IRC07:53
*** [ack]_ has quit IRC07:53
*** nevele has quit IRC07:53
*** jakedahn has joined #openstack07:55
*** karmabot has joined #openstack07:55
*** exlt has joined #openstack07:55
*** letterj has joined #openstack07:55
*** _cerberus_ has joined #openstack07:55
*** [ack]_ has joined #openstack07:55
*** nevele has joined #openstack07:55
*** sets mode: +vvv exlt letterj _cerberus_07:55
*** kashyapc has joined #openstack07:56
*** ar1 has joined #openstack07:56
*** eday has joined #openstack07:56
*** coredump|br has joined #openstack07:56
*** PiotrSikora has joined #openstack07:56
*** tomo_bot has joined #openstack08:00
*** allsystemsarego has joined #openstack08:08
*** eisenhower has joined #openstack08:09
zykes-alekibango: here ?08:16
*** sanjib has quit IRC08:17
*** eisenhower has quit IRC08:19
*** miclorb has joined #openstack08:35
*** jakedahn has quit IRC08:39
*** mtaylor has joined #openstack08:48
*** mtaylor has joined #openstack08:48
*** ChanServ sets mode: +v mtaylor08:48
*** BK_man has joined #openstack09:08
*** kashyapc has quit IRC09:21
*** odyi has quit IRC09:26
*** odyi has joined #openstack09:26
*** kashyapc has joined #openstack09:26
*** irahgel has joined #openstack09:55
*** irahgel has left #openstack09:57
*** miclorb has quit IRC10:27
*** jonwood has joined #openstack10:34
*** irahgel has joined #openstack10:38
*** kashyapc has quit IRC10:41
*** hisaki has joined #openstack10:47
*** dizz has joined #openstack10:56
alekibangozykes-: ?10:57
*** hazmat has joined #openstack10:57
zykes-alekibango: did you get to start on the setup of a 4 node thing ?11:03
zykes-how did you think that that should look ?11:03
dendrobatesrc has been respun and uploaded to
dendrobatesplease test this could be it.11:08
zykes-dendrobates: why a respin ?11:10
dendrobatesthere were merges yesterday11:11
dendrobateswe need a final version to test11:11
*** ctennis has quit IRC11:16
zykes-i can download it11:17
zykes-want to try to install a 4 node + cluster11:17
*** terje has joined #openstack11:17
*** terje_ has quit IRC11:18
*** PiotrSikora has quit IRC11:19
*** DubLo7 has quit IRC11:19
*** eday has quit IRC11:20
*** eday has joined #openstack11:20
zykes-dendrobates: can i install that without using a tarball ?11:21
dendrobatesyou can use vish's script to pull trunk11:21
*** dizz is now known as dizz|away11:22
zykes-will do11:23
*** PiotrSikora has joined #openstack11:24
*** al-maisan is now known as almaisan-away11:26
zykes-dendrobates: would it be possible to add a feature req for RHEV support ?11:28
dendrobatesyes you can add a blueprint, but the fedora guys are working on it already11:29
dendrobatesso I don;t think it is necessary11:29
dendrobatesat least packages11:29
zykes-it would be nice11:30
dendrobatesI would imagine they might do it. but a blueprint would be good11:30
zykes-a customer that i work for and my company is using RHEV / VMWare11:30
zykes-+ KVM on ubuntu11:30
alekibangozykes-: today i installed waterpipes :) (hardware)11:30
zykes-waterpipes ?11:31
zykes-like in house stuff ?11:31
alekibangoyes :)11:31
zykes-dendrobates: what does a blueprint mean ?11:31
alekibangogot my hands dirty in hardware11:31
alekibangozykes-: dishwasher, etc11:32
*** ctennis has joined #openstack11:32
alekibangobut i will be on the task soon11:32
dendrobatesa blueprint is a overview of a requested feature, but they are mostly for devs that intend to do the  work themselves11:33
dendrobatesyou can also file a wishlist bug11:33
zykes-hmm k11:34
*** kashyapc has joined #openstack11:37
*** sophiap has joined #openstack11:44
zykes-isn't it dendrobates ?11:57
zykes-which is nova specific11:57
dendrobatestrue, openstack contains all openstack blueprints11:58
zykes-where did you see that the fedora folks where working on RHEV integration ?11:58
*** perestrelka has quit IRC12:01
*** perestrelka has joined #openstack12:02
dendrobatesI didn't, but I thought if anyone was going to work on RHEV support it would be them12:16
zykes-rhev is commercial from redhat, kind of like H-V from ms / VMWare12:17
dendrobatesIIUC, RHEV is more of a management framework for various hypervisors12:17
zykes-so doubt the fedora team will be involved there12:17
zykes-unless that RedHat launches a project that's open sourced then like RHEV on linux12:17
dendrobatesand we talk to the hypervisors directly, or via libvirt.12:17
zykes-can't you talk to the vCenter / RHEV-M servers instead ?12:18
zykes-if wanted12:18
dendrobatesI would have to know more about  RHEV to say for sure.12:18
zykes-it's the same as vCenter basically.12:18
zykes-it hosts an api that let's you talk to it to control actions12:19
dendrobatesWe would accept patches  from RH, though12:19
zykes-not from others that would make it ?12:19
dendrobatesahh, is it basically rebranded vcenter?12:19
zykes-dendrobates: nopp, it's redhat's equivelant of vCEnter12:20
dendrobatesI'd have to learn more about it to comment intelligently12:20
*** terje has quit IRC12:20
zykes-do you know about vCenter ? it's a management tool for ESXi / ESX12:21
zykes-manage storage etc on the ESXi nodes, RHEV-M does the same thing for RHEV / RHEL with KVM12:21
dendrobatesI have heard of t, butI am not up to date on vmwares offerings12:21
dendrobatesis there a libvirt driver for it?12:22
zykes-don't think so12:22
zykes-not sure though, think the new planned version serves an REST api though12:22
jonwooddendrobates: I think libvirt supports vCenter, at least to an extent.12:22
dendrobatesit sounds like it might duplicate some of openstacks functionality12:22
*** DubLo7 has joined #openstack12:24
zykes-dendrobates: it does some of the functionality yes, but for business that allready have it up and running i guess it would be a nice thing to have12:24
dendrobatesthe question is should we support them or should they support us?  Would people want to use the RHEV frontend to control openstack12:25
zykes-i would rather want openstack to control rhev12:26
zykes-to some extent12:26
zulis rhev the windows thing?12:27
*** westmaas has joined #openstack12:27
zykes-zul: it's getting ported in the near future if i'm not wrong12:27
zykes-to rhev alla linux12:28
zulzykes-: so yes :)12:28
*** johnpur has joined #openstack12:29
*** ChanServ sets mode: +v johnpur12:29
*** ar1 has quit IRC12:32
*** almaisan-away is now known as al-maisan12:35
*** aliguori has joined #openstack12:35
*** sagi_ has joined #openstack12:41
*** gustavomzw has joined #openstack12:46
dendrobatesnow I know what RHEV is.  It's the thing redhat did to try to monetize the Qurmranet purchase quickly12:51
*** pvo has joined #openstack12:53
*** ChanServ sets mode: +v pvo12:53
*** terje has joined #openstack12:56
zykes-dendrobates: it's the redhatification of qumranet yes12:57
zykes-i would wonder what version they are going to open source...13:05
*** jdarcy has joined #openstack13:13
*** dendrobates is now known as dendro-afk13:14
*** DubLo7 has quit IRC13:20
*** Podilarius has joined #openstack13:21
*** dendro-afk is now known as dendrobates13:22
*** pvo has quit IRC13:23
*** ambo has joined #openstack13:23
*** schisamo has joined #openstack13:29
*** mdomsch has joined #openstack13:35
*** guigui has joined #openstack13:36
*** ppetraki has joined #openstack13:44
annegentlemorning all, my doc changes are in the RC2 tarball (thanks!) but not pushed to yet, is that just a timing issue (tarball is 2 hours old) or do we need to trigger something?13:48
*** pvo has joined #openstack13:48
*** ChanServ sets mode: +v pvo13:48
zykes-question, can the queue mechanism be spread to multiple nodes ?13:54
*** gundlach has joined #openstack13:58
*** theron_ has joined #openstack14:00
*** Ryan_Lane|away is now known as Ryan_Lane14:10
*** dendrobates is now known as dendro-afk14:11
*** w00 has joined #openstack14:20
*** rnirmal has joined #openstack14:20
creihtmtaylor: did the swift rc1 packages ever get put in a PPA somewhere?14:21
*** dendro-afk is now known as dendrobates14:24
*** abecc has joined #openstack14:28
*** BK_man has quit IRC14:28
*** abecc has joined #openstack14:28
*** hisaki has quit IRC14:29
*** rnirmal has quit IRC14:30
*** rnirmal has joined #openstack14:31
*** BK_man has joined #openstack14:33
*** adjohn has joined #openstack14:34
*** adjohn has quit IRC14:34
*** JohnBergoon has joined #openstack14:37
*** JohnBergoon has quit IRC14:48
*** JohnBergoon has joined #openstack14:49
*** sanjib has joined #openstack14:54
*** sirp has joined #openstack14:55
*** pharkmillups has joined #openstack15:01
*** JohnBergoon has quit IRC15:05
*** jrussell has joined #openstack15:06
dendrobatescreiht: around?15:08
*** Kami_ has quit IRC15:08
gholtApparently the tweetosphere is all a twitter that we've officially made our first release.15:08
creihtdendrobates: yes (in case you are in a different channel and didn't see my reply :D)15:10
*** Kami_ has joined #openstack15:11
dendrobateswould you mind if I created a swift milestone named 2010.1 for the austin release.  I wouldn't change any of you version numbers15:11
* creiht gives in15:11
creihtdendrobates: at what time are you guys expecting to cut the nova release?15:12
dendrobatesafter, the meeting today15:13
gholtSeriously, that's a lot of tweetering, which is cool. But I guess my definition of "candidate" is outdated. ;)15:13
creihtdendrobates: cool15:14
zuldendrobates: are you guys having a release party?15:15
*** sophiap_ has joined #openstack15:15
dendrobatesyour definition is correct, ours is not15:15
dendrobateswe'll fix it next release15:16
dendrobateszul: at UDS15:16
zuldendrobates: me likey15:16
*** sophiap has quit IRC15:18
gholtdendrobates: Heh, all part of the fun of the first. :)15:18
*** BK_man has quit IRC15:19
*** BK_man has joined #openstack15:19
*** sophiap_ has quit IRC15:20
*** sophiap has joined #openstack15:21
*** leimy has joined #openstack15:24
gholtdendrobates: I just saw your futile response. Hehe15:26
creihtswift 1.1.0-rc2 has been uploaded15:28
creihtmtaylor: any way to get packages based on that in a ppa?15:28
mtaylorcreiht: yup...15:29
creihtthat would be awesome15:30
leimyThis is probably only tangentially on topic for this channel but, what's the advantage of bzr over say git?15:31
*** rnirmal_ has joined #openstack15:32
leimyit might be too flaimbaity to answer really.15:32
leimyer flamebaity :-)15:32
*** HouseAway is now known as AimanA15:32
leimyit's just that we're trying to decide between hg and git here, and I just saw this was using bzr, and I wondered what criteria led to that.15:32
gholtAh, I believe the surrounding launchpad is what really drove that.15:33
leimyConsistent UI?15:33
leimyoh... wait, launchpad.  I understand.15:33
*** anotherjesse_ has joined #openstack15:34
*** mdomsch has quit IRC15:34
leimywell that's a good enough answer for me :-).15:34
*** dendrobates is now known as dendro-afk15:34
leimyDon't want to degenerate into vi vs emacs or anything.15:35
creihtleimy: My short answer to that type of questions, is I don't care as long as it is distributed :)15:35
gholtCentrally. j/k15:35
leimyMy company just announced support of OpenStack today, so I'm trying to get up to speed a bit.15:35
*** rnirmal has quit IRC15:35
*** rnirmal_ is now known as rnirmal15:35
gholtAh, cool. Which company is that?15:35
uvirtbotNew bug: #664567 in swift "Memcached race condition and test cleanup" [Undecided,New]
creihtleimy: they all have their advantages and disadvantages, so you really just have to try them out15:36
*** spectorclan has joined #openstack15:39
leimygholt: oops went AFK.  Cirrascale.15:39
leimycreiht: yeah.  Even Linus Torvalds had a hard time picking on Hg vs Git.  He said "eh, they're about the same".15:40
leimyIt really depends on what workflow you want to implement.15:40
*** metcalfc has joined #openstack15:40
*** pvo has quit IRC15:41
*** thom has left #openstack15:41
uvirtbotNew bug: #664572 in nova "AttachVolume doesn't validate input" [High,New]
*** kevnfx has joined #openstack15:41
gholtleimy: Neat. And welcome to our crazy world. :)15:42
*** metcalfc has quit IRC15:43
spectorclanDesign Summit Attendees - working on developer party for third night downtown San Antonio - any video games you want at party? planning to get PS2, Wii, etc ; what do you want?15:44
*** metcalfc has joined #openstack15:44
leimygholt: thanks!  We're like a lighter and more focused version of Verari.  I'm on the container management engineering team for FOREST and our FE products.15:44
*** guigui has quit IRC15:46
creihtspectorclan: guitar hero/rock band are good party games15:47
spectorclancreiht Yep, already on the list15:47
creihtleimy: the FOREST container looks cool15:47
glange_spectorclan: dwarf fortress or minecraft?15:48
* creiht smacks glange_ 15:48
leimycreiht: It's the best project I've ever gotten to work on :-)15:48
spectorclanglange: interesting, never hard of them15:48
creihtleimy: yeah that looks cool15:48
anotherjesse_spectorclan: rockband15:48
glange_spectorclan: I'm joking, they are not good for parties15:48
creihtI would love to see swift running in one of those15:48
leimyThe lead designer for the container is the guy responsible for Cray's best selling machine :-)15:48
spectorclanglange ok, I am a bit old fashioned in my games, still love 1941 and Galaga15:49
creihtspectorclan: you would like my mame cabinet then :)15:49
spectorclancreiht: mame?15:50
*** dendro-afk is now known as dendrobates15:50
*** dfg_ has joined #openstack15:50
creihtspectorclan: arcade emulator15:50
*** silassewell has joined #openstack15:50
glange_spectorclan: it might be cool if there was room for board games though15:50
spectorclancreith: interesting - i have Sinistar on my PS2 - love that game15:50
creihta pic of it in progress15:51
spectorclanglange: we can do board games; bar will have room for that; which ones? RISK15:51
glange_spectorclan: maybe bring your own board games would be best15:51
glange_we in san antonio have a bunch we like to play15:51
spectorclancreiht: very cool - we should add that picture to the new OpenStack Flickr site
creihtspectorclan: heh15:52
spectorclanthat may be the only time we see Mark at his desk working15:52
*** glange_ is now known as glange15:53
jaypipesspectorclan: my only request is to have it indoors.  I remember melting my butt off on top of Whole Foods ;)  Then again, it is November, so it should only be 90 degrees or so instead of 110.15:56
* leimy goes into lurk mode.15:56
leimyNice meeting you all!15:56
spectorclanjaypipes: it will be indoors at a bar near the river; no problem on that. I come from Florida and like AC15:57
jaypipesspectorclan: down by the river!? :) nice.15:57
* jaypipes pictures Chris Farley participating in games.15:57
spectorclanjaypipes: waiting to get the sites to choose from ; will let everyone know the site when I get the info15:58
*** littleidea has joined #openstack15:59
*** littleidea has quit IRC16:00
*** theron__ has joined #openstack16:00
*** theron_ has quit IRC16:03
*** eday has quit IRC16:04
*** eday has joined #openstack16:04
*** theron__ has quit IRC16:05
*** arif-ali has joined #openstack16:05
anotherjesse_annegentle: is someone figuring out what is up with your doc updates and trunk?16:05
*** eday has quit IRC16:05
*** eday has joined #openstack16:05
annegentleanotherjesse_: not that I've seen, yet.16:05
annegentleannotherjesse_: if you can take a look that would be great (and teach me to fish as needed) :)16:06
anotherjesse_annegentle I'll see what I can do - I was going to take a pass at docs as well16:07
anotherjesse_I'm not a BZR/launchpad expert16:07
*** ChanServ sets mode: +v eday16:08
*** gustavomzw has quit IRC16:08
*** gustavomzw has joined #openstack16:09
annegentleanotherjesse_: great, thanks, me neither, but another pair of eyes is always good.16:09
*** krish has joined #openstack16:11
*** calavera has quit IRC16:11
*** blpiatt has joined #openstack16:18
*** anotherjesse_ has quit IRC16:23
*** arif-ali has quit IRC16:24
*** daleolds has joined #openstack16:26
dendrobatesjaypipes: any new bugs?16:26
jaypipesdendrobates: nope.16:26
*** arif-ali has joined #openstack16:27
edaydendrobates: were you able to reproduce the issues you were seeing last night?16:27
dendrobateseday: I was not.  I was extremely tired, so it was probably just me.16:27
*** krish is now known as krish|FoodHuntin16:28
*** krish|FoodHuntin is now known as krish16:29
edaydendrobates: ahh, thats good (not you being tired) :)16:29
*** kashyapc has quit IRC16:33
*** littleidea has joined #openstack16:33
*** littleidea has quit IRC16:34
*** krish is now known as krish_FoodHuntin16:35
*** annegentle has left #openstack16:36
*** annegentle has joined #openstack16:36
jaypipesvishy: ping16:44
*** joearnold has joined #openstack16:47
*** kashyapc has joined #openstack16:50
jaypipesvishy: I've been struggling to find where in the code base an Instance's host attribute is set. :( can you help?, pls16:54
jaypipesvishy: I'm trying to do an to the compute manager to create an instance, but in order to do that I need the instance's host... kind of a chicken and egg thing.16:54
vishyyou should be doing a cast to the scheduler16:55
vishycreate should be in ComputeManager, but the code itself should run in api or scheduler16:57
jaypipesvishy: k :(  I guess I'm missing something then, because I don't see how that will work.  I need to send a call to the compute node to update the instance data for a specific instance.  Why would I ask the schedyuler to do that?16:57
*** kakoni has joined #openstack16:58
*** anotherjesse_ has joined #openstack16:58
vishyi don't think the compute node is the one responsible for talking to the database16:58
jaypipesvishy: because of the silly ec2 ID translation crap, I need to first create the instance on the compute node, send the instance identifier back to the API node, then send a call back to the copmute node to update the EC2 id of the instance after its host is addigned :(16:58
kakoni creiht: around?16:58
jaypipesvishy: compute manager.16:58
*** blpiatt has quit IRC16:59
vishyscheduler or some other cluster level process should ultimately be responsible for updating the db16:59
jaypipesvishy: if the compute node is not responsible for talking to the database about the instances that the compute node is managing, then who is?  I don't understand that.16:59
*** anotherjesse_ has quit IRC16:59
vishycompute node will send updates through the queue16:59
vishythe idea being that the workers don't have write access to the db17:00
vishybut anyway that is a ways off17:00
vishyfor now, i thought the ec2-id is generated from internal id now?17:00
jaypipesvishy: but the API server is currently updating the database directly.  I'm trying to set it up so the API node sends an update request to the scheduler to update an instance's data.  But the problem is I have no clue how to send a request to the scheduler to update an instance that I don't know the host of.17:00
vishywhy doesn't the scheduler just update the db?17:01
vishythat seems like a good place for db writes to occur?17:01
*** khussein has joined #openstack17:01
jaypipesvishy: that means the scheduler will be the bottleneck to the system. seems like an awful choice for a bottleneck.  why not have the worker node update the database, allowing the scheduler to serve more requests?17:02
vishywell it could be a separate db writer worker, I just think you're offloading a ton of work to the queue17:02
jaypipesvishy: sorry, I'm not following you.  what work am I offloading to the "queue"?17:03
vishysending updates through the queue to compute17:03
vishyinstead of writing directly from api or scheduler process17:03
jaypipesvishy: alright, lemme start from scratch here...I think we're getting off topic.  My task is to move database write calls to *the compute manager*.  Doesn't the compute manager live on the compute node? :(17:04
vishyit lives in multiple places17:04
jaypipeswell there's my problem I guesss :)17:04
* jaypipes was confused by that.17:04
vishyso managers can run in process or through a service17:05
jaypipesservice, manager, driver, backend.  there's about 4 different ways we describe "things that do shit" .17:05
jaypipesvishy: so how are managers "started up"? by a service? or by some bin/nova-xxx tool?17:06
vishynova-compute loads a service that wraps the ComputeManager17:06
vishyand exposes its public methods via rpc17:06
jaypipesvishy: ok, so is nova-compute the "compute node"?17:06
jaypipesso it *does* have the compute manager, then.17:06
vishynova-compute runs on the host that is actually running vms17:07
vishybut ComputeManager can be (and is) loaded without a wrapping service17:07
vishyso it can be used in process directly17:07
vishyso all the code relating to vms should be in the ComputeManager17:08
vishybut that code won't necessarily all run on the nova-compute host17:08
vishyis that clear?17:09
jaypipesvishy: OK, so if I move code to create instances in the database from the API server to the compute manager, that would be fine, no?17:09
vishyyeah that is defintely phase one17:09
vishythen you can just call the code directly from api/ec2/cloud.py17:10
vishywe already do that with network_manager17:10
jaypipesvishy: the problem with that, though, is that the EC2 API node fiddles with the instance data in the database itself and only the ec2 node knows how to translate the internal id to an ec2 id.17:10
vishybleh, I haven't looked at the ec2_id stuff17:10
vishylet me look real quick17:10
*** mdomsch has joined #openstack17:10
jaypipesvishy: but I think I might be able to work around it..17:10
jaypipesvishy: lines 857 through 86317:11
jaypipesvishy:  in api/ec2/cloud.py17:11
*** rnirmal has quit IRC17:11
jaypipesvishy: another question that's really bugging me...hopefully you can shed light on.  Why is it that *some* things are called on the network_manager directly (like network_manager.allocate_fixed_ip(), but other things aren't (like rpc.cast on network topic for setup_fixed_ip)?  This confuses the heck out of me.17:13
vishythat particular one is going to move as per my doc17:14
vishycalling directly is for stuff that can be done in process17:14
vishycast is for stuff that has to happen on the remote host17:14
jaypipesvishy: so, really, calling directly is only for database stuff.17:15
vishydirectly is for stuff that can or must happen locally17:15
vishya better example is17:15
vishyvolume code17:15
*** ibarrera has quit IRC17:16
vishythe compute host needs to discover remote volumes17:16
jaypipesvishy: discovering remote volumes is a local task?17:17
vishybut ComputeManager shouldn't have to understand how the particular commands work17:17
vishyso it calls a local version of VolumeManager17:17
vishyjaypipes: sure it needs to call aoe-discover on the local host for example17:17
vishyjaypipes: and it needs to find out the local path of the volume to attach17:18
vishythat is a volume related command that runs on the compute host17:18
jaypipesvishy: sorry, I'm completely lost :(  mostly because I've never used aoe and I don't know anyone who has.17:19
vishydoing that through rpc would make the actual discover code ("sudo aoe-discover") run on the volume host, which isn't what we want17:19
vishywell it is the same for iscsi17:19
vishythe volume host exports the volume17:19
*** silassewell has quit IRC17:19
vishyand the compute host has to connect to it17:19
*** rnirmal has joined #openstack17:19
jaypipesvishy: ok, that's a little clearer now. thx.17:20
*** al-maisan is now known as almaisan-away17:20
*** westmaas has quit IRC17:21
jaypipesvishy: I guess my biggest confusion stems from things called manager that don't actually manage things.  Like the VolumeManager example above.  That doesn't manage anything.  It just finds out what volumes are managed by some other node.17:21
*** blamar has quit IRC17:22
jaypipesvishy: but on the volume node itself, the VolumeManager actually *does* manage something...17:22
*** westmaas has joined #openstack17:26
*** BK_man_ has joined #openstack17:28
jaypipesvishy: thx for your help this morning.  I know I sound like I'm bitching, but I'm really just trying to understand everything and provide some enhancements that might make it easier for new contributors to grapple with the code.17:29
*** BK_man has quit IRC17:31
vishyjaypipes: no problem. I know it is a bit confusing with service / manager /driver, I was hoping my discussion of networking would make it clearer17:32
*** BK_man_ has quit IRC17:33
jaypipesvishy: it did somewhat.  what wasn't clear (because you weren't talking about this...) was that the manager classes behave and do different things depending on whether they are called in-process or via a service (a service which listens for specific messages...)17:34
jaypipesvishy: and now that is clear to me.  I'll work with annegentle to put some docs together about that.17:34
*** blamar has joined #openstack17:38
*** krish_FoodHuntin is now known as krish17:39
*** arif_ali has joined #openstack17:43
*** Ryan_Lane is now known as Ryan_Lane|food17:43
*** arif-ali has quit IRC17:43
*** spectorclan has quit IRC17:46
*** leimy has quit IRC17:47
*** pharkmillups has quit IRC17:50
*** aliguori has quit IRC17:55
creihtkakoni: howdy17:57
creihtkakoni: update your code from trunk, then try the bench again :)17:57
*** aliguori has joined #openstack17:57
creihtturns out in my effort to clean up that code, I was a little over zealous and made the benchmark put to just one container17:58
creihtwhich makes that one container the weak point, so the updated code will create several (20 by default) and assign the objects randomly to each container17:59
*** aliguori has quit IRC17:59
creihtmtaylor: any update on the ppa?18:02
*** arif_ali has quit IRC18:03
kakonicreiht: hmm. yes, saw that. With hmm.. 3 object nodes I was (+1 proxy) I was seeing something like 500 put reqs/sec. So pretty decent18:04
*** kevnfx has quit IRC18:04
creihtI ran the test here in our lab to just one proxy, and was able to max the proxy out at close to 1500 reqs/sec18:05
kakonicreiht: anyway. question. Any hints/advices/tuning tips how one should arrange zones?18:05
creihtthat proxy was an 8 core box, running 16 workers18:05
creihtkakoni: zones are used to isolate the data as much as possible18:06
mtaylorcreiht: sorry - not feeling well today, went back to sleep right after we talked earlier18:06
mtaylorcreiht: I'll get that done in a few minutes...18:06
creihtmtaylor: doh... no worries  hope you get better soon18:06
kakonicreiht: question 2. How do you guys sync builder files in your clusters?18:06
creihtmtaylor: maybe at the dev summit you can walk me through how a lot of that stuff is set up18:06
*** Ryan_Lane|food is now known as Ryan_Lane18:07
creihtkakoni: so further on question 1:  So you should take things like power, network availability, etc. into account18:08
creihtso if you can put the zones on different power sources, that would be a good idea18:08
creihtbasically push the zone as far up the stack as possible18:08
creiht2.  The builder files only have to be on an admin server (though it is a good idea to keep those backed up as they are needed to build the rings18:09
creihtfor the ring files, we use a push method by building the ring, and then pushing the ring out to all the nodes with scp18:09
creihtthe servers are designed that they will reload the ring if they notice the modified time has changed18:10
zykes-what's image warehouse daemon ?18:13
dendrobatesvishy: are you guys still testing.18:13
* jdarcy wakes up.18:13
*** Ryan_Lane is now known as Ryan_Lane|away18:15
kakonicreiht: okey cool. Thanks for the info. I did notice that the my initally bad perf figures are really caused by the bad IO capabilities on the storage servers (they were like I said pretty crappy machines anyway;)18:15
creihtkakoni: either way you helped me find a small problem with the benchmark tool, so thank you :)18:15
zykes-what bench mark tool is that ?18:16
zykes-dendrobates: are you here ?18:16
*** silassewell has joined #openstack18:17
*** jero has quit IRC18:17
kakonicreiht: no prob. But really, need to get myself together and proceed to the next level=>start committing to the swift tree;)18:18
annegentlejaypipes: ah process vs. service, I never would have thought that, feel free to 'splain where that distinction goes in the doc... I have networking info in, feel free to start a new section there if you see fit.18:19
vishyyeah sorry had crazy meetings18:20
*** khussein_ has joined #openstack18:21
dendrobateszykes-: am now18:21
annegentlecan someone help me with dependencies? I'm installing the packaged Nova on Ubuntu 10.04...18:22
annegentleAfter getting ImportError: No module named twisted.internet.threads for the Nova RC2, I installed twisted.18:22
annegentleNow I get "No module named gflags" - what do I do to get that installed? It's a python dependency of some sort?18:23
zulannegentle: sudo apt-get install python-gflags18:23
*** pharkmillups has joined #openstack18:23
*** khussein has quit IRC18:23
*** khussein_ is now known as khussein18:23
annegentlezul: uhhh now I get E: Couldn't find package python-gflags (I think I saw a forum post about this problem?)18:25
zuluniverse enabled?18:25
annegentledoesn't look like it...18:27
*** spectorclan has joined #openstack18:34
*** rlucio has joined #openstack18:36
*** joearnold has quit IRC18:36
uvirtbotNew bug: #664656 in swift "object auditor can error for some objects" [High,In progress]
*** kashyapc has quit IRC18:37
annegentleOk, documenting that on 10.04 you have to download gflags to install it (it's available on 10.10 though)18:38
rlucioannegentle: by 'download' you mean apt-get install python-gflags from the nova PPA right?18:41
annegentlethat didn't work for me, so does the Nova PPA need to change it sounds like?18:43
zuldendrobates: ping you are missing the python-gflags on 10.04 it probably should be in the nova ppa18:43
*** gundlach has quit IRC18:44
rluciopython-gflags_1.3-2ubuntu0ppa1_all.deb is in the nova ppa already18:45
rluciounless version dependencies changed recently, it should work18:45
zulannegentle: yeah then you might want to add to your instructions on how to get python-gflags from the nova ppa18:46
annegentleokay, thanks18:46
annegentleand uh how do I do that? :)18:47
zullike that blog post you showe me :)18:48
zykes-dendrobates: time for a message ?18:48
annegentleoh! Okay. :)18:49
zykes-what blog is that ?18:49
annegentlezykes-: here it is
mtaylorcreiht: source packages uploaded18:51
zykes- < wonder what's up with folks bothering to write comments like the comment on that post.18:51
mtaylorcreiht: I uploaded them to both the openstack and the swift ppa18:52
creihtmtaylor: cool18:52
mtaylormy pleasure!18:52
mtaylorcreiht: I suppose there is really nothing stopping you from doing that yourself... perhaps I should write a doc page18:52
zykes-'Even "Free" has a price', yeah sure, but then again it's not going to be better then the people making an effort in the community18:53
*** Grizzletooth_ has joined #openstack18:53
creihtyeah that would be cool18:53
*** Grizzletooth_ has quit IRC18:54
spectorclanzykes: I agree, pretty poor comment18:55
spectorclanzykes: he didn't have the guts to say I am from VMWARE18:55
zykes-i'm trying to get my company to go with openstack instead of vcloud18:56
zykes-so hoping that i'll win :)18:56
spectorclanzykes: i am sure I can get people to help you if you need18:56
*** sirp1 has joined #openstack18:56
zykes-heh, i've got a plan :)18:56
zykes-just need an approval for something first.18:57
spectorclan Openstack press release18:58
zykes-i'm based in norway so (:18:58
zykes-dunno how many folks here are from europe atm18:58
*** blpiatt has joined #openstack18:59
mtaylorzykes-: some18:59
mtaylorzykes-: soren is in denmark19:00
*** sirp has quit IRC19:00
zykes-ok :)19:00
*** littleidea has joined #openstack19:00
*** sirp1 has quit IRC19:01
jdarcyLooks like they're trying to attract the people who leave a lightweight instance running somewhere else because it's cheaper.19:01
*** sirp has joined #openstack19:02
zykes-if we go ahead we're gonna try to make a "genericable" billing system i think.19:02
zykes-that was the discussion with some folks at a meeting at least19:03
*** blpiatt has quit IRC19:03
dendrobateszul: ok, I'll upload it soon19:05
*** arif-ali has joined #openstack19:11
dendrobatessirp jaypipes:  is glances ready to go.  I used it to make the RC, but did not test it.19:13
*** BK_man has joined #openstack19:17
*** arif-ali has quit IRC19:17
*** aliguori has joined #openstack19:20
_0x44sirp, jaypipes, dendrobates: Sorry that I didn't get the swift client done before I left for Italy.19:22
dendrobates_0x44: all is forgiven19:22
*** jonwood has quit IRC19:22
sirpdendrobates:  afaik, will need to check with jaypipes or test it to be sure19:22
sirp_0x44:  no worries man, porting it to use swift's new fangled client was pretty simple19:23
_0x44sirp: Awesome, I wasn't sure if it had made it in for RC19:24
dendrobatessirp: fyi, I have not tested it at all.  but it is optional19:24
dendrobatesI hope we see some glance bugs submitted in the next few days19:25
*** Lukey575 has joined #openstack19:26
jaypipes_0x44: dude, no worries :)19:26
jaypipesdendrobates: it should be, yes, but would be good to get mtaylor's input on it.  It's not my specialty :)19:27
mtaylorjaypipes: aroo?19:27
jaypipesmtaylor: the for glance.  perhaps give it a look?19:27
jaypipesmtaylor: it's modeled after nova's. shouldn't be much difference.19:27
mtaylorjaypipes: cool19:27
dendrobatesnot mine either.  Python distribution seems a bit magical compared to19:28
jaypipesmtaylor: but I don't know much about distutils stuff.. :)19:28
jaypipesdendrobates: mtaylor is awesome with magic ponies.19:28
dendrobatessoren fixed up the  nova one for me.19:28
*** gundlach has joined #openstack19:29
mtaylorjaypipes: you may want to add the things in bin as scripts19:29
mtaylorin the scripts=[]19:29
jaypipesmtaylor: k. sirp can you do that?19:29
Lukey575join #ganglia19:30
mtaylorjaypipes: and then, I'd recommend doing python sdist (which will make a source tarball) untarring that somewhere and checking to make sure it has everything you expect19:30
mtayloryou may need to add a file listing additional files19:30
*** khussein has quit IRC19:30
jaypipesmtaylor: k19:30
mtaylorsuch as your doc dir19:30
_0x44Lukey575: Don't you tell me what to join. :|19:30
mtaylorbut again, you can steal one of those from nova19:30
dendrobatesdoes anyone have any issues with me creating Authors file from the commit logs to include in the nova tarball?19:32
Lukey5750x44: my bad, forgot the '/' =)19:33
_0x44Lukey575: :)19:33
*** Ryan_Lane|away is now known as Ryan_Lane19:35
*** perestrelka has quit IRC19:35
*** perestrelka has joined #openstack19:36
jaypipessirp: ok, never mind. i'll do the things mtaylor suggested and propose for merging.19:37
mtaylordendrobates: no. sounds fine19:40
mtaylordendrobates: you see in where I'm already making a ChangeLog yeah?19:40
*** krish has quit IRC19:41
mtaylordendrobates: I think that's an excellent idea19:41
*** rlucio has quit IRC19:41
dendrobatesbut it is nice to have a concise unique sorted list19:41
mtaylordendrobates: oh - no, I meant - you see where there is already an sdist hook you can add to19:42
dendrobatesFirst Last <Email>   ok for the format19:42
mtaylorworks for me19:42
gundlachjaypipes: do you have your test-overhaul branch up yet?  i'm running out of tests to convert that don't require a reactor running :)19:42
dendrobatesmtaylor: I was thinking of ninja patching.19:42
mtaylordendrobates: well, you could - but there's already a hook that runs during sdist - but really, it's up to you :)19:43
jaypipesgundlach: :) hehe, ok, gimme 30 minutes.  going to finish glance packaging stuff and I will then push my work.19:43
gundlachtx :)19:43
dendrobatesI'd rather not touch until after the re;ease19:44
_0x44dendrobates: Is it not possible to tag a certain commit as the release, so that later commits can be applied to trunk?19:44
vishyrc2 seems to be working for me19:45
dendrobatesvishy: awesome19:45
vishyfixing up my install script19:45
sirpback, catching up19:45
vishylooks like there are still permissions issues19:46
vishyyup booted19:47
dendrobates_0x44: it is but this beeds to be changed before the release.  I'll ninja patch in the Authors and generate then merge the changes later19:47
_0x44dendrobates: Ah okay, I didn't realize there needed to be additional changes to the stuff.19:47
vishysigh: i think i'm listed with an old email in a bunch of files19:48
dendrobatesvishy: I pulled out all but your gmail, I think19:48
*** silassewell has quit IRC19:48
vishyoo that would be was listed as yahoo19:49
dendrobatesYes I used the gmail address for you19:49
dendrobatesis that ok?19:49
*** mtaylor has quit IRC19:50
*** joearnold has joined #openstack19:57
*** kakoni has quit IRC19:59
jaypipesgundlach: alrighty, I pushed my recent work to lp:~jaypipes/nova/cleanup-unittests.  The single revision has the work I've been doing on creating a fresh test environment based on fixtures and a unit test for that work.  It does *not* include the work I did previously.  So, feel free to propose a merge of your own work into the cleanup-unittests branch and I'll merge it in :)20:03
vishydendro: updated my repo with simpler install script20:04
vishydoesn't use pip at all anymore, just the ppa20:04
jaypipesvishy: nice.20:04
*** littleidea has quit IRC20:05
spectorclanDesign Summit Party for Developers ONLY on 3rd night at Coyote Ugly Saloon in SA (VIP ROOM)
ttxdendrobates: have a date for the openstack party at UDS ? Looke like the server team is having dinner wednesday now20:13
dendrobatescan someone who really knows review jaypipes MP20:13
dendrobatesttx: it's tuesday20:13
ttxah, tuesday I can't make it :)20:13
dendrobatesttx: WHAT!20:14
ttxyou asked me and I told you :)20:14
dendrobatesIt was not completely up to me.20:14
annegentlevishy: awesome re: ppa, I'm going to take a look at your script20:15
dendrobatesI think it's tuesday, my brain is fried, so wait for the official announcement20:15
*** silassewell has joined #openstack20:15
dendrobatesguys we NEED someone to review jaypipes changes to glances for the release.  Any volunteers?20:19
edaydendrobates: I can look, but I've not looked at many files :/20:20
dendrobateseday: me neither.20:21
edayspectorclan: is the food coming from that bar too?20:22
spectorclaneday: yes - food and drinks and we will video games and pool20:22
spectorclaneday: is there a food issue? I have not been there before20:24
dendrobatesspectorclan: we have vegans among us.20:24
dendrobatesand vegetarians20:24
_0x44dendrobates: I reviewed, it looks good.20:25
_0x44dendrobates: Do we need a second?20:25
dendrobates_0x44: how good are you with
*** littleidea has joined #openstack20:26
dendrobatesI just want to make sure some one who understands python dist has looked20:26
_0x44dendrobates: I've written a few of them, most recently for that pdx app you don't like the name of20:26
*** rlucio has joined #openstack20:26
*** khussein has joined #openstack20:26
_0x44dendrobates: Hold one, I'll pull his branch and run it just to make sure.20:27
edayjaypipes: should we just rename the bin/* files instead of symlink?20:27
dendrobates_0x44: thanks20:27
_0x44eday: With python build it doesn't look like they were symlinked20:28
edayspectorclan: was just curious, my vegan & feminist side was screaming out in pain :)20:28
eday_0x44: in the merge req the symlinks were added20:29
edaysee lines 10-1320:29
edayerr, 1101420:29
spectorclaneday: we will certainly have a veggie spread for you. I will take care of it!20:29
edaydamnit, 11-1420:29
*** khussein has quit IRC20:29
edayspectorclan: no worries :)20:29
*** sparkycollier has joined #openstack20:29
_0x44eday: I see now, I glossed over them because there wasn't any syntax20:29
*** khussein has joined #openstack20:29
_0x44eday: Yeah, look at it now, I think I agree that those should be moved instead of symlinked.20:30
_0x44dendrobates/jaypipes: looks like it works correctly20:31
*** rlucio has quit IRC20:31
creihtspectorclan: yeah so, would it be possible to have a place that would be a bit more friendly to both sexes?20:32
jaypipes_0x44: cool, thx20:32
dendrobatesor even all 4 sexes.20:32
annegentleeday: thanks for your feminist side remark, hee hee.20:32
*** pvo has joined #openstack20:33
*** ChanServ sets mode: +v pvo20:33
_0x44dendrobates: 5, there's an A* movement to be added to glbt20:33
jaypipes_0x44, eday: so, I should remove the symlinks?20:33
_0x44jaypipes: I think it would be cleaner20:33
jaypipes_0x44: sure, no prob. :)20:33
spectorclancreith: I have never been to that place before and we wanted a nice bar near the river that people could easily get to. I will check with the event planner to see what other options are20:34
*** blpiatt has joined #openstack20:34
*** schisamo has quit IRC20:34
_0x44_cerberus_: What was the name of that fancy mexican restaurant the #ozone team took Jared and I to when we first visited you guys?20:34
_0x44_cerberus_: Didn't they have a bar?20:34
creihtI've never been there either, but it is pretty easy to tell from the web site... :)20:34
jaypipesdendrobates, _0x44: ok, symlinks removed. pushed.20:35
dendrobates_0x44: la tuna20:35
spectorclancreiht: i agree; however the VIP room does look nice. I will let you know what the event folks have to say when I hear back from them20:35
*** DubLo7 has joined #openstack20:35
dendrobatesjaypipes: thanks20:36
_0x44eday: You want to approve now that the symlinks are gone?20:37
*** rlucio has joined #openstack20:37
creihtspectorclan: thanks20:37
edayjaypipes: approved20:37
edaycreiht, annegentle: hehe :)20:38
jaypipesspectorclan: I'd prefer someplace a little friendlier for the ladies, too... for the record. :)20:39
spectorclanjaypipes: no problem, i have given the word back to the event folks. If anyone has been to SA river area and has a place they like, let me know and I can suggest20:40
pvodoes it have to be on the river? imo, that is the best place to buy a $5 margarita for $1020:40
eday_0x44: I think I want to the same nice mexican place down there, it was pretty good. can't remember the name though20:40
creihtyeah I'm not sure I've been to any place on the river walk that was really good :)20:41
spyall the mexican places on the river aren't that great heh20:41
creihtyeah especially the mexican food places20:41
spectorclanpvo: we wanted something downtown that we could walk to from the hotel20:41
_0x44spy: La Tuna was pretty good.20:41
pvoLa Tuna is my vote20:41
pvoif there is a vote20:41
spectorclanHow about Ticket Sports Pub? I will ask about La Tuna20:42
pvoits a short  cab ride from downtown, but is kid/family/nerd friendly20:42
jaypipespvo: if it has Golden Tee, you're on. ;P20:42
spy_0x44: la tuna isn't on the river :)20:43
spyla tuna is pretty good though20:43
_0x44spy: You're right, what was that place you took Jared and I to when we first came to visit?20:43
pvojaypipes: it isn't a "bar" per se, but is a good venue20:43
spectorclanspy: working it20:43
_0x44spy: It started with an A?20:43
pvo_0x44: rosarios?20:43
edaylets just go to The Cove, good veggies burgers and we can do out laundry too :)20:43
creihtThe cove is awesome20:43
jaypipespvo: heh, I was just joking and trying to provoke a GT war. :)20:43
creihta bit aways though20:43
_0x44pvo: Ah, acenar20:44
_0x44jaypipes: your merge prop is approved.20:44
jaypipes_0x44: cheers20:44
johnpurcreight: casa rio is OK, agree on the cove (awesome)20:44
glangeAcenar has decent food, it's owned/run by the same people that do Rosarios20:44
dendrobatesthe cove is great20:44
_cerberus_+1 on the Cove20:44
spythe cove is pretty good too20:44
* jaypipes should just become a veggie. would be so much easier to hang with eday and dendrobates.20:44
johnpureday: you can wash your car there too!20:45
spythe cove has veggie stuff20:45
_0x44jaypipes: You should!20:45
_cerberus_jaypipes: plenty of awesome non-veggie burgers at The Cove as well ;-)20:45
pvocove also works. less room than la tuna20:45
spectorclani thought all texans ate MEAT20:45
pvospectorclan: not all.20:45
dendrobates_0x44:  adelante20:45
johnpurcove has the best fish tacos in sa.20:45
jaypipes_0x44: hey, I've actually given it some serious thought recently.  packed on 25 lbs since I quit smoking last year...20:45
edayjohnpur: hehe20:45
*** sparkycollier_ has joined #openstack20:45
*** sparkycollier has quit IRC20:46
*** sparkycollier_ is now known as sparkycollier20:46
pvojaypipes: it is a bit harder in south texas to be veggie20:46
pvoand even harder to be vegan20:46
jaypipespvo: indeed.20:46
_0x44dendrobates: adelante?20:46
_cerberus__0x44: great vegetarian mexican place20:46
edayspectorclan: I live in Portland, OR, so the texan rule doesn't apply20:46
_cerberus_Not sure if you went or not20:46
dendrobatesthe cool mexican restaurant20:47
spectorclanegay: I can see being a Vegan in Portland, great city20:47
spectorclanoops eday sorry20:47
_0x44Wait, there's a vegetarian mexican restaurant in SAT and you guys have never taken me?20:47
pvooh dear20:47
* _0x44 is glad he quit. ;)20:47
* _0x44 kids (to be clear)20:47
_0x44Alright, I'm out for the night. Good night kids20:48
edayyeah adelante is really good too20:48
glangefor friendly service, we could go to Dick's Last Resort, it's right on the river20:50
spectorclanglange: if we don't do Coy. Ugly than Dick's Last Resort must be more sexist?? never been there so I don't know20:50
*** jxta has joined #openstack20:50
johnpurspectorcaln: yes. underwear hanging from the ceiling (or so i've been told:))20:51
glangespectorclan: I was kind of joking, their "thing" is that they are rude to the customers in a kind of joking/friendly way20:51
glangeI don't know if they have a party room either20:52
spectorclanglange: no problem. I can check them out but the underwear thing may be too much for people20:52
*** ctennis has quit IRC20:56
*** blpiatt has quit IRC20:57
dendrobatesaustin meeting in 2 minutes20:58
*** adjohn has joined #openstack20:58
creihtlet's do this :)20:58
dendrobatesin #openstack-meeting20:58
*** tariqne has joined #openstack20:58
*** tariqne has quit IRC20:59
*** westmaas has quit IRC21:01
*** tasimne has joined #openstack21:02
tasimnewill there be more documentation available for network configuration21:03
annegentletasimne: short answer: yes, longer answer: for now there's a current explanation in the mailing list: that I'd like to turn into doc... your thoughts?21:06
*** ctennis has joined #openstack21:08
*** kevnfx has joined #openstack21:13
*** jxta has quit IRC21:13
*** openstack has joined #openstack21:19
*** sirp has quit IRC21:20
*** sirp has joined #openstack21:21
edayannegentle: we can probably update to remove all the deprecation warning msgs too, since those were fixed21:23
annegentleeday: oh yeah, I'll do that21:24
*** brainproxy has joined #openstack21:27
*** Pentheus has joined #openstack21:30
tasimneone more thing, i read somewhere that there would be a web UI, i guess it is not included in this release?21:31
*** jdarcy has quit IRC21:31
*** spectorclan has quit IRC21:32
annegentletasimne: check out, it's lightweight21:32
*** mtaylor has joined #openstack21:34
*** ChanServ sets mode: +v mtaylor21:34
PentheusI'm trying to figure out if I'm looking at a bug or if I'm doing something stupid.  Trying to "manually" build swift rings within .py and it seems to have some code to deal with None .devs entries (infact it makes them on its own) but other code which freaks out on NoneTypes:
edaydendrobates: are we going to create a new branch for bexar series, or keep moving within the same branch?21:39
creihtPentheus: it is quite possible that the None's may not be caught everywhere21:41
*** sparkycollier has quit IRC21:41
Pentheuscreiht: yeah, that was my feeling21:44
Pentheuscreiht: seems to hum along with
creihtPentheus: ahh yes, the ring builder assumes that when you are first creating the ring, that there would be no empty spots (and this would never happen if you are using just the ring builder code)21:44
creihtPentheus: please file a bug... should be easy to fix21:44
*** schisamo has joined #openstack21:45
*** gundlach has quit IRC21:46
*** miclorb has joined #openstack21:46
Pentheuscreiht: will do, thanks21:46
creihtPentheus: no problem, and thanks for poking at it :)21:47
creihtPentheus: out of curiosity, are you manually building the ring like that to learn more about it, or do you have other ideas?21:50
*** mtaylor has quit IRC21:52
Pentheuscreiht: We do automagic inventory of everything, down to the disks themselves (everything has its own ID) so we prefer to use our own IDs and provisioning system to build the ring and distribute it automatically everywhere it needs to go21:52
Pentheusthis has been very easy since swift is py and we're py, so I'm just working with the rings directly and distributing when storage arrays detect a new disk21:53
creihtcan you say who "We" is?21:54
Pentheuscreiht: NephoScale, but we're changing the name - everyone hates it :)21:54
Pentheuscreiht: www.nephoscale.com21:54
*** mdomsch has quit IRC21:54
creihtbtw, the ring does have a metadata field that you could use to store ids21:55
creihtnot sure if that is useful21:55
Pentheuscreiht: Yeah, I was going to store it in there if this gave me any trouble, but that small one liner seemed to solve the issue21:56
Pentheuscrieht: I just need to determine if there are any side effects before I file the report :)21:56
*** khussein_ has joined #openstack22:04
*** khussein has quit IRC22:08
*** khussein_ is now known as khussein22:08
vishysoren: ping!22:13
vishyno network love on uml :(22:14
*** Lukey575 has quit IRC22:14
*** khussein has left #openstack22:15
dendrobatesvishy: I'll get in touch with him.22:20
vishyjust trying to use uml with the current build and it almost works22:20
dendrobatesI only tested it on cloud instances that don't have 8021q22:20
grizzletoothI was just setting up the SAIO swift dev stack, and got an odd error during the functests pass.  I don't see it mentioned anywhere online, or in the irclogs:
dendrobatesso it expected failure22:21
grizzletoothinstalled on vanilla ubuntu 10.04 fresh install per the docs22:21
*** mtaylor has joined #openstack22:21
*** mtaylor has joined #openstack22:21
*** ChanServ sets mode: +v mtaylor22:21
*** Ryan_Lane is now known as Ryan_Lane|away22:23
*** londo__ has joined #openstack22:24
*** gondoi has quit IRC22:24
*** rnirmal has quit IRC22:24
*** londo_ has quit IRC22:27
*** adjohn has quit IRC22:27
*** adjohn has joined #openstack22:28
*** kevnfx has quit IRC22:31
*** tasimne has quit IRC22:34
creihtgrizzletooth: looking22:39
creihtof course all the tests pass for me22:42
*** ppetraki has quit IRC22:44
creihtgrizzletooth: did you add the third testing account?22:44
creihtswift-auth-add-user -K devauth test tester3 testing322:44
grizzletoothcreiht: of course they do :)22:44
grizzletoothcreiht: yes - it is in my bash history22:45
creiht(step 14 under "Setting up scripts for running Swift")22:45
*** pvo has quit IRC22:46
grizzletoothswift-auth-add-user -K devauth -a test3 tester3 testing322:46
dendrobatesvishy: did you create a bug about the uml networking?22:46
grizzletoothurp - I think typo - i did test3 tester322:46
creihtand it needs no -a22:46
creiht-a makes it an admin22:47
vishydendrobates: no I wanted to see from soren if i had missed something first22:47
grizzletoothah - I edited the prebious lines thinking I saw the pattern22:47
creihtthe 3rd account has no priviledges, and that test it fails tries to change the perms on a container to make sure they can't22:48
creihtin this case, the user was on a different account, so it was getting an unauthorized error instead of forbidden22:48
dendrobatesvishy: that is actually the issue I was having last night.22:48
grizzletoothcreiht: cool - everything else looked like it was working22:49
dendrobatesI switched to qemu, accidentally and could not reproduce it.22:49
creihtgrizzletooth: awesome22:50
dendrobatesor actually forgot to switch to UML when I tried to reproduce it.22:50
*** pharkmillups has quit IRC22:51
creihtI'm out... thanks everyone for their help getting this done today22:53
creihtdendrobates: the only thing that I'm not sure about is what to do to get the ppas built22:53
creiht(and the version will need to be updated in the debian branch)22:55
*** prabindatta has joined #openstack22:57
dendrobatescreiht: I'll work on that the next few days23:00
*** adjohn has quit IRC23:00
*** DubLo7 has quit IRC23:08
*** AimanA is now known as HouseAway23:09
*** prabindatta has quit IRC23:10
*** dendrobates is now known as dendro-afk23:17
*** johnpur has quit IRC23:17
*** adjohn has joined #openstack23:23
*** rlucio_ has joined #openstack23:26
*** miclorb has quit IRC23:28
*** ArdRigh has joined #openstack23:28
*** allsystemsarego has quit IRC23:28
*** rlucio has quit IRC23:29
*** silassewell has quit IRC23:31
*** abecc has quit IRC23:33
*** Ryan_Lane|away is now known as Ryan_Lane23:33
*** adjohn has quit IRC23:45
*** DesiJat has joined #openstack23:48
*** arthurc has quit IRC23:53
*** grizzletooth has quit IRC23:56

Generated by 2.14.0 by Marius Gedminas - find it at!