Tuesday, 2010-12-07

*** Ryan_Lane has joined #openstack00:03
vishyJordanRinke: yes, but I'm a little concerned with doing that dynamically00:04
vishyJordanRinke: I wonder if i could just add the extra ip to the interface directly without using a bridge00:04
vishyJordanRinke: the problem is that we've settled on vlans for our deployment00:05
vishyJordanRinke: so figuring out how the other modes should work and testing them is more than I can handle.  I made flatdhcp to bridge some of the gap, but it really should be taken over by others who are actually going to use those modes in production00:06
*** aliguori has joined #openstack00:07
*** johnpur has quit IRC00:12
*** dendrobates is now known as dendro-afk00:18
*** jimbaker has quit IRC00:19
*** jc_smith has quit IRC00:24
*** joearnold has quit IRC00:27
*** joearnold has joined #openstack00:27
*** dubsquared has quit IRC00:34
*** metoikos has joined #openstack00:35
*** JordanRinke has quit IRC00:45
*** jc_smith has joined #openstack00:50
*** aliguori has quit IRC00:54
*** hazmat has quit IRC00:56
*** ar1 has joined #openstack01:07
*** dragondm has quit IRC01:11
*** dirakx has joined #openstack01:11
*** ar1 has quit IRC01:12
*** ar1 has joined #openstack01:13
*** jc_smith has quit IRC01:16
*** rlucio has quit IRC01:16
*** iammartian_ has joined #openstack01:20
*** Cybodog has joined #openstack01:23
*** joearnold has quit IRC01:29
*** schisamo has quit IRC01:31
*** maplebed has quit IRC01:33
*** rds__ has quit IRC01:39
*** Ryan_Lane has quit IRC01:40
*** dendro-afk is now known as dendrobates01:45
*** kashyapc has joined #openstack01:47
*** littleidea has quit IRC01:58
*** littleidea has joined #openstack02:13
*** jimbaker has joined #openstack02:35
*** BK_man has quit IRC02:49
*** dubsquared has joined #openstack03:12
dubsquaredjordanrinke:  in the house?03:13
dubsquaredvishy:  you run into issues with dnsmasq getting killed off, but when it respawns after a nova-network restart it comes up with a process owned by 'nobody'03:15
vishyit is supposed to do that03:15
vishy2 processes03:15
vishyone is owned by nobody03:15
vishythe second process runs nova-dhcpbridge with lowered privileges03:16
dubsquaredhrm, these two appear to be the same03:17
dubsquaredyeah, killall dnsmasq, service nova-network restart, then i get two of the same process03:17
vishythat is how dnsmasq works03:18
dubsquaredodd, felt like there was something different on the second one03:18
vishynope, one is a chid process, runs with lowered privilegew03:19
vishybut the command line looks exactly the same03:19
*** kashyapc has quit IRC03:22
*** MarkAtwood has joined #openstack03:22
*** gaveen has quit IRC03:29
*** jbryce_ has joined #openstack03:31
*** jbryce has quit IRC03:33
*** jbryce_ is now known as jbryce03:33
*** sophiap has quit IRC03:44
*** Cybodog has quit IRC03:44
*** jimbaker has quit IRC03:56
*** kashyapc has joined #openstack03:56
*** mdomsch has joined #openstack03:59
*** dubsquared1 has quit IRC04:41
*** omidhdl has joined #openstack04:53
*** MarkAtwood has quit IRC05:03
*** BK_man has joined #openstack05:35
*** mdomsch has quit IRC05:38
*** jimbaker has joined #openstack05:42
*** spectorclan has quit IRC05:47
*** dabo has quit IRC05:47
*** dabo has joined #openstack05:50
*** spectorclan has joined #openstack05:51
*** daleolds has quit IRC05:52
*** that1dude has quit IRC05:55
*** ramkrsna has joined #openstack05:56
*** ramkrsna has joined #openstack05:56
*** BK_man has quit IRC05:59
*** f4m8_ is now known as f4m806:17
*** jfluhmann_ has quit IRC06:21
*** soosfarm has quit IRC06:42
*** soosfarm has joined #openstack06:43
*** jimbaker has quit IRC06:49
*** rlucio has joined #openstack06:52
*** rlucio has quit IRC06:53
*** MarkAtwood has joined #openstack06:55
*** sirp1 has quit IRC07:02
*** krish has joined #openstack07:03
*** krish has left #openstack07:04
*** omidhdl has quit IRC07:06
*** jfluhmann has joined #openstack07:14
*** jimbaker has joined #openstack07:18
*** guigui has joined #openstack07:18
*** jbaker_ has joined #openstack07:18
*** jimbaker has quit IRC07:18
*** jfluhmann has quit IRC07:23
*** omidhdl has joined #openstack07:25
*** jbaker_ has quit IRC07:25
*** jfluhmann has joined #openstack07:26
*** omidhdl1 has joined #openstack07:26
*** omidhdl has quit IRC07:26
*** stewart has quit IRC07:34
*** jfluhmann has quit IRC07:36
*** jfluhmann has joined #openstack07:37
*** ibarrera has joined #openstack07:40
*** jfluhmann has quit IRC07:49
*** omidhdl has joined #openstack07:54
*** omidhdl1 has quit IRC07:54
*** guigui has quit IRC08:03
*** GasbaKid has joined #openstack08:09
*** jfluhmann has joined #openstack08:10
*** metoikos has quit IRC08:13
*** calavera has joined #openstack08:14
*** ramkrsna has quit IRC08:16
*** zaitcev has quit IRC08:18
*** joshuamckenty has left #openstack08:23
*** ctennis has quit IRC08:28
*** almaisan-away is now known as al-maisan08:37
*** HugoKuo_ has joined #openstack09:11
*** HugoKuo has quit IRC09:13
*** kashyapc has quit IRC09:19
*** kashyapc has joined #openstack09:20
*** stewart has joined #openstack09:22
*** irahgel has joined #openstack09:22
*** larstobi has joined #openstack09:23
*** ArdRigh has joined #openstack09:24
*** jkakar has joined #openstack09:32
*** allsystemsarego has joined #openstack09:37
*** allsystemsarego has joined #openstack09:37
*** arthurc has joined #openstack09:45
*** Adiantum has quit IRC09:48
*** ctennis has joined #openstack09:57
*** dubsquared has quit IRC09:58
*** omidhdl1 has joined #openstack09:58
*** omidhdl has quit IRC09:59
*** reldan has joined #openstack10:00
*** stewart has quit IRC10:06
*** reldan has quit IRC10:15
*** befreax has joined #openstack10:18
*** omidhdl1 has quit IRC10:35
alekibangosoren: are u here? there are no nova  packages in  http://ppa.launchpad.net/nova-core/ppa/ubuntu/pool/main/10:37
ttxalekibango: I think you want to use the "trunk" ppa.10:38
alekibangoah, sorry i got confused prolly10:39
*** ctennis has quit IRC10:39
alekibangoi somehow got lost in the internet lol10:39
ttxit does that sometimes10:39
alekibangoty ttx10:40
*** omidhdl has joined #openstack10:41
*** vivekbabs has joined #openstack10:46
*** ramkrsna has joined #openstack10:54
*** kashyapc has quit IRC10:56
*** ArdRigh has quit IRC10:58
*** kashyapc has joined #openstack10:59
*** reldan has joined #openstack11:03
*** kashyapc has quit IRC11:04
*** ianweller has quit IRC11:05
uvirtbotNew bug: #686438 in glance "Cannot create image: Json Circular reference detected" [Undecided,New] https://launchpad.net/bugs/68643811:06
*** ianweller has joined #openstack11:06
chmouelany 'glance' gurus around ?11:08
chmoueli am wondering at first what the teller and the parallax, i am assuming parallax is for uploading/update/post etc.. and teller is to serve quickly the files?11:13
chmoueli can see only metadatas as well there, is there already a facilty to upload image itself ?11:14
*** kashyapc has joined #openstack11:16
sorenchmouel: I /think/ the idea is that you put the image somewhere by yourself (in swift, S3, whatever) and register it with glance.11:16
vivekbabsi am trying to use nova compute (revision 440 from trunk ppa on Ubuntu Maverick) with XenServer.  in nova.conf i've set up "connection_type=xenapi" as well as xenapi_connection_{url, username, password} and restarted all nova services after that.  if i do euca-run-instances (as described in http://wiki.openstack.org/RunningNova) i get "instance not present" exception (http://paste.openstack.org/show/242/).11:20
chmouelyeah i think I just realized it from the teller code, I am trying to figure out how I can that reference on the parallax entry so far I have come to that point http://paste.openstack.org/show/243/11:21
vivekbabsperhaps i'm missing a crucial step.  can someone help?11:21
chmouelvivekbabs: virsh list by any chance ?11:21
vivekbabsvirsh list is empty11:21
chmouelanything in the logs that would give u the hint ?11:22
*** ramkrsna has quit IRC11:22
vivekbabschmouel: logs are normal (i've turned on verbose in nova.conf) except for the exception i pasted at http://paste.openstack.org/show/242/11:24
vivekbabsperhaps a basic question, should the steps to run with XenServer be much different from local node + kvm combo, apart from modifying nova.conf with xenapi params?11:25
chmouelvivekbabs: not entirely sure to be honest11:32
*** rds__ has joined #openstack11:50
*** fabiand has quit IRC11:51
*** miclorb_ has quit IRC12:01
*** ramkrsna has joined #openstack12:03
*** reldan has quit IRC12:11
*** reldan has joined #openstack12:17
*** westmaas has joined #openstack12:22
*** miclorb_ has joined #openstack12:26
*** kashyapc has quit IRC12:33
*** guigui1 has joined #openstack12:54
vivekbabsjust noticed that i am getting "exceptions.DeprecationWarning: Use of empty request context is deprecated" from  is_admin_context and then "raise Exception('die')".  this is perhaps the real reason for an instance not getting launched on XenServer.  what could be wrong?12:57
*** Podilarius has joined #openstack12:57
*** fabiand has joined #openstack12:58
*** GasbaKid has quit IRC13:01
*** tdn has joined #openstack13:01
*** GasbaKid has joined #openstack13:01
*** stewart has joined #openstack13:03
*** hazmat has joined #openstack13:07
*** Cybo has quit IRC13:08
alekibangovivekbabs: i would check why the request context is empty, but thats only guessing  (in source)13:11
*** dirakx has quit IRC13:12
*** omidhdl has quit IRC13:18
*** omidhdl1 has joined #openstack13:18
vivekbabsalekibango: yes, i was just doing that.  i've put debug statements and gotten some stack traces for cases when there is a valid context and when there is none.  Trying to make sense of those now.13:25
*** ramkrsna has quit IRC13:29
*** hazmat has quit IRC13:35
*** reldan has quit IRC13:36
*** aliguori has joined #openstack13:44
*** omidhdl1 has left #openstack13:47
sorenvivekbabs: Perhaps the Xen API driver hasn't been updated to always pass a context.13:49
vivekbabssoren: yes, soren, i too spotted it :)13:49
sorenvivekbabs: Not doing so is an error, and during development we made it a fatal error so that we'd catch it before release. You've just caught one of those cases :)13:50
vivekbabsi'm going to change it and run again ... i am going to use context.get_admin_context() as done in libvirt_conn.py13:50
vivekbabshope that's OK13:50
sandywalshvivekbabs, I think your instance is running13:50
sandywalshbut it may just need to be rebooted13:51
sandywalshwe've been seeing the same thing13:51
sandywalshif your xensource.log isn't reporting any problems13:51
sandywalsh(the die message isn't related, I get that too)13:51
sandywalshif you reboot the image, it may start properly13:52
sandywalshvivekbabs, i.e. http://paste.openstack.org/show/244/13:52
vivekbabsand now i'm getting "Exception: Found no network for bridge br100" from "/usr/lib/pymodules/python2.6/nova/virt/xenapi.py", line 250, in _find_network_with_bridge .. going to look at that13:52
sandywalshvivekbabs, networking doesn't work with xenserver yet. known issue13:53
vivekbabssandywalsh: i'll check out my XenServer logs13:55
vivekbabssandywalsh: "known issue" ... is there a bug for it?13:57
*** hazmat has joined #openstack13:59
sandywalshvivekbabs, hmm, good question. Lemme see13:59
sandywalshvivekbabs, not that I see. I should note as well that we're using the lp:~armando-migliaccio/nova/xenapi-refactoring branch for the freshest xen support14:02
*** ctennis has joined #openstack14:03
vivekbabssandywalsh: ok, that's a nice hint :).  i'll download that branch and use it14:04
larstobiHi, I'm trying to get Swift-1.1.0 to work, but I'm stuck. I can't get auth-server to run — it dies right after writing it's pid to /var/run/swift/auth-server.pid. Also, the unit tests are failing: http://paste.openstack.org/show/245/14:06
vivekbabssandywalsh: btw, what do you mean by "networking doesn't work with xenserver yet"?  has it been addressed in xenapi-refactoring?14:07
sandywalshvivekbabs, not that I'm aware of. afaik the only way into the instance is via the console for now. I could be wrong.14:08
vivekbabssandywalsh: ok.  thank you.14:08
sandywalshvivekbabs, np14:08
*** gundlach has joined #openstack14:15
*** ppetraki has joined #openstack14:18
*** gondoi has joined #openstack14:20
larstobihere's the output of "strace swift-init auth-server start": http://paste.openstack.org/show/246/14:20
*** reldan has joined #openstack14:20
*** novahelp has joined #openstack14:21
novahelpi have a problem on installing nova,can you help me14:24
*** kashyapc has joined #openstack14:28
sorenNoone can help you if you don't explain what the problem is.14:28
alekibangonovahelp: try asking real questions14:29
annegentlenovahelp: sure go ahead. I recently rearranged the wiki pages for Nova installation so hopefully you're testing out the instructions on http://wiki.openstack.org/NovaInstall14:29
chmouelcreith: so generally speaking should we use swift.common.client or python-cloudfiles when interfacing with swift?14:31
*** gundlach has quit IRC14:33
novahelpthank you!my problem is nova-network service cannot load.The article says create a new network 1 64. i install nova on virtualbox(default setting),can you tell me how to config the ip address of virtualbox to match the network
*** gundlach has joined #openstack14:33
novahelpyes,i follow the article http://wiki.openstack.org/NovaInstall,but i cannot run it,can you tell me more about the network configuration.14:35
alekibangonovahelp: it was creating network definitions in nova  :) that is not ifconfig line14:35
annegentlenovahelp: when I type ifconfig in my virtual box, I see
annegentlenovahelp: that's the local loopback14:36
alekibangonovahelp: which article are u following14:36
annegentleI'm guessing http://wiki.openstack.org/RunningNova14:37
novahelpthe ip of lo is, but the eth0 will get a ip from dhcp,can you tell me which method do you select and which nics do you use?14:37
alekibangonova-manage network create 3 1614:38
*** ctennis has quit IRC14:38
* chmouel think you probably want to have basic linux knowledge before playing with nova at this stage14:38
alekibangoannegentle:  this line somehow is missing in your docs :)14:38
alekibangochmouel: that would be great, but the article forgot to tell how to add networks to nova14:39
chmouelyeah true that14:39
*** littleidea has quit IRC14:39
annegentlealekibango: yep just added it LOL14:39
annegentlealekibango: I had fixed it on the nova.openstack.org site and neglected the wiki - time to put RST in the wiki...14:40
novahelpin my virtualbox,the ip of lo is, the eth0 is and the virbr0 is That is right?14:41
*** littleidea has joined #openstack14:41
novahelpannegentle: can you add the the network configuration of virtualbox? i am a new player..so,dont smile me.14:42
*** littleidea has quit IRC14:42
novahelpchmouel: yes,i think so.i will try my best to study linux knowledge.thank you14:43
chmouelnovahelp: lo is loopback interface on linux a standard thing that you don't have to worry about for openstack14:44
*** littleidea has joined #openstack14:46
novahelpchmouel: ok,i dont care it. but i dont make sure my virtualbox configuration whether is right. And i try many times to install nova, and failed, so i come to here for help.14:46
*** pandemicsyn has quit IRC14:46
*** _0x44 has quit IRC14:46
*** karmabot has quit IRC14:46
*** _0x44 has joined #openstack14:47
*** ChanServ sets mode: +v _0x4414:47
*** pandemicsyn has joined #openstack14:47
annegentlenovahelp: I'm glad you're trying Nova. Have you tried alekibango's suggestion? sudo nova-manage network create 3 16?14:48
*** zheng_li has joined #openstack14:48
alekibangothat creates 3 networks having size 16 addresses iirc14:48
annegentlealekibango: 3 networks having 16 IP addresses, yep14:48
*** littleidea has quit IRC14:49
novahelpannegentle: yes.i try to create that network,but still cannot load service nove-network.And i would like to confirm my network setting of virtualbox whether is right?14:49
alekibangoannegentle: its not explained well in the install manual imho14:49
alekibangoor running manual14:49
*** karmabot has joined #openstack14:49
annegentlealekibango: yep, agreed.14:50
alekibangowe need to cover all those nova-manage commands14:50
annegentleso help me fix it :)14:50
alekibangoannegentle: having my own problems right now, but i will be back on it next week14:50
annegentlealekibango: no worries. Just using your technique :)14:51
novahelpannegentle: in my virtualbox,the ip of lo is, the eth0 is and the virbr0 is The configuration is right? Network access method is using NAT.14:52
*** ppetraki has quit IRC14:53
alekibangoannegentle: problem is that i do not know enough about nova networking to write the doc. some nova devels should do those...14:54
*** ppetraki has joined #openstack14:54
annegentlenovahelp: ok, here's the thing. You're running from a script called nova.sh. Within that script, the "nova-manage network create ..." command has already been run for you.14:55
*** ppetraki has quit IRC14:55
annegentlenovahelp: so the RunningNova page isn't quite what you need in your situation14:55
alekibangoah, yes, good tip14:55
*** ppetraki has joined #openstack14:55
novahelpannegentle: the first i use that nova.sh script to install nova,failed..14:55
*** reldan has quit IRC14:55
alekibangonovahelp: which os?14:55
*** reldan has joined #openstack14:56
novahelpubuntus 64bit server14:56
*** jfluhmann has quit IRC14:56
alekibangohow it failed?14:56
novahelpand virtualbox 3.2.1214:56
*** reldan_ has joined #openstack14:56
annegentlenovahelp: I have also seen "Permission denied /var/openstack/nova-network.pid" errors on VirtualBox when I didn't shut down all the Nova services properly14:57
novahelpannegentle: whatever i can not load nove-network service.14:58
annegentlenovahelp: could you paste what you see in the network screen into http://paste.openstack.org/ and then put that URL in the channel?14:59
novahelpso i guess whether my network configuration of virtualbox is right.can you tell me the right configuration about virtualbox?14:59
*** zheng_li has quit IRC14:59
novahelpannegentle: OK. please wait a moment.15:00
*** reldan has quit IRC15:00
*** reldan_ is now known as reldan15:00
alekibangobtw i am trying to make installation of nova automated using fai and i think i will finally have it tonight15:00
annegentlenovahelp: yes, your network config is fine. Your virtualbox is probably fine also. We'll get to the root of it. :)15:00
*** vivekbabs has quit IRC15:01
*** dubsquared has joined #openstack15:02
*** f4m8 is now known as f4m8_15:03
novahelpannegentle: now i am pasting my log,can you wait me?15:03
*** daleolds has joined #openstack15:04
annegentlenovahelp: sure, take your time.15:04
annegentlenovahelp: no pressure :)15:04
*** zheng_li has joined #openstack15:04
alekibangonovahelp:  clouds are still not  easy to use. if you have small number of servers, you might wait with clouds until it gets easier.15:04
alekibangoannegentle: did you see 'no pressure' 10:10    video?15:05
alekibangoannegentle: i  would say authors should be in prison now...15:06
annegentlealekibango: eek, no.15:06
novahelpalekibango: i bought a few server to try cloud environment, and i would not wait for long time.i like chalenge15:07
alekibangonovahelp: few = 2, 5, 30, 100, or more?15:08
*** allsystemsarego has quit IRC15:08
*** allsystemsarego has joined #openstack15:10
alekibangoannegentle 5 minutes of real tough propaganda, enjoy: http://www.youtube.com/watch?v=5FkB4uiizVo   (and remember, this is really official piece from   http://www.1010global.org/no-pressure  )15:11
notmynamechmouel: re client.py vs python-cloudfiles: I would suggest client.py as included in swift. it's faster. python-cloudfiles has some nice features, but it's geared more to Rackspace's specific product (CDN etc). also python-cloudfiles is slow15:12
novahelpalekibango: 2..15:12
*** zheng_li has quit IRC15:13
chmouelnotmyname: sweet, cheers..15:13
*** dubsquared has quit IRC15:13
alekibangoannegentle:  and if you really think co2 is a problem, see http://scienceandpublicpolicy.org/   or google climategate :)15:14
*** zheng_li has joined #openstack15:14
notmynamechmouel: in truth, we have grand plans to rewrite the python-cloudfiles library to use client.py (cause it's just that much better) :)15:14
*** zheng_li has joined #openstack15:14
chmouelyeah that was going to be my next question :)15:14
*** zheng_li has quit IRC15:15
notmynameno idea when that will happen, but it's on the "we should probably do this sooner than later" list15:15
*** zheng_li has joined #openstack15:15
novahelpannegentle: soon i will be ok15:17
novahelpannegentle: i pasted my log. the url is http://paste.openstack.org/show/247/15:19
annegentlenovahelp: thanks. Okay. I will try my best. Let's start from the beginning. Did you download the .iso or the .zip from http://wiki.openstack.org/NovaVirtually?15:23
annegentleand I will need help - I'm no expert :)15:23
alekibangoannegentle: but you are brave :)  thats often enough :)15:23
novahelpannegentle: i downloaded both15:23
annegentlenovahelp: okay. Then did you unzip the .zip? and find a .vmdk file?15:24
novahelpannegentle: okey,i unzip the zip file,please wait a moment.15:24
annegentlenovahelp: It's named Ubuntu 64 Server.vmdk. Sorry if I'm going too fast. In real life I talk too fast too. :)15:25
novahelpdont be mind.i can follow you~15:25
chmouelis it worthwhile to log bunch of bugs report against glance integration and component or there is an active branch being worked on somewhere else than trunk ATM ?15:25
novahelpbefore the installion,can i confirm the virtualbox setting?15:26
novahelpannegentle: before the installion,can i confirm the virtualbox setting?15:26
annegentlesure - I am running 3.2.1015:27
annegentlenovahelp: but that should be okay15:27
annegentlenovahelp: we can compare15:27
novahelphow to compare?15:27
novahelpannegentle: how to compare?15:28
*** lorin has joined #openstack15:28
annegentlenovahelp: if you look at the screenshots on http://wiki.openstack.org/NovaVirtually, you can compare to your screens. Tell me if they differ.15:28
*** lorin is now known as lorinh15:28
*** lorinh is now known as bubble-bobble15:28
annegentlenovahelp: and, I want to improve that page with steps - so I will write down as we go.15:28
annegentlenovahelp: so, first, unzip the zip file to any directory.15:29
annegentlenovahelp: then launch Virtualbox15:29
novahelpok.now unziping...15:29
novahelpnow 85%15:29
novahelpannegentle: 87%15:29
*** bubble-bobble is now known as lorin-hochstein15:30
*** mdomsch has joined #openstack15:30
*** lorin-hochstein has quit IRC15:30
ttxjaypipes: about Glance "clients" spec, looks complete to me -- can I mark it "Implemented" ?15:32
ttxjaypipes: also teller-api looks almost ripe ?15:32
alekibangothat sounds promising :)15:34
*** crago has joined #openstack15:34
*** johnpur has joined #openstack15:36
*** ChanServ sets mode: +v johnpur15:36
*** crago has quit IRC15:37
*** littleidea has joined #openstack15:43
*** jfluhmann has joined #openstack15:50
creihtchmouel: client choice is up to you, but I recommend swift.common.client :)15:52
creihtlarstobi: Check that you have write permissions to /etc/swift/auth.db15:52
*** GasbaKid has quit IRC15:54
*** sirp1 has joined #openstack15:55
larstobicreiht: thanks or responding. I tracked it down to the auth.db as well. It fixed the auth-server start dying. The auth-server now runs. However, the ./.unittests still reports the same 6 errors.15:55
creihtit is probably a permissions issue as well... let me go refresh my memory as to where those tests are running15:56
larstobiokay, thanks15:56
uvirtbotNew bug: #686621 in nova "Context is not passed when creating a new server via OpenStack API" [Undecided,New] https://launchpad.net/bugs/68662115:56
*** GasbaKid has joined #openstack15:56
creihthrm, I thought all the db tests were done in memory though...15:57
creihtlarstobi: what version of webob do you have installed?15:58
larstobicreiht: 1.015:58
creihtand are you running the latest swift trunk, or an older version?15:59
larstobiI'm running 1.1.015:59
*** xfaf has joined #openstack16:00
creihtthere may be a bug with webob 1.0 that wasn't backported to 1.1.016:00
*** zul has quit IRC16:01
*** xfaf is now known as zul16:01
creihtlarstobi: give me a few minutes to look into it16:01
larstobithanks :)16:01
creihtshort answer is that I don't think it should prevent you from experimenting with it16:02
larstobiyes, it seems to work so far16:02
creihtIt just wont work entirely correctly for certain error cases16:02
*** dragondm has joined #openstack16:07
creihtlarstobi: ok confirmed... I'll get that fix into the 1.1 code branch shortly16:07
creihtIt was just some fixes to the unit tests16:07
creihtthe code should work fine16:08
larstobicreiht: good :) thanks!16:08
*** miclorb_ has quit IRC16:08
*** jimbaker has joined #openstack16:11
*** jbaker_ has joined #openstack16:12
*** jimbaker has quit IRC16:12
*** guigui1 has quit IRC16:12
*** BK_man has joined #openstack16:14
*** jdarcy has joined #openstack16:18
*** calavera has quit IRC16:19
creihtttx: what's the rule for releasing an incremental update to a past releease?16:20
ttxcreiht: i don't think we have any rule set yet16:20
creihtttx: for example we are beginning to get to a point to where we have enough bug fixes in the 1.1 branch that we should make a 1.1.1 release at some point16:20
*** ibarrera has quit IRC16:21
creihtThese are all backported fixes from the 1.2 current working branch16:21
*** BK_man has quit IRC16:21
ttxcreiht: I have no strong opinion on that (yet), i'll have to sync with dendrobates to get his view16:21
ttxdunno how much of this was already discussed and decided16:22
creihtttx: ok no rush, but it is on the horizon for us, so it would be nice to get that figured out16:22
creihtI don't think it has at all16:22
ttxcreiht: I'll discuss it with dendrobates when he comes back from vacation16:22
*** larstobi has quit IRC16:23
*** ibarrera has joined #openstack16:24
*** fabiand has quit IRC16:25
*** miclorb has joined #openstack16:26
*** fabiand has joined #openstack16:32
creihtinteresting: http://www.database.com/16:33
*** kashyapc has quit IRC16:34
chmouelnot sure about product but at least they have a good domain name16:35
alekibango... it is service16:35
ttxcreiht: the free offer sounds interesting for a number of scenarios16:35
*** vivekbabs has joined #openstack16:40
*** bfschott has joined #openstack16:44
*** fabiand has quit IRC16:46
*** kashyapc has joined #openstack16:50
vishyvivekbabs, soren: i don't know if you are still around, but xenapi is broken16:51
vishythe exception die was supposed to get removed at some point as well, but in any case: https://code.launchpad.net/~armando-migliaccio/nova/xenapi-refactoring/+merge/42133 fixes the empty context16:55
vivekbabsvishy, do you know if merging of lp:~armando-migliaccio/nova/xenapi-refactoring will fix xenapi?  (there may still be some networking issues as mentioned by sandywalsh.)16:55
*** ilya has joined #openstack16:56
vishyvivekbabs: afaict, yes but I haven't personally tested it as all16:56
*** ilya is now known as Adiantum16:56
vivekbabsvishy: ok16:56
*** dirakx has joined #openstack16:57
*** piken is now known as piken_lunch16:58
creihtmtaylor: would it be possible to get a webob 1.0 ppa in the swift ppas?16:59
jk0vishy: we were hoping that xenapi-refactor would make it into trunk today :)16:59
creihts/ppa/debian package16:59
mtaylorcreiht: sure16:59
jk0vivekbabs: that new branch will likely fix any problems you're seeing. we've been using it internally for several days now17:00
vivekbabsjko: that's great to hear17:00
vishyjk0: yeah I'm looking, there seems to be a bit of contention around novadeps17:00
creihtmtaylor: https://launchpad.net/ubuntu/+source/python-webob17:00
jk0vishy: I thought he removed those in the latest merge17:01
creihtlooks like they have 1.0 packaged for natty, so would be nice to just package that for lucid17:01
jk0maybe not tho17:01
creihtmtaylor: the main reason for that is so that people will dev against webob 1.0, so that we don't accidently write tests that don't work with 1.0 :)17:02
mtayloryes. I think this is a good idea17:03
creihtand once we get the package available, it would be nice to upgrade hudson17:03
*** jonkelly has joined #openstack17:03
* mtaylor actually needs to upgrade the hudson machine to maverick...17:03
*** rlucio has joined #openstack17:05
vishyjk0: ah you are correct17:06
*** hggdh has quit IRC17:07
* ttx disappears for a couple of hours17:07
bfschotthey guys, just got our team steered to transition from Euca to OS, so if you see a bunch of new USC-ISI faces on the irc/lists, say hello.17:09
bfschottwe're going to be submitting some blueprints for multi-architecture support x86+gpu, SGI ultra-violet shared memory machines, and Tilera processors17:09
alekibangobfschott: you are welcome17:10
alekibangobfschott:   http://ai.isi.edu/ ?17:11
*** hggdh has joined #openstack17:11
alekibangorather http://www.isi.edu/17:11
bfschottbut yes, we're a group in Arlington, VA17:12
bfschottwow, my web page is out of date :-)17:12
edaybfschott: what are USC folks doing all the way out in VA? :)17:13
annegentlebfschott: excellent. Feel free to use the wiki for longer specs for the blueprints.17:14
bfschottwill do17:14
*** vivekbabs has quit IRC17:15
alekibangobfschott: interesting... :)     can you please tell why did you pick openstack? what do you expect from the move?17:20
openstackhudsonProject nova-tarmac build #41,587: ABORTED in 12 min: http://hudson.openstack.org/job/nova-tarmac/41587/17:20
*** openstackhudson has quit IRC17:20
*** jkakar has quit IRC17:21
*** kashyapc has quit IRC17:21
creihtmtaylor: btw, I just found out that we also need this patch in the deb for webob: http://bitbucket.org/ianb/webob/changeset/4a767f7d165917:21
mtaylorright after I uploaded the source packages...17:22
bfschottalek: many struggles with getting Eucalyptus to do what we want17:23
bfschottcode, compile, deploy mechanism hard to work with17:23
chmouelwhat's so good about those SGI ultra-violet hardware compared to dell or other hardware?17:24
* mtaylor stabs creiht 17:24
bfschottwe added some architecture changes to the node controller XML spec and Axis2/c to pass back architecture field, list of accelerators, etc. as part of "describe-resources" poll17:24
chmouelyeah i should have google first17:24
chmouelhttp://www.sgi.com/products/servers/altix/uv/ => "with architectural support to 262,144 cores "17:24
mtaylorholy cripes! that's a lot of cores17:25
chmouelwould love to see ostack scale on that17:25
bfschottSGI UV is big memory machine (we are getting a 1TB main memory machine, 256 cores)17:25
alekibangobfschott: so its adaptability i guess17:25
chmouelor even linux actually :)17:25
bfschottin a single system image, or partitioned down into smaller VMs17:25
*** openstackhudson has joined #openstack17:26
bfschottwhat we want to do is make sure when somebody does a run-instances with architecture dependencies the scheduler goes to the right physical nodes17:26
bfschottwill work for toys :-)17:27
*** BK_man has joined #openstack17:27
bfschotthardware is coming week of 12/20 so I can't wait17:28
bfschottalek: yes modular architecture is key17:29
alekibangobfschott: i believe you made a good choice, even if this is still so young project.. :)17:29
*** JordanRinke has joined #openstack17:33
*** Ryan_Lane has joined #openstack17:34
*** novahelp has left #openstack17:34
*** iammartian has joined #openstack17:35
bfschottalek: I was also impressed with the core team's guitar hero or was it rockband skills17:36
alekibangobfschott: i do play guitar ;)17:40
edayalekibango: I play drums, but failed horribly at rockband. it kinda didn't seem to help :)17:40
mtaylorsoren, creiht: hudson machine upgraded to maverick17:40
alekibangoeday: please join #ninjam if you want to play over internet :)17:40
alekibangohere i play some guitar  --  live ninjam session over internet http://autosong.ninjam.com/mp3.php/20101105_2131_0000-Ogner-AD_Bass-Linuxer-itsme.mp317:41
alekibangobfschott: but really, this team is great.17:42
*** BK_man has quit IRC17:43
alekibangoha, ninjam official channel is not on this server :) but it is really fun, if you are musician, you can play with live people over internet...  there is over 30,185 hours of music recorded via ninjam... and sometimes the jam is pretty great17:44
*** befreax has quit IRC17:45
*** arthurc has quit IRC17:46
*** brd_from_italy has joined #openstack17:52
*** joearnold has joined #openstack17:55
*** al-maisan is now known as almaisan-away18:03
*** bfschott has quit IRC18:04
rlucioi'm still looking for approval on the two-liner fix for  lp677475 .. anyone have a minute to look?  https://code.launchpad.net/~rlucio/nova/lp677475/+merge/4258418:06
*** zheng_li has quit IRC18:06
mtaylorcreiht: ok. new packages uploded - should hit soon I guess18:07
alekibangomtaylor: which packages? trunk ppa?18:08
*** robrightscale has joined #openstack18:08
mtayloralekibango: new python-webob packages18:08
*** bfschott has joined #openstack18:09
*** gundlach has quit IRC18:10
edayanyone here know enough about nova ldap to review/approve rlane's ldap branches? (vishy already did)18:18
edayxtoddx: perhaps you? :)18:18
xtoddxi'll pretend18:19
xtoddxi'll check it out after our scrumm in a few minutes18:19
edayxtoddx: thanks! https://code.launchpad.net/~rlane/nova/ldap-user-modify-only/+merge/4197118:20
vishyeday: are you still in the bay area?18:20
edayvishy: I am for this week, I leave sunday18:20
vishyeday: cool, perhaps we could meet up for a hack day this week before you leave18:21
edayvishy: when will you guys be at ames?18:22
*** rlucio has quit IRC18:27
Ryan_Laneif anyone needs to talk to me about them, I'm generally about :)18:30
Ryan_Lane(the ldap branches, that is)18:31
*** rlucio has joined #openstack18:35
*** piken_lunch is now known as piken18:36
*** zheng_li has joined #openstack18:36
jaypipesttx: hey, sorry, late start today... xmas shopping delays.  as for clients API spec, it's almost complete, as is teller-api.  Should be done by tomorrow or  Thursday.18:38
*** bfschott has quit IRC18:39
*** bfschott has joined #openstack18:39
vishyRyan_Lane: btw, the python standard is to not use \ inside of parentheses, and to line up the next line after the (18:42
alekibangojaypipes: i stopped doing this years ago and it makes me and my family happier, especially when i see others going crazy in this season...18:43
vishyanyone familiar with fdisk/partitioning?18:45
*** iammartian has quit IRC18:48
*** iammartian_ is now known as iammartian18:48
vishyI'm trying to figure out why I'm getting the following physical/logical extents messages for my partitioned disks: http://pastie.org/135639418:49
alekibangojaypipes:  Btw Jesus was not borned on 25th anyway -- it was the babylonian ho having baby that day...   To understand how 'funny' this is, please try to Imagine whole Israel celebrating birth of Moses on 20th april, wearing torches in the night, with black hindu symbols of peace in white circles on red flags...   (hint: Adolf Hitler was born that day)...    i am not joking. the contrast is similar.18:50
jaypipesalekibango: hehe, well, we try to limit the presents...mostly get gifts for kids in the extended family. :)18:50
alekibangoyes i always do give  something  to kids just to not make them sad... but thats it18:50
*** reldan has quit IRC18:51
jaypipesvishy: never seen that before. :(18:52
edaymtaylor: nova tarmac seems to be hung.. should I just kill or do you want to inspect?18:52
alekibangovishy: gentleman without cylinders?  did you try other tools?18:53
vishyalekibango: the cylinder thing is standard when using a disk image18:54
alekibangouhm it looks strange...18:54
*** ibarrera has quit IRC18:54
alekibangogpart might maybe detect something?18:54
vishyalekibango: the logical/physical extents is strange18:54
*** fabiand has joined #openstack18:54
jaypipeschmouel: this is quite bizarre: https://bugs.launchpad.net/glance/+bug/686438.  What is trying to be serialized there?  Could you attach the object to the bug report so I can include it in a bug fix?18:55
uvirtbotLaunchpad bug 686438 in glance "Cannot create image: Json Circular reference detected" [Undecided,New]18:55
edayxtoddx, vishy: if you guys have any time today, could you also try to review my 3 compute refactor branches? want to get those in before I move forward18:56
vishyalekibango: hmm gpart is giving slightly odd results18:56
alekibangosfdisk/cfdisk might also be different18:56
alekibangoi dont know, i never seen this too18:56
alekibangoit might be the silent data corruption maxwell daemon18:57
alekibangoyou know maxwell's daemon, do you :)18:57
jaypipeseday: I can do a review, too :) or at least give it a shot.18:57
alekibangothats the malevolent creature responsible for data corruption :))18:58
*** bfschott has quit IRC18:58
*** iammartian_ has joined #openstack18:59
edayjaypipes: oh, you too.. didn't know if you were busy with glance :)18:59
*** bfschott has joined #openstack18:59
jaypipeseday: :)18:59
edaythe more the better, it's a realy party over on the merge reviews page today19:00
*** allsystemsarego has quit IRC19:00
mtayloreday: just kill it19:00
creihtmtaylor: awesome, and thanks!19:01
mtayloreday: I upgraded the machine to maverick today19:01
*** allsystemsarego has joined #openstack19:01
openstackhudsonProject nova-tarmac build #41,588: ABORTED in 1 hr 35 min: http://hudson.openstack.org/job/nova-tarmac/41588/19:01
edaymtaylor: ahh, ok19:01
alekibangosoren: would you accept my patches making your packages more debian compatible?  (if i will do them)19:05
*** BK_man has joined #openstack19:05
*** iammartian_ has quit IRC19:05
*** BK_man_ has joined #openstack19:06
*** BK_man has quit IRC19:06
*** BK_man_ is now known as BK_man19:06
sandywalshany objectstore pro's here?19:08
notmynamesandywalsh: what's up?19:08
sandywalshnotmyname, hey! the xenapi is requesting an image with the following url:
sandywalshnotmyname, but objectstore is 404'ing because the filename is .../image.encrypted19:09
openstackhudsonProject nova build #283: SUCCESS in 1 min 13 sec: http://hudson.openstack.org/job/nova/283/19:09
sandywalshnotmyname, I assume decryption would be done client-side (from the caller) correct?19:10
notmynamesandywalsh: I think I may have responded in error :-) to me objectstore == swift19:10
sandywalshnotmyname, sorry nova objectstore19:10
Ryan_Lanevishy: ah, I didn't realize I used \ inside of any parentheses. On lines without parentheses, I should be using \, correct?19:11
vishyyes, or wrapping in parens19:11
Ryan_LaneI'm not really used to pep8, though I did read it a couple of times19:11
vishywrapping in parens is a bit more common i think19:11
Ryan_Laneah. ok. will do that instead then19:12
Ryan_Lanevishy: any thoughts on my proposed changes on the schema?19:13
vishyI commented already, no?19:14
Ryan_LaneI added some more comments19:14
jk0any chance we can get one more look at https://code.launchpad.net/~armando-migliaccio/nova/xenapi-refactoring/+merge/42133 ?19:14
Ryan_Lanevishy: I understand about isAdmin, and agree it is likely a good attribute to keep, though still believe it should be renamed as isNovaAdmin19:20
vishyi like the new name19:20
vishyjust trying to avoid changing ldap schema unless it is absolutely necessary19:20
sorenalekibango: Depends on the patches.19:20
Ryan_Lanewell, the schema changes should make things simplier19:21
alekibangosoren: ic, i think i will try19:21
Ryan_Lanemaking changes now is easier than later19:21
sorenalekibango: How are they not Debian compatible now?19:21
Ryan_Lanevishy: I think we should make it as ideal as possible, and finalize it19:21
alekibangofor example, libvirt  group in debian -> ubuntu libvirtd19:22
sorenalekibango: Then no, of course not. That would break them on Ubuntu.19:22
alekibangosoren:  that can be automated to work both ways..19:22
sorenalekibango: Then sure. Why wouldn't I?19:22
Ryan_Lanevishy: and from an LDAP admin's perspective, I think the isAdmin name change is a neccessary one19:22
sorenAw, crap.19:23
Ryan_Lanevishy: and for projects/roles, requiring they exist in separate OUs is likely a good solution for separating them. heirarchy in LDAP can be problematic, but can be useful on occasions like this19:25
vishyRyan_Lane: I agree19:26
*** reldan has joined #openstack19:27
edayRyan_Lane: can you update your opendj branch with your name in authors? (see merge req output)19:27
Ryan_Laneless custom schema is almost always better, if it is avoidable :)19:27
Ryan_Laneeday: ah. yes. I'll do that.19:27
edayRyan_Lane: actually, have you signed the CLA already too?19:29
Ryan_LaneI'm somewhat unfamiliar with launchpad19:29
Ryan_Lanehow do I add myself as an author?19:30
ttxTeam meeting in 90 min in #openstack-meeting !19:30
edaynot sure if those checks are in yet or not, but wanted to make sure thats not it either19:30
edayedit the Authors file :)19:30
*** jaypipes has quit IRC19:32
sandywalshso, any nova objectstore people in the room?19:33
*** Ryan_Lane has quit IRC19:35
edaysandywalsh: that is mainly soren, but others might be able to answer :)19:35
sandywalsheday, thx19:36
*** Ryan_Lane has joined #openstack19:36
Ryan_Lanevishy: ok. I'll change the branch to make the schema changes to project, and I'll add a version flag so that people can use the old schema until they migrate19:36
*** BK_man_ has joined #openstack19:36
*** BK_man has quit IRC19:36
*** BK_man_ is now known as BK_man19:36
Ryan_Laneeday: added myself to Authors19:37
sorenRyan_Lane: I already filed an mp for a patch that adds you.19:41
sorenJust needs approval.19:41
sorenI'm working on a making the check also check pending merges, otherwise it's a complete fail.19:42
xtoddxsandywalsh: whats the objectstore question?19:43
sandywalshxtoddx, the xenapi is requesting an image from objectstore with the following url:
sandywalshxtoddx, but objectstore is 404'ing because the filename is .../image.encrypted19:44
Ryan_Lanesoren: ah. cool. thanks19:45
rluciosoren: thanks for catching the authors file update19:45
sorenrlucio: I didn't. Hudson did.19:45
sandywalshxtoddx, I don't see anything in the code seems to support encrypted images19:45
sandywalshxtoddx, (other than storing them encrypted)19:45
sorenThey're always encrypted.19:45
sorenWhen uploaded.19:45
xtoddxsandywalsh: i think they are encrypted when bundled and uploaded, but shouldn't be on disk19:45
rluciosoren: oh.  well still, thanks for adding me.  btw i used the guide you wrote on bzr, maybe the Authors thing should be added there?19:45
sorenthe objectstore decrypts them.19:46
*** BK_man has quit IRC19:46
sorenrlucio: https://code.launchpad.net/~soren/nova/no-more-networks-error/+merge/42482  shows the failure from hudson.19:46
sandywalshxtoddx, soren so, is it supposed to accept .../image as a url, but read/decrypt and transmit the encrypted version?19:47
sandywalshxtoddx, soren, because, as I say, it just 404's on http://.../_images/.../image19:48
sorensandywalsh: It's supposed to be decrypted once it's registered.19:48
xtoddxi don't see where .encrypted would be use19:48
xtoddxin nova/objectstore/handler.py19:48
sorenIt's in image.py19:49
openstackhudsonProject nova build #284: SUCCESS in 1 min 19 sec: http://hudson.openstack.org/job/nova/284/19:49
sandywalshyes, I see where it gets encrypted, the images are registered, but remaining encrypted19:50
xtoddxsandywalsh: what is the 'imageState' field in info.json for that image?19:51
*** burris has quit IRC19:51
*** Ryan_Lane is now known as Ryan_Lane|lunch19:51
sandywalsh{"imageOwnerId": "darksecret", "isPublic": false, "imageId": "ami-vwbcoull", "imageState": "decrypting", "architecture": "x86_64", "imageLocation": "mybucket/ubuntu-lucid.img.manifest.xml", "imageType": "machine"}19:51
sandywalshah, something failing on the decrypt perchance19:52
xtoddxhas the image been uploaded for a while, eg long enough for decryption to finish?19:52
sandywalshit has been now, but perhaps something failed earlier19:53
sandywalshthis is good though ... tells me where to look19:53
xtoddxcool, after that is should build image.tar.gz, then change state to untarring and then save out the "image" file19:53
xtoddxgood luck19:53
sandywalshthanks xtoddx ... that's a great help!19:53
sandywalsh(didn't know what the flow during that sequence should be)19:54
sorenIt's really not very obvious :)19:55
*** joearnold has quit IRC19:57
*** reldan has quit IRC19:57
*** reldan has joined #openstack19:57
openstackhudsonProject nova build #285: SUCCESS in 1 min 16 sec: http://hudson.openstack.org/job/nova/285/20:04
jk0soren / eday: should this pass now that Ryan_Lane|lunch added himself to Authors? https://code.launchpad.net/~armando-migliaccio/nova/xenapi-refactoring/+merge/4213320:06
*** reldan has quit IRC20:08
sandywalshxtoddx, 2010-12-07 12:06:11-0800 [HTTPChannel,4,] 14609:error:04065072:rsa routines:RSA_EAY_PRIVATE_DECRYPT:padding check failed:rsa_eay.c:592:20:12
xtoddxsandywalsh: thats new to me.  perhaps some sort of key issue?  Did you set --privatekey and --cert when bundling?20:14
sandywalshxtoddx, hmm, nope ... didn't know that was required. I did the euca-add-keypair dance20:16
xtoddxsandywalsh: i think if you source the novarc in the project zipfile it aliases the bundle command for you20:17
xtoddxi'm at a loss as to how to debug that, though20:17
sandywalshxtoddx, oh, yes it does ... cool, I can take it from there. thx again!20:18
*** brd_from_italy has quit IRC20:22
*** miclorb has quit IRC20:28
*** jaypipes has joined #openstack20:29
ttxTeam meeting in 30 minutes in #openstack-meeting20:32
edayRyan_Lane|lunch: can you remove the last commit in your opendj branch? soren already has it20:34
sorenhttps://code.launchpad.net/~soren/nova/check-authors-for-pending-merges/+merge/43006 should solve this in the future20:36
sorenIt makes it so that ryan's branch would have failed to merge rather than the subsequent one doing so.20:36
xtoddxsoren: adds a dep on bzrlib, or is that always available to anyone with bzr installed?20:39
openstackhudsonProject nova build #286: SUCCESS in 1 min 15 sec: http://hudson.openstack.org/job/nova/286/20:39
sorenxtoddx: Always available to anyone with bzr installed.20:39
sorenxtoddx: bzr is a rather thin wrapper around bzrlib.20:40
xtoddxit is possible to install from .deb and not have bzr, right?  probably could use documentation.20:40
termiesoren: if they download a tarball20:40
*** befreax has joined #openstack20:41
sorentermie: then no .bzr.20:41
sorentermie: The import is inside a check for whether .bzr exists.20:41
sorenThe code I'm replacing would just call "bzr" which is equally likely to fail.20:42
sorenUnder the exact same conditions, even.20:42
sorenxtoddx: If there's not bzr history, the check just succeeds.20:42
xtoddxlooks good then20:43
sorenOh, hang on.20:43
sorenerr.. Never mind :)20:44
xtoddxhaha, good, cause i just approved20:45
sorenWow, I suck at taking days off.20:47
chmouelsoren: is bzrlibs package being deprecated ? this is from debian experimental http://paste.openstack.org/show/250/20:47
chmoueli mean bzrtools20:48
sorenchmouel: I doubt it. It's probably just lagging a tiny bit compared to bzr.20:50
*** bfschott has quit IRC20:51
*** bfschott has joined #openstack20:52
*** masumotok_ has joined #openstack20:53
openstackhudsonProject nova build #287: SUCCESS in 1 min 12 sec: http://hudson.openstack.org/job/nova/287/20:54
ttxMeeting in one minute in #openstack-meeting !!20:59
*** befreax has quit IRC20:59
*** alekibango has quit IRC21:01
*** alekibango_ has joined #openstack21:01
*** miclorb_ has joined #openstack21:02
*** ctennis has joined #openstack21:03
*** alekibango_ is now known as alekibango21:03
spectorclan ttx: the plan is to keep all the other lists up and running for now21:04
*** joearnold has joined #openstack21:10
Ryan_Lane|luncheday: fixed21:14
*** Ryan_Lane|lunch is now known as Ryan_Lane21:14
edayRyan_Lane: cool, thanks. fyi, you can bzr uncommit and then bzr push --overwrite too21:16
Ryan_Laneah. neat. thanks21:17
sorenYeah, but you really shouldn't.21:17
*** irahgel has left #openstack21:18
sorenNo. Really. You shouldn't.21:18
Ryan_Lanemakes merges more difficult?21:18
sorenOnce stuff has been pushed, don't rewrite its history.21:18
edaysoren: well, for branches that has no deps (ie, not trunk), it's fine :)21:18
sorenIt's really not.21:18
sorenYou don't know if someone has already pulled it.21:19
*** hggdh has quit IRC21:19
Ryan_Laneit isn't much work to revert and push21:19
edayRyan_Lane: it's fine how it is now21:19
* Ryan_Lane nods21:19
edaysoren: sure, for bigger branches, but for smaller patches I don't think it's a big deal IMO. sometimes it's a lot easier to uncommit than revert all your changes in another rev21:20
sorenWhat is the use case here?21:22
soreneday: ^21:24
edayI make a change to a dev branch on top of other outstanding commits, push it, realize it shouldn't be there for one reason or another, and want to back it out before proposing it for merge21:26
edayfor larger commits, easier to uncommit than to commit again with reversing the changes21:26
edayknowing that no one probably touched it and branched from it since it's just my dev branch21:27
openstackhudsonProject nova build #288: SUCCESS in 1 min 14 sec: http://hudson.openstack.org/job/nova/288/21:29
mtaylorsoren: did you catch my note about upgrading hudson machine21:32
sorenmtaylor: Yup.21:32
sorenmtaylor: Cool beans.21:33
*** hggdh has joined #openstack21:33
mtaylorsoren: I also moved to using tarmac from ppa rather than installed from source dir21:33
*** BK_man has joined #openstack21:33
sorenmtaylor: Great. I also see updating bugs status now works.21:33
mtaylorsoren: which brings me to - is there a dist-upgrade tool I'm not aware of which knows how to change all of the ppa sources.lists that have been added by add-apt-repository?21:34
sorendo-release-upgrade is the canonical (har har) tool to do upgrades.21:34
sorenIf it doesn't do the right thing, I doubt anything will.21:34
mtaylorsweet. glad to know about it21:34
BK_manhi all. quick question. should command 'nova-manage user admin admin' create something in db? I'm using MySQL server as db instead of sqlite21:35
sorenBK_man: Yes.21:36
sorenBK_man: Well, unless you're using ldap for your user db.21:36
*** BK_man has quit IRC21:36
*** BK_man has joined #openstack21:36
BK_mansoren: khm. how can I check what I'm using - LDAP or something?21:37
sorenBK_man: If in doubt, you're not.21:38
Ryan_LaneBK_man: unless you are trying to use LDAP, you aren't :)21:38
BK_mannova-manage user admin admin completed OK without any errors but still NOTHING in DB (select * from users returns zero)21:39
sorenBK_man: If you have to ask, you're not using LDAP. It's not something you do by accident.21:39
ttxjaypipes: around ?21:39
*** crago has joined #openstack21:40
BK_mansoren: I have my DB specified as --sql_connection=mysql://nova:nova@ - is that correct?21:41
sorenI don't know. Is it? :)21:42
sorenIt's your database :)21:42
sorenNot mine.21:42
BK_mansoren: however, going to mysql (mysql -unova -pnova -h nova) and quering it (select * from users) returns empty set21:42
Ryan_LaneBK_man: did you create a nova user and give the nova user permissions on the database?21:43
BK_mansoren: and I can't issue "nova-manage user admin admin" command again21:43
chmouelis that working as nova user ?21:43
BK_manRyan_Lane: yep. I'm able to connect to that db using mysql command21:43
BK_manRyan_Lane: mysql -unova -pnova -h nova - working21:44
Ryan_Lanemaybe it's using redis?21:44
ttxjaypipes: going to bed now, see PM21:44
BK_manRyan_Lane: not think so - never installed it21:44
BK_manbtw, I'm running on bzr44021:45
Ryan_LaneBK_man: run strace21:45
sorenBK_man: Did you install from ubuntu packages or did yo udo it yourself?21:45
* Ryan_Lane thinks like a sys-admin ;)21:45
*** crago has quit IRC21:45
sorenRyan_Lane: redis?!?21:45
Ryan_Lanewas that not used for users?21:46
Ryan_Laneor was it sqllite before?21:46
sorenRyan_Lane: Redisectomy completed at bzr rev 355.21:46
sorenBK_man: My guess is that your nova-manage is using a different sql connection than you expect.21:47
BK_mansoren: what conf file nova-manage is using?21:48
BK_mansoren: I got that trace: sqlalchemy.exc.IntegrityError: (IntegrityError) column id is not unique u'INSERT INTO users (created_at, updated_at, deleted_at, deleted, id, name, access_key, secret_key, is_admin) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)' ('2010-12-07 14:44:37.968687', None, None, False, 'admin', None, 'eaa1a3ab-f05c-4296-864d-f702a2dd072a', 'c1f798cf-b2a8-471b-8216-8d3396625f15', True)21:49
BK_mansoren: however, there is no in DB. Let me check with strace21:49
sorenBK_man: That depends on my question to you earlier that you did not yet answer.21:50
soren21:45 <+soren> BK_man: Did you install from ubuntu packages or did yo udo it yourself?21:50
*** reldan has joined #openstack21:50
BK_mansoren: I'm using my own RPM packages which I build for RHEL6. They specfile is based on silas work for Fedora21:51
uvirtbotNew bug: #686772 in nova "get_info and list_instances to be made async" [Undecided,New] https://launchpad.net/bugs/68677221:51
*** iRTermite has quit IRC21:51
sorenBK_man: Then I don't know.21:52
*** dirakx has quit IRC21:53
BK_mansoren: all my cnf files in /etc/nova/ contains sql_connection strings referring to mysql and not to sqlite21:54
BK_mansoren: however, nova-manage still uses SQLite instead of MySQL :-(21:54
sorenBK_man: Does nova-manage use any of those files?21:55
sorenBK_man: I know what I did in the Ubuntu packages to make it work. I have no clue what someone called silas did for Fedora and you then built on top off.21:56
BK_mansoren: how could I check? I'm not a Python guru :(21:56
*** jdarcy has quit IRC21:56
BK_mansoren: do you have a kind of RPM specfile for Ubuntu available? I need to take a look on it to get an idea21:56
sorenBK_man: Check with strace.21:56
BK_mansoren: my stuff is here: https://github.com/abrindeyev/openstack-nova-rhel621:57
sorenBK_man: https://code.launchpad.net/~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu21:57
jaypipesgah, I keep missing the meeting... my calendar has had it at starting now...21:59
dabojaypipes: DST FTW!21:59
BK_mansoren: I see only single attemp to load nova.conf from /usr/bin instead from /etc/nova22:00
jaypipesdabo: yeah, no shit, eh? :)22:00
sorenBK_man: That's your problem then.22:00
BK_mansoren: got it working. what should I patch in nova code to enable it with right path?22:01
sorenBK_man: http://bazaar.launchpad.net/~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu/annotate/head:/debian/patches/nova-manage_flagfile_location.patch22:01
sorenBK_man: That's one way. I'm going to change that soo, though, but it works.22:01
*** iammartian_ has joined #openstack22:02
alekibangodoes it use /etc/nova ? :)22:06
*** BK_man has quit IRC22:06
*** BK_man_ has joined #openstack22:07
vishysoren: the code you added for a single path to CA etc. does it use the cwd if not specified?22:11
sorenvishy: It uses... err..22:12
*** jfluhmann has quit IRC22:13
sorenvishy: os.path.join(os.path.dirname(__file__), '../')22:13
sorenvishy: So the directory above the directory holding utils.py.22:13
sorenflags.py, I mean.22:14
sorenSame directory, though.22:14
vishysoren: hmm i can't figure out why dirakx is still having trouble then22:14
sorenvishy: His backtrace suggests it's looking in the right place.22:14
sorenvishy: ...but for some reason, it's not being installed.22:14
*** jfluhmann has joined #openstack22:15
vishysoren: wait he is still running from the egg22:15
sorenvishy: No.22:16
sorenIt's not an egg per se.22:16
vishywait, he didn't post traceback for the last thing he tried22:16
sorenIt's an unpacked egg.22:16
sorenvishy: We're talking about https://answers.launchpad.net/nova/+question/136966 right?22:16
sorenOk. So his novarc problem seems to have been resolved.22:17
BK_man_gents, what we will do with new openssl? At the moment we have 'California' != 'California' bug in openssl >0.9.8. At the moment I only relaxed match policy in openssl.cnf and bypassed that check.22:17
sorenvishy: The next problem is probably due to not having run nova-api yet.22:17
sorenvishy: ...so genrootca.sh hasn't been run (which is what creates openssl.cnf from openssl.cnf.template22:17
vishyah good point22:17
sorenI'm about to head to bed, though, so if you could tell him that, that'd be great.22:18
alekibangosleep well, soren22:18
sorenalekibango: ta22:18
* soren wanders off22:18
*** GasbaKid has quit IRC22:20
*** jfluhmann has quit IRC22:28
*** BK_man has joined #openstack22:37
*** BK_man_ has quit IRC22:37
*** allsystemsarego has quit IRC22:39
Ryan_Lanevishy: I'm confused about what the get_user_roles function is doing in the ldap driver...22:44
vishygetting the user roles...22:45
*** jfluhmann has joined #openstack22:45
Ryan_Laneright. but if project_id is passed in, it searches for a project instead?22:45
Ryan_Lanebut it is searching for a project dn and explicitly saying "(!(objectclass=novaProject))" in the search22:46
Ryan_Lanefind a project that isn't a project?22:46
Ryan_Laneshould that be role_id?22:46
Ryan_Laneis it supposed to be searching for a specific role?22:47
*** burris has joined #openstack22:49
Ryan_Laneis that function, when called that way, really saying: "is this user in this role?"22:50
Ryan_Lanevishy: ^^22:50
vishyit is finding the local roles22:51
vishythere are global (user) roles22:51
vishyand project roles22:51
vishyproject roles are groupOfNames under the novaProject dn22:51
*** reldan has quit IRC22:52
vishyhence my comment about having to change the logic of how they are searched to not include the project if you change novaProject to groupOfNames22:52
Ryan_Lanemy terminology is slightly off then :)22:53
*** jbaker_ has quit IRC22:53
Ryan_LaneI'm trying to separate out the roles into two separate OUs22:53
Ryan_Lanethere's the ldap_project_subtree flag, and the role_project_subtree flag22:53
Ryan_Lanethose are the two you are referring to, correct?22:54
Ryan_Lane(i'm working on changing the logic right now, which is why I ask :) )22:56
*** mdomsch has quit IRC22:57
*** gondoi has quit IRC23:01
*** ctennis has quit IRC23:07
*** BK_man_ has joined #openstack23:07
*** BK_man has quit IRC23:07
*** BK_man_ is now known as BK_man23:07
Ryan_Lanevishy: ldap_project_subtree = project roles, role_project_subtree = global user roles?23:07
vishyldap_project_subtree is the project ou23:08
vishythe project roles go under the individual project that they are associated with23:08
*** ctennis has joined #openstack23:10
*** ctennis has joined #openstack23:10
*** ppetraki has quit IRC23:11
Ryan_Lanevishy: are those not the two things that would need to be separated? currently ldap_project_subtree and role_project_subtree share the same OU. even if those two are separated, is there still something else to be separated?23:13
vishyyou have to change the logic for get project roles23:14
vishyis not going to work23:14
vishyif project is also a groupOfNames23:14
Ryan_LaneI'm having a conceptual misunderstanding of how things work. I can work out the logic if I understand what each object does.23:15
vishya project is basically a group of names with a projectManager23:16
vishybut under the project dn23:16
vishyare one or more role group of names23:16
* Ryan_Lane nods23:16
vishyso those are the users that have the 'developer' role for myproject23:17
vishyright now it searches cn=myproject;ou=projects for all group of names to find all the roles for the project23:18
Ryan_Laneyou are placing objects underneath a cn?23:18
* Ryan_Lane twitches23:18
vishyis that bad?23:18
Ryan_Lanegenerally, yes ;)23:18
Ryan_Lanenot saying it isn't done23:18
vishyhow do you define a subgroup without doing that?23:19
Ryan_Lanegroup inheritance23:19
Ryan_Lanerecursive group membership23:19
Ryan_Lanegimme a sec23:19
vishydoes it make sense to do roles that way.  I just did the easiest that made sense to me, but if there is a more ldapy way to do it, cool23:20
Ryan_Lanecn=myproject would have: member: cn=myproject-developers,ou=projects23:20
Ryan_Laneor some such23:20
Ryan_Lanethen when you search for the groups, you do a recursive search23:20
vishyi see23:20
vishybut what if you have multiple roles23:20
vishywon't you end up with the same user in the project multiple times?23:21
Ryan_Lanethe code should eliminate any duplications23:21
vishyk that works23:22
Ryan_Lanethat explains a lot of my confusion in the code :)23:22
vishyas long as you can somehow locate project roles using project_id and role_name23:22
vishyit was just far to genius for you :)23:22
vishy* too23:23
*** aliguori has quit IRC23:23
Ryan_Laneit's an interesting way to do it23:23
Ryan_Lanejust one that is out of the ordinary enough that I didn't expect it ;)23:23
xtoddxdon't python list comprehensions automatically filter out None?23:26
*** jimbaker has joined #openstack23:27
creihtxtoddx: don't think so23:28
xtoddxyea, i just convinced myself it doesn't.  nor False23:28
xtoddxi don't know where i got that impression23:28
*** aliguori has joined #openstack23:36
*** BK_man_ has joined #openstack23:37
*** BK_man has quit IRC23:37
*** BK_man_ is now known as BK_man23:37
xtoddxRyan_Lane: I'm looking at https://code.launchpad.net/~rlane/nova/ldap-user-modify-only/+merge/4197123:39
Ryan_Lanextoddx: ok. any issues?23:39
xtoddxOn line 60, isn't that check already done at the top of the function with __user_exists?23:39
Ryan_Lanelemme see. gimme a sec23:40
xtoddxthat would be line 60 of the diff, not of the file23:40
Ryan_Laneyeah :)23:40
Ryan_Lanextoddx: not necessarily23:43
Ryan_Lanewait. misread23:44
Ryan_Laneyes. being done23:44
Ryan_Lanelemme fix that and push it23:44
Ryan_LaneI'm noticing another problem now though :)23:48
Ryan_Lanenothing happens and nothing is returned if the user doesn't exist23:48
Ryan_Lanean exception should be thrown there23:48
xtoddxRyan_Lane: __to_user has always just returned none23:50
xtoddxi think we should keep that behavior23:51
Ryan_Laneok, so return none on failure?23:51
xtoddxunless i'm not understanding where you're wanting to raise an error23:51
Ryan_Laneif the user account doesn't exist, it can't "create" the user23:51
xtoddxah, in modify mode, i see.23:51
Ryan_Laneso it fails to add the user23:52
xtoddxi think a raise might be appropriate in that case23:52
Ryan_Laneany idea which one I should use?23:52
Ryan_Laneok. cool. that's what I put there :)23:52
Ryan_Laneok. pushed.23:53
Ryan_Lanextoddx: think that error message may be too confusing?23:54
xtoddxstill waiting on launchpad to refresh23:54
xtoddxi'll pull locally23:54
Ryan_Lanehmm. that error message may be confusing as hell23:55
*** jordandev has joined #openstack23:56
Ryan_LaneMaybe "User's LDAP object %s doesn't exist" would be better23:57
Ryan_LaneMaybe "LDAP object for %s doesn't exist" would be better23:58

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