Tuesday, 2011-03-15

*** dfg has quit IRC00:05
*** adjohn has joined #openstack00:07
*** Ryan_Lane has joined #openstack00:07
*** azneita has joined #openstack00:11
*** dendro-afk is now known as dendrobates00:15
*** Ryan_Lane has quit IRC00:16
*** cascone has joined #openstack00:19
*** dragondm has quit IRC00:27
*** reldan has quit IRC00:32
*** jc_smith has left #openstack00:33
openstackjenkinsProject nova build #640: SUCCESS in 2 min 0 sec: http://hudson.openstack.org/job/nova/640/00:33
openstackjenkinsTarmac: Implements basic OpenStack API client, ready to support API tests00:33
*** joearnold has quit IRC00:36
*** romain_lenglet has joined #openstack00:38
*** adiantum has joined #openstack00:51
*** bcwaldon has quit IRC00:54
*** dendrobates is now known as dendro-afk00:58
*** adiantum has quit IRC00:58
*** adiantum has joined #openstack01:04
*** ewindisch has joined #openstack01:11
*** gregp76 has quit IRC01:14
*** aliguori has quit IRC01:17
*** aliguori has joined #openstack01:32
*** MotoMilind has quit IRC01:40
*** jtimberman has quit IRC01:42
winston-dnotmyname: hi, r u around?01:43
jarrodman i love freeswitch01:50
jarrodits my favorite opensource project in existence01:50
HugoKuodoes anyone's archi is [1 machine : API +mysql+ objetstore]   [1machine network + scheduler] [1machine compute node]?    I got a problem on nova-networl, could someone plz  show me your "nova.conf" and how many NIC using on nova-networl component.01:51
*** adiantum has quit IRC01:56
*** adiantum has joined #openstack02:01
*** jakedahn has joined #openstack02:04
*** aliguori has quit IRC02:05
*** hazmat has quit IRC02:08
*** dendro-afk is now known as dendrobates02:23
*** ewindisch has quit IRC02:25
*** nelson has quit IRC02:36
*** kie has joined #openstack02:44
*** ccustine has quit IRC02:47
HugoKuoStcuk on this problem for several days.     could someone give me some suggestion?      https://answers.launchpad.net/nova/+question/149115         thanks02:50
*** paltman has joined #openstack02:56
*** bcwaldon has joined #openstack03:02
*** jakedahn_ has joined #openstack03:07
*** jakedahn has quit IRC03:09
*** jakedahn_ is now known as jakedahn03:09
*** ewindisch has joined #openstack03:12
*** paltman has quit IRC03:18
*** paltman has joined #openstack03:18
*** hazmat has joined #openstack03:24
*** adiantum has quit IRC03:35
*** cascone has quit IRC03:37
*** paltman has quit IRC03:40
*** dendrobates is now known as dendro-afk03:41
*** adiantum has joined #openstack03:41
*** thielmann__ has joined #openstack03:47
*** adiantum has quit IRC03:49
*** cascone has joined #openstack03:50
*** thielmann_ has quit IRC03:51
*** adiantum has joined #openstack03:55
*** iammartian has quit IRC03:59
*** paltman has joined #openstack04:03
*** romain_lenglet has quit IRC04:13
*** sandywalsh has quit IRC04:16
*** cascone has quit IRC04:30
*** westmaas1 has quit IRC04:49
*** cascone has joined #openstack04:50
*** maplebed has quit IRC04:53
annegentlecreiht: actually, Ryan_Lane is right on about the logo and creative commons... but he can certainly talk to spectorclan about branding and uses of the logo.04:54
*** adiantum has quit IRC04:54
*** adiantum has joined #openstack04:55
*** bcwaldon has quit IRC05:01
*** kashyap has joined #openstack05:02
*** Ryan_Lane has joined #openstack05:06
*** paltman has quit IRC05:07
*** gregp76 has joined #openstack05:09
*** omidhdl has joined #openstack05:14
*** paltman has joined #openstack05:24
*** adiantum has quit IRC05:30
*** adiantum has joined #openstack05:36
*** Ryan_Lane has quit IRC05:40
*** fysa has quit IRC05:42
*** kie has quit IRC06:01
mtayloreday: awake?06:03
edaymtaylor: yo06:05
mtayloreday: hey - you remember that place you took me for burritos in portland?06:06
mtayloreday: the one with the vegan burritos I can't shut up about?06:06
mtayloreday: if I wanted to tell someone to go there, what would I tell them?06:06
edaymtaylor: yup, http://blpdx.com/06:06
mtayloreday: thanks06:07
edaymtaylor: np :)06:08
mtayloreday: friend of a friend just moved to portland, which came up after I mentioned that burrito (or wrap, as it might have been) in conversation06:10
edaymtaylor: ahh, cool06:10
edaymtaylor: it moved from the small cafe place we went, it's not on the east side06:11
edayerr, now06:11
*** f4m8_ is now known as f4m806:11
edaybigger, dedicated space (not being shared with yoga studio like before)06:11
mtaylorgood for them06:11
*** zenmatt has quit IRC06:30
*** gregp76 has quit IRC06:48
*** daveiw1 has quit IRC07:03
*** berendt has joined #openstack07:08
uvirtbotNew bug: #735295 in nova "class SchedulerManager not found while trying to start nova-scheduler" [Undecided,New] https://launchpad.net/bugs/73529507:12
uvirtbotNew bug: #735298 in nova "can't start nova-compute: Invalid: Invalid xml. '<cpu>' must be 1,but 5" [Undecided,New] https://launchpad.net/bugs/73529807:12
*** zykes- has quit IRC07:18
*** ivan has quit IRC07:18
*** ryker has quit IRC07:18
*** Failbaitr has quit IRC07:18
*** freeflyi1g has quit IRC07:19
*** Failbaitr has joined #openstack07:19
*** freeflying has joined #openstack07:25
*** mgoldmann has joined #openstack07:27
*** ivan has joined #openstack07:27
*** miclorb has quit IRC07:30
*** ryker has joined #openstack07:30
*** drico_ has joined #openstack07:34
*** soren has quit IRC07:35
*** _0x44 has quit IRC07:35
*** karmabot` has quit IRC07:36
*** laurensell has quit IRC07:36
*** drico has quit IRC07:36
*** huleboer has joined #openstack07:38
*** soren has joined #openstack07:38
*** _0x44 has joined #openstack07:38
*** ChanServ sets mode: +v soren07:39
*** ChanServ sets mode: +v _0x4407:39
*** karmabot has joined #openstack07:42
*** ramkrsna has joined #openstack07:42
*** omidhdl1 has joined #openstack07:46
*** omidhdl has quit IRC07:46
*** ramkrsna has quit IRC07:47
*** lionel_ has quit IRC07:48
*** miclorb has joined #openstack07:49
*** lionel_ has joined #openstack07:52
*** laurensell has joined #openstack07:55
*** ramkrsna has joined #openstack08:01
*** MarcMorata has joined #openstack08:07
*** rcc has joined #openstack08:09
HugoKuois there anyone trying to use same archi as mine ?     could you plz show me your nova.conf https://answers.launchpad.net/nova/+question/14911508:13
*** adiantum has quit IRC08:19
*** guigui has joined #openstack08:19
*** mgoldmann has quit IRC08:19
*** skiold has joined #openstack08:23
*** allsystemsarego has joined #openstack08:29
*** allsystemsarego has joined #openstack08:29
*** adiantum has joined #openstack08:31
alekibango...sheepdog server is down (located in japan).  now after the earthquake and tsunami, when 3 reactors exploded, i am afraid that life in japan is hard. Nuclear fallout will probably hit west coast of usa too..08:32
*** DigitalFlux has joined #openstack08:32
DigitalFluxHi Guys08:32
DigitalFluxI was wondering if there is a ready test environment for someone working on a feature that uses the combination of kvm/libvirt/Glance/OpenStack API08:33
DigitalFluxI can see there is a http://wiki.openstack.org/XenServerDevelopment08:34
DigitalFluxand that Novascript for objectstore + QEMU/Libvirt08:34
DigitalFluxAnyone working on a kvm/libvirt/Glance/OpenStack API env ?08:34
alekibangoDigitalFlux: there will be, in the future...  but for now afaik its not ready08:34
alekibangoi am using fai to automatically install all stuff in a cluster, in few minutes08:35
alekibangoi am willing to setup testing environment, if someone provides 5-10 servers08:35
alekibango... which could be used to do automated functional testing of every merge request on multiple configurations (and operating systems)08:36
alekibangoas all can be reinstalled in few minutes08:36
alekibangoi can help with setting up fai server -- someone needs to provide machines, configs for hosts and tests08:37
alekibangoDigitalFlux: for basic testing, nova.sh script can be used08:38
*** daveiw has joined #openstack08:38
*** azneita has quit IRC08:44
*** uksysadmin has joined #openstack08:44
berendtalekibango: could you provide your FAI configuration in the wiki?08:45
alekibangoberendt: i plan to release it somehow when it will go gold (working well)08:47
alekibangofor now i still have to fight with some parts of the system08:47
alekibangofor example, i used debian squeeze and i had to package lots of stuff myself...08:48
alekibangonow trying to use ubuntu on servers, getting different bugs08:48
alekibangoberendt: i will prolly put the config into some public git repository08:48
berendtalekibango: why not yet? think it's not problem if it's not fully working, you can maintain the configuration in the git repository. think it would be a good starting point for someone interessted in using FAI in combination with openstack08:50
alekibangohmm you are right. release early and often :)08:52
alekibangoberendt: are u interested?08:53
alekibangoi will put that  higher in my todo list...08:54
berendtalekibango: we'll try an other solution in the next weeks. but i want to take a look on FAI, too.09:01
alekibangowhat will you try?09:08
nerensHi all, I need some help understanding how openstack fits together pls. I've got 3 machines, pollux with swift for storage, ares as the cloud controller with postgresdb and hera as compute node. So nova-network only runs on one instance so run that on ares right? where do i run nova-api and scheduler?09:13
*** berendt has quit IRC09:15
*** arefaey has joined #openstack09:17
*** irahgel has joined #openstack09:19
*** befreax has joined #openstack09:26
*** miclorb has quit IRC09:31
*** mgoldmann has joined #openstack09:37
nerensi guess i'm not the only one that's confused then?09:42
*** HugoKuo_ has joined #openstack09:49
alekibangonerens: you can run as many services on as many hosts as you wish... :) 3 machines are not much.  rather use mysql. i have seen there were some issues with postgress09:50
*** HugoKuo__ has joined #openstack09:50
alekibangonerens: its confusing... you are not the only one, for sure. but now most people sleep09:50
alekibangoand i need to get some sleep too09:50
alekibangoin 4-5 hours there will be more people09:51
alekibangonerens: you can run compute  and swift and everything on all your nodes09:51
alekibangorunning compute on 1 node is not that great :)09:52
alekibangothe same for swift, at least 3-4 machines for swift only is good09:52
nerensalekibango: thx for the help, so nova-api I can run on all hosts, is nova.conf the same on all hosts too?09:52
alekibangomight be... depends on config09:52
alekibangonerens: it is very open and adaptable technology09:53
alekibangothat gives also problems...09:53
alekibangoi mean, you need to know what you want09:53
alekibangowhich you prolly do not know for now...09:53
nerenswhat problems have you seen running postgres?09:53
*** HugoKuo has quit IRC09:53
alekibangoi didnt, but i have seen something here or on maillinglist09:53
alekibangomysql is tested first...09:54
alekibangoby most devels09:54
alekibangomaybe drizzle will be good09:54
alekibango(in future)09:54
alekibangothey should go stable soon09:54
*** HugoKuo_ has quit IRC09:54
nerensso if i want to run multiple compute nodes, will I still run all nova services on all nodes?09:55
alekibangonova-api -- you really need only one09:55
alekibangobut if that host fails, another one might be good to have09:55
alekibangoyes you run nova-compute and also other services on other nodes09:55
nerensok, so nova-api and nova-network can run on one host right?09:55
alekibangocan be also on all of them09:56
nerensok, and the scheduler?09:56
alekibangoi run that on all nova hosts :)09:56
nerensalekibango: thanks for your help.. i'll let you get some rest, will bother the guys a little later when I get stuck :)09:58
alekibangook, good luck09:58
alekibangoyou can also ask on launchpad questions page09:59
alekibangoor using maillist09:59
*** adjohn has quit IRC09:59
nerensthanks, will do10:00
*** openstackjenkins has quit IRC10:01
*** openstackjenkins has joined #openstack10:01
*** alekibango has quit IRC10:04
HugoKuo__Does any one success to isolate nova-network to a single machine ?  I10:05
*** miclorb has joined #openstack10:08
*** adiantum has quit IRC10:08
*** adiantum has joined #openstack10:11
*** adjohn has joined #openstack10:23
*** daveiw has quit IRC10:26
*** daveiw has joined #openstack10:27
*** adjohn has quit IRC10:27
*** skiold has quit IRC10:31
*** koolhead11 has joined #openstack10:32
koolhead11hi all10:32
sorenHugoKuo__: What does that mean?10:35
HugoKuo__soren :10:35
HugoKuo__I try to isolate nova-network and nova-scheduler into on machine10:36
HugoKuo__but I face some network problem10:36
sorenWhat does it mean to isolate it?10:37
sorenIt's not like they usually accidentally start running on other machines.10:37
HugoKuo__I success to deploy nova on one machine ....everything is running fine10:39
HugoKuo__so I try to make it better performance10:39
soren"-A PREROUTING -d -p tcp -m tcp --dport 80 -j DNAT --to-destination"10:39
sorenI think this line is wrong , nova-api is on
sorenThat was meant to look like a quite.10:40
HugoKuo__nova-api is on
sorenquote, even.10:40
*** reldan has joined #openstack10:40
HugoKuo__I know why that linux_net.py  assign to iptables10:41
sorenHugoKuo__: ec2_dmz_host is the flag that is used to populate that IP, apparently.10:41
HugoKuo__default is $my_ip   right ?10:41
HugoKuo__after that I write a fixed string to replace metadata iptable rule to
headoffsoren: You mentioned to me yesterday that there is no possibility to manually set an IP when alunching instances , If i log in to a machine and change /etc/network/interfaces it would be changed on teh instance, right? Do you have any idea of how to make the instance being able to get the metadata server info, because the instance is getting stuck when trying to access the EC2metadata server10:42
HugoKuo__ok I can curl on nova-compute node...10:43
HugoKuo__but while instance try to get metadata from nova-api , it shows url error....10:43
zulsoren: http://bazaar.launchpad.net/~zulcss/nova/nova-lxc/revision/76110:44
sorenheadoff: No, that will not work. Of course not.10:44
sorenheadoff: If it did, people could just steal each other's ip's.10:44
HugoKuo__mr. soren , I'll try to use ec2_dmz flag ....... thanks10:45
sorenzul: Uh..10:45
*** hadrian has quit IRC10:45
zulsoren: i also noticed that if you arent running qemu/xen/qemu-kvm you will get list index out of range with the latest trunk as well10:46
headoffsoren: yeah thats true but for testing purpose, i was wondering what's happening behind the scenes why it cant be accessed. thanks anyway10:46
*** adiantum has quit IRC10:50
sorenheadoff: If you attempt to change your mac or your IP, your traffic will be filtered.10:52
sorenheadoff: It'll never leave your VM.10:52
sorenheadoff: I'm still puzzled why you're adding the complexity of Nova. You seem to be specifically looking for standard virtualisation.10:53
*** koolhead11 is now known as koolhead1710:58
sorenzul: What do you mean?11:00
zulsoren: see https://bugs.launchpad.net/nova/+bug/73539011:01
uvirtbotLaunchpad bug 735390 in nova "Live migration breaks UML (and will break LXC)." [High,Confirmed]11:01
*** diegoparrilla has joined #openstack11:06
uvirtbotNew bug: #735390 in nova "Live migration breaks UML (and will break LXC)." [High,Confirmed] https://launchpad.net/bugs/73539011:06
diegoparrillaAnybody in the house running a Nova multinode configuration with FlatDHCPManager ?11:07
*** kashyap has quit IRC11:17
arefaeyI'm running nova on single node using nova.sh, But I get an exception in the Network and volume screen11:18
arefaey(nova): TRACE:     raise TypeError('f(ile) should be int, str, unicode or file, not %r' % f)11:18
arefaey(nova): TRACE: TypeError: f(ile) should be int, str, unicode or file, not <open GreenPipe '<fd:4>', mode 'wb' at 0x2067f80>11:18
HugoKuo__hello diegoparrilla11:20
HugoKuo__let's jumo to pm11:20
*** miclorb has quit IRC11:30
*** matclayton has joined #openstack11:37
sorenarefaey: Your eventlet is out-dated.11:40
sorenarefaey: https://answers.launchpad.net/nova/+faq/148511:40
zulsoren: i might not be awake yet but im not sure where you are getting kpartx from the mounting of the container doesnt use kpartx anymore it just uses the loopback device to mount the image11:41
sorenzul: Oh.11:42
sorenzul: /me re-reviews11:42
*** omidhdl1 has left #openstack11:47
*** pyros_xx has joined #openstack11:49
*** hggdh has quit IRC11:51
*** cascone has quit IRC11:51
*** shrewduh has joined #openstack11:52
zulsoren: one last thing the losetup --find --show just tries to find the loopback device so it can clean it up by removing it as well, ill add more testcases as well12:00
*** hggdh has joined #openstack12:01
*** dprince has joined #openstack12:02
*** hazmat has quit IRC12:05
*** hazmat has joined #openstack12:05
sorenzul: Can't you comment on the merge proposal instead? It's very hard to follow the thread when half the conversation is on IRC, the other on launchpad.12:06
zulsoren: sure12:06
*** shrewduh has quit IRC12:07
*** shrewduh has joined #openstack12:08
arefaeyroot@arefaey-desktop:/usr/local/lib/python2.6/dist-packages/eventlet# curl https://bitbucket-assetroot.s3.amazonaws.com/which_linden/eventlet/20110214/77/greenpipe-wrap.patch | patch12:11
arefaey  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current12:11
arefaey                                 Dload  Upload   Total   Spent    Left  Speed12:11
arefaey100  1212  100  1212    0     0   1000      0  0:00:01  0:00:01 --:--:--  168512:11
arefaeypatching file subprocess.py12:11
arefaeyHunk #1 FAILED at 27.12:11
arefaey1 out of 1 hunk FAILED -- saving rejects to file subprocess.py.rej12:11
arefaeysoren : I tried to patch the eventlet but it failed12:11
*** h0cin has joined #openstack12:14
sorenarefaey: Maybe apply it manually.12:14
sorenarefaey: Which version of eventlet are you running?12:14
arefaeyspren: the Network exception now :12:17
arefaeysoren: (nova): TRACE: Traceback (most recent call last):12:17
arefaey(nova): TRACE:   File "/home/arefaey/openstack/nova/nova/utils.py", line 437, in _inner12:17
arefaey(nova): TRACE:     self.f(*self.args, **self.kw)12:17
arefaey(nova): TRACE:   File "/home/arefaey/openstack/nova/nova/service.py", line 196, in periodic_tasks12:17
arefaey(nova): TRACE:     self.manager.periodic_tasks(context.get_admin_context())12:17
arefaey(nova): TRACE:   File "/home/arefaey/openstack/nova/nova/network/manager.py", line 151, in periodic_tasks12:17
arefaey(nova): TRACE:     time)12:17
arefaey(nova): TRACE:   File "/home/arefaey/openstack/nova/nova/db/api.py", line 348, in fixed_ip_disassociate_all_by_timeout12:17
arefaey(nova): TRACE:     return IMPL.fixed_ip_disassociate_all_by_timeout(context, host, time)12:17
arefaey(nova): TRACE:   File "/home/arefaey/openstack/nova/nova/db/sqlalchemy/api.py", line 98, in wrapper12:17
arefaey(nova): TRACE:     return f(*args, **kwargs)12:17
arefaey(nova): TRACE:   File "/home/arefaey/openstack/nova/nova/db/sqlalchemy/api.py", line 660, in fixed_ip_disassociate_all_by_timeout12:17
arefaey(nova): TRACE:     'leased': 0})12:17
arefaey(nova): TRACE:   File "/usr/lib/python2.6/dist-packages/sqlalchemy/orm/query.py", line 2145, in update12:18
arefaey(nova): TRACE:     "Could not evaluate current criteria in Python. "12:18
arefaey(nova): TRACE: InvalidRequestError: Could not evaluate current criteria in Python. Specify 'fetch' or False for the synchronize_session parameter.12:18
sorenPlease use a pastebin instead of pasting here.12:18
arefaeyOK, sorry12:19
sorenI have no idea what that error is.12:19
arefaeyHow can I query the current eventlet version, I used easy_install eventlet12:19
sorenNo idea.12:20
sorenttx: See all that green? https://jenkins.linux2go.dk/job/Nova-user-test/12:23
sorenttx: That's from trunk.12:23
*** shrewduh has quit IRC12:30
*** zenmatt has joined #openstack12:30
*** adiantum has joined #openstack12:30
ttxsoren: cool12:30
*** zul_ has joined #openstack12:32
*** dendro-afk is now known as dendrobates12:32
*** zul has quit IRC12:33
*** zul_ is now known as zul12:35
*** zul has joined #openstack12:36
*** dendrobates is now known as dendro-afk12:40
ttxsoren: ah! some red !12:41
*** dendro-afk is now known as dendrobates12:42
sorenttx: It's probably not an actual failure.12:43
sorenttx: /me looks12:43
*** patcoll has joined #openstack12:43
sorenttx: Yeah, it's not. It's because my test sometimes fail due to certain circumstances outside of Nova's control.12:44
sorenttx: I should really fix that.12:47
sorenttx: Muahahah. Ok, so now I'm monitoring the PPA. When stuff there changes, it pulls the new packages and starts testing them.12:48
DigitalFluxWhat is the state of this branch https://code.launchpad.net/~justin-fathomdb/nova/virtualbox-support12:50
DigitalFluxDoes Bexar support Vbox as a hypervisor ?12:50
sorenDigitalFlux: No.12:54
DigitalFluxsoren: so this was not merged ..12:54
uksysadminHi DigitalFlux, I emailed Justin on this and the response was that although it seemed interesting, there wasn't that much need for it... certainly not in the commercial space.12:55
uksysadminso one of the reasons it never got merged (yet...?)12:55
uksysadminoutside of standalone testing, would you use vbox in production?12:56
DigitalFluxuksysadmin: Thanks, actually some of Vbox features can be considered needed in a commercial product :)12:56
DigitalFluxLike their RDP server12:56
*** diegoparrilla has quit IRC12:56
DigitalFluxTheir snapshots feature are pretty cool tool12:56
uksysadminI don't disagree, but I'd never be the one to suggest using vbox in production where I work... not unless I liked to join the dole queue12:56
*** arefaey has quit IRC12:57
sorenttx: Ok, I think I've addressed that now.12:57
DigitalFluxcomparing the speed of VBox snapshots vs kvm/libvirt snapshot, VBox surprisingly wins !12:59
*** skiold has joined #openstack12:59
uksysadminguess it comes down to what you want out of virtualisation... I've been running virtualisation for years in prod and never needed to do any snapshotting so that feature isn't a requirement for the way the environments I've work on are set up13:00
uksysadminmust be good for windows updates and stuff where you might want to roll back environments, etc.?13:00
sorenttx: I thought wrong.13:01
*** reldan has quit IRC13:01
*** aliguori has joined #openstack13:06
*** ppetraki has joined #openstack13:06
*** shentonfreude has joined #openstack13:10
*** ramkrsna has quit IRC13:14
*** ramkrsna has joined #openstack13:14
*** sandywalsh has joined #openstack13:15
*** adiantum has quit IRC13:20
*** adiantum has joined #openstack13:21
*** guigui has quit IRC13:23
*** kashyap has joined #openstack13:23
ttxsoren: saw bug 735390 ?13:33
uvirtbotLaunchpad bug 735390 in nova "Live migration breaks UML (and will break LXC)." [High,Confirmed] https://launchpad.net/bugs/73539013:33
ttxLooks like Live migration makes a few libvirt = qemu assumptions.13:33
sorenttx: Yeah.13:33
sorenttx: I'll look into it soon. Not right now.13:34
*** shentonfreude has left #openstack13:39
*** bkkrw has joined #openstack13:40
*** sandywalsh has quit IRC13:41
*** spectorclan_ has joined #openstack13:44
*** bkkrw has quit IRC13:46
*** axelis has joined #openstack13:48
*** guigui3 has joined #openstack13:48
*** adiantum has quit IRC13:57
*** jfluhmann_ has quit IRC13:57
*** adiantum has joined #openstack13:58
*** koolhead17 is now known as koolhead17|afk13:59
*** kyzh has joined #openstack14:08
*** rpedde has joined #openstack14:08
*** sandywalsh has joined #openstack14:13
sorenttx: \o/ I seem to have beaten the false negatives. https://jenkins.linux2go.dk/job/Nova-user-test/14:13
sorenttx: it's odd, really. I would ssh into an instance, wget a url, echo something, exit 0. Every once in a while the stuff I echoed just never came across. I think it the connection got severed without the output buffer got flushed.14:14
*** reldan has joined #openstack14:16
*** bcwaldon has joined #openstack14:17
ttxsoren: cool!14:21
*** kyzh has quit IRC14:21
*** kyzh has joined #openstack14:21
*** kyzh is now known as Guest2220414:22
*** ewindisch has quit IRC14:23
*** hadrian has joined #openstack14:26
*** msassak has joined #openstack14:30
*** scotticus has joined #openstack14:31
*** thatsdone has joined #openstack14:32
*** ramkrsna has quit IRC14:33
*** ramkrsna has joined #openstack14:34
*** kashyap has quit IRC14:39
*** jtimberman has joined #openstack14:41
*** dmshelton has quit IRC14:43
*** matclayton has joined #openstack14:45
*** Ryan_Lane has joined #openstack14:45
*** pyros_xx has quit IRC14:46
*** Guest22204 has quit IRC14:47
*** westmaas has quit IRC14:47
*** tr3buchet has quit IRC14:48
*** westmaas has joined #openstack14:48
*** Guest22204 has joined #openstack14:50
*** f4m8 is now known as f4m8_14:50
*** tr3buchet has joined #openstack14:51
*** MarkAtwood has joined #openstack14:52
*** rnirmal has joined #openstack14:53
*** pvo has joined #openstack14:53
*** enigma1 has joined #openstack14:59
*** pvo has quit IRC14:59
*** pvo has joined #openstack15:01
*** dragondm has joined #openstack15:02
*** dmshelton has joined #openstack15:02
*** shrewduh has joined #openstack15:05
*** shrewduh has quit IRC15:06
*** rnirmal_ has joined #openstack15:09
*** Failbaitr has quit IRC15:10
*** rnirmal has quit IRC15:11
*** rnirmal_ is now known as rnirmal15:11
*** Ryan_Lane has quit IRC15:22
*** mrsrikanth has joined #openstack15:23
*** BK_man has joined #openstack15:24
*** reldan has quit IRC15:25
*** lorinh1 has joined #openstack15:26
lorinh1Anybody know where I can find a repo with the python-multiprocessing package for lucid?15:29
creihtlorinh1: as of python 2.6, multiprocessing is built in15:29
lorinh1Hmmm, then python-multiprocessing shouldn't be a dependency on lucid, should it?15:30
*** jfluhmann has joined #openstack15:30
lorinh1creiht: I'm getting package issues because of a python-multiprocessing dependency, but it sounds like it doesn't need to be there: https://bugs.launchpad.net/nova/+bug/73552615:31
uvirtbotLaunchpad bug 735526 in nova "Can't build packages on lucid box because of missing python-multiprocessing" [Undecided,New]15:31
*** Ryan_Lane has joined #openstack15:32
uvirtbotNew bug: #735526 in nova "Can't build packages on lucid box because of missing python-multiprocessing" [Undecided,New] https://launchpad.net/bugs/73552615:32
creihtlorinh1: ahh interesting, I will have to defer to someone else to answer that question then15:33
lorinh1creiht: Is Chuck Short in the room? He made the change: http://bazaar.launchpad.net/~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu/revision/10815:34
creihtI'm not sure who he is15:34
creihtMy guess is zul?15:34
zulwhats wrong?15:35
dendrobateszul: what did you mess up this time15:35
zuldendrobates: everything!!15:35
lorinh1zul: I can't build packages on lucid because it now requires python-multiprocessing http://bazaar.launchpad.net/~openstack-ubuntu-packagers/ubuntu/natty/nova/ubuntu/revision/10815:35
zullorinh1: thats because of the live-migration needs python-multiprocessing now15:36
*** reldan has joined #openstack15:36
lorinh1zul: But, apparently, multiprocessing ships with Python 2.6, which is the base install of Lucid (?)15:36
zulyou sure?15:36
creihtzul: http://docs.python.org/library/multiprocessing.html15:37
zulcreiht: cool ill revert it15:37
lorinh1zul: thanks15:37
zullorinh1: done15:39
zuldendrobates: happy birthday btw15:39
lorinh1zul: Going to test it out right now...15:39
openstackjenkinsProject nova build #641: SUCCESS in 2 min 3 sec: http://hudson.openstack.org/job/nova/641/15:39
openstackjenkinsTarmac: PEP8 0.5.0 cleanup.15:39
dendrobateszul: thanks15:40
*** ramkrsna has quit IRC15:43
*** bkkrw has joined #openstack15:45
lorinh1zul: It works again.15:46
zullorinh1: cool15:46
*** Ryan_Lane has quit IRC15:47
*** lorinh1 has left #openstack15:47
*** paltman has quit IRC15:47
openstackjenkinsProject swift build #220: SUCCESS in 29 sec: http://hudson.openstack.org/job/swift/220/15:48
openstackjenkins* Tarmac: Fixed an import that was causing Connection to not retry on HTTPExceptions.15:48
openstackjenkins* Tarmac: Adds documentation how to the SAIO on how to setup rsyslog to have separate logs for each node15:48
openstackjenkins* Tarmac: s3api: add get acl object and bucket support15:48
openstackjenkinsA simple response that gives the owner to FULL_CONTROL always returns15:48
openstackjenkinsbecause Swift doesn't support a fine acl.15:48
openstackjenkinsThis also adds get acl unit tests.15:48
openstackjenkins* Tarmac: s3api: fix PUT Object Copy15:48
openstackjenkins* Tarmac: Fixes problem with creating a new processed files list in log processing when the download of that file fails.15:48
*** bkkrw has quit IRC15:50
*** dragondm has quit IRC15:52
*** dragondm has joined #openstack15:52
openstackjenkinsProject nova build #642: SUCCESS in 2 min 2 sec: http://hudson.openstack.org/job/nova/642/15:55
openstackjenkinsTarmac: Removed duplicated tests.15:55
openstackjenkinsOne was a straight copy-and-paste, one was different, but the masked test didn't pass.15:55
*** guigui has joined #openstack15:57
*** kashyap has joined #openstack15:58
*** dfg has joined #openstack15:58
*** guigui3 has quit IRC15:58
*** dendrobates is now known as dendro-afk16:00
*** thatsdone has quit IRC16:00
*** dendro-afk is now known as dendrobates16:00
*** enigma1 has quit IRC16:01
openstackjenkinsProject nova build #643: SUCCESS in 2 min 3 sec: http://hudson.openstack.org/job/nova/643/16:01
openstackjenkinsTarmac: Make Authors check account for tests being run with different os.getcwd() depending on how they're run. Add missing people to Authors.16:01
*** dragondm_ has joined #openstack16:02
*** dragondm has quit IRC16:03
*** dragondm_ is now known as dragondm16:03
openstackjenkinsProject swift build #221: SUCCESS in 27 sec: http://hudson.openstack.org/job/swift/221/16:03
openstackjenkinsTarmac: Remove DevAuth16:04
*** mgoldmann has quit IRC16:04
*** enigma1 has joined #openstack16:07
*** Ryan_Lane has joined #openstack16:09
*** enigma1 has quit IRC16:12
*** guigui3 has joined #openstack16:17
*** enigma1 has joined #openstack16:18
*** guigui has quit IRC16:19
*** adjohn has joined #openstack16:19
*** guigui3 has quit IRC16:20
jaypipesjustinsb: around?16:25
*** Guest22204 is now known as kyzh16:26
*** kyzh has joined #openstack16:26
*** nerens_ has joined #openstack16:28
*** befreax has quit IRC16:28
justinsbjaypipes: Hi!16:28
justinsbjaypipes: Sorry about my bazaar b*tching16:29
jaypipesjustinsb: no worries. sorry if I b*tched at you. :) Could you run those commands in the git-bzr bridge and let me know if you have problems?16:29
*** nerens has quit IRC16:30
*** nerens_ is now known as nerens16:30
jaypipesjustinsb: latest comment on the merge prop... tried to demonstrate in bzr commands the scenario you described, and show what *should* happen :)16:30
justinsbjaypipes: Cool - didn't see that email until just now16:30
justinsbjaypipes: I'll give it a go... should help me understand bazaar better!16:30
jaypipesjustinsb: alternately, of course, let me know if those commands *don't* represent the scenario you were trying to describe, ok?16:31
*** kashyap has quit IRC16:32
jaypipessoren, mtaylor: hey packaging gurus, please give https://code.launchpad.net/~devcamcar/nova/adminclient-setup/+merge/53352 some of your love and experience.16:32
* mtaylor loves and experiences jaypipes 16:33
* mtaylor wipes off jaypipes16:33
jaypipeseww. :)16:33
mtaylorjaypipes: ok, I'm going to vote strongly against having a second setup.py in a subdir inside of nova16:34
mtaylorjaypipes: I would argue that it either needs to be its own project, or it needs to be part of nova's setup.py16:34
ttxmtaylor: taht's my opinion as well, though I voiced it in a less definitive way16:35
mtaylorif it needs its own setup.py because PyPI might want to install it and not something else, then it's its own thing with its own release cycle16:35
edaymtaylor: can you build two separate packages/eggs from one setup.py?16:35
ttxmtaylor: it currently has its own binary package in the ubuntu packaging though16:35
mtayloreday: no16:36
ttxeday: that's what it currently does.16:36
mtaylorttx: well, that's different- debian packaging is designed to handle multiple binary packages from a single source- pypi/eggs not so much16:36
ttxright. yes for ubuntu, no for eggs16:36
edayahh, ok16:36
mtaylorbut it's python and doesn't require building c++ - so what's the downside to just having things that need adminclient grabbing nova egg?16:37
mtaylorit's not like it's a huge amount of code or anything16:37
ttxmtaylor: comment on the BMP, please... since the proposer is not around16:37
mtaylorttx: ok. will do16:37
ttxyou're arguing in front of a crowd agreeing with you (my favorites too)16:38
edaybecause that's how I'm setting up burrow... one setup.py in root, another in clients/python/ (and clients/<language>/ will each have it's own config thing).16:38
ttx...except eday.16:38
edaymtaylor: main reason is the much longer dep list for nova than a simple client perhaps?16:38
edayttx: haha, thanks :)16:38
mtayloreday: sure. I mean, I hear that - but setuptools really just doesn't support that model sensibly16:40
mtayloreday: and operates under the assumption that things with their own setup.py are really their own projects16:40
edaymtaylor: how so? just build the dist from the root, and another from a client subdir?16:40
*** imsplitbit has joined #openstack16:41
creihtmtaylor: so why can't you just have more than one python project in a launchpad project?16:41
*** skiold has quit IRC16:41
mtaylorcreiht: you can have as many things as you want in a launchpad project- but in a single branch having recursive self-contained sub-modules is a thing that isn't really well modeled in setuptools16:42
edaymtaylor: ahh, I can see if it's nested in another package (nova/novaadminclient)... I guess I'm talking about if you had /burrowd and /clients/python/burrow packages16:43
openstackjenkinsProject nova build #644: SUCCESS in 2 min 0 sec: http://hudson.openstack.org/job/nova/644/16:43
openstackjenkinsTarmac: just fixing a small typo in nova-manage vm live-migration16:43
mtaylorin this case we're really talking more about setuptools and python project organization ... I mean - I am of course only one person and can be overruled - I'm just saying it's gonna wind up being _way_ more wonky that I think the benefit is worth16:43
mtayloreday: I would argue you will have less headache if you have a burrowd project, and a python-burrow-client project and don't keep them in the same tree16:44
*** ryker has quit IRC16:44
mtayloreday: because there isn't going to be a good way at the toplevel to say "give me everything in this tree" (like there would be in an autotools or cmake-based build)16:44
edaymtaylor: hmm, I wanted to keep the two in one tree since there is such a tight dependency (burrowd depends on burrow for proxy abilities)16:44
mtaylorthen in that case, burrowd and burrowclient should both be serviced by one setup.py - imho16:45
edaymtaylor: but burrowclient doesn't need nearly the same deps as burrowd :)16:45
mtaylorand don't try to model burrowclient and the other language clients as being peers16:45
edaybut they are :)16:46
mtaylorthey aren't - because burrowd is tightly dependent on burrowclient but not on, say, the perl client16:46
edaymtaylor: sure, but outside of that dep, they are peers16:46
mtaylorsure - but in this case we're talking about source management and build systems16:47
mtaylorperl is gonna want to do MakeMaker, ruby is going to want to use the ruby make thing16:47
jaypipessirp_: those comments on the related_images review all make sense?16:47
mtaylortrying to lump all of those into a single package doesn't gain you a whole lot - whereas bundling burrowclient with burrow _does_ actually let you achieve the tight source-level binding you need16:48
edaymtaylor: yup.  I was thinking perhaps a top-level Makefile if you wanted to try installing everything regardless of language (main setup.py, and cd && <lang-specific-pkg-mgmt-thing> for clients/*/16:48
sirp_jaypipes: yeah, great review, zero'd in on the what i thought were the weaknesses too16:48
*** uksysadmin has quit IRC16:48
mtaylorbut then you've made a frakenstein of a build system for very little gain (I'm speaking from experience - I did this with ndb-bindings and drizzle-bindings and no-one other than me could figure out how to deal with it16:48
edaymtaylor: I would say you do get something, a multi-language test suite that can be run on every commit16:48
edayoh, I liked drizzle-bindings :)16:49
jaypipessirp_: ok, cool. :) btw, after 6 days, not a single ping or comment on that bug I filed with sqlalchemy-migrate about the index being removed on a different column when dropping a column. :(16:49
mtayloreday: drizzle-bindings was way too baroque and I've been splitting them into their own projects :)16:49
mtayloreday: it broke ALL THE TIME16:49
sirp_jaypipes: ugh16:50
mtaylorbecause I was trying to interface with build systems that were _not_ designed to be sub-systems16:50
*** kashyap has joined #openstack16:50
mtayloreday: obviously, as I said- it's your project - but for my money, I would a) not care about pulling in extra burrow deps - people are installing either from pypi or debs and will really never notice16:51
edaymtaylor: well, they don't need to be.. it's the same as if you just cd client/x/ and run the right thing16:51
edayI'm not saying they all need to be tightly coupled16:51
edaybut I trust your experience too :)16:51
mtayloreday: right- but then that is a NIGHTMARE for debian packaging16:51
edayreally? I thought it would be simple for each client dir to have it's own deb pkg rules?16:52
jaypipesjustinsb: could you take a gander at https://code.launchpad.net/~blamar/nova/lp732866/+merge/53101 and see if any of it intersects/interferes/conflicts with your test-login stuff?16:52
mtaylorit will be an absolute shitstorm16:52
edaywhats the difference between bzr branch lp:perl-burrow-client and bzr branch lp:burrow && cd client/perl  ?16:52
jaypipessirp_: yeah, no sh*t.16:52
mtayloreday: well - ok - are you talking about releasing multiple different tarballs from each thing- so that client/perl would cut a tarball based in the client/perl dir?16:53
edaymtaylor: yes.. consider each client/*/ to be it's own project16:53
edayjust happens to be in one source repo16:53
mtayloreday: ok - I would be able to do debs for released tarballs16:53
mtayloreday: I would not sensibly be able to do debs from trunk16:53
edayi'm certainly not trying to share build/packaging, just trying to keep the source under one trunk16:54
markwasheday: do you think my personality merge prop still needs fixing?16:54
edaymtaylor: can you exaplin that last statement?16:54
mtaylortrunk packages are built via bzr bd which assumes that the root of the branch is the root of the source thing16:55
mtaylorso it will build debs from a bzr branch based on the root16:55
edaymarkwash: I'll need to look, I assume it's in an active mp?16:55
mtaylorthere is no way to tell it to do its thing from subdirs of branches16:55
edaymtaylor: ahh, bummer. is this a deb pkg tool thing?16:56
markwasheday: as far as I know, it does--but I might be missing something, still a lp newbie16:56
mtaylorwe can, of course, write more scripts to generate tarballs of sub-modules on trunk commits and then apply packaging to those ... but see now we're getting in to having to write complex things to accomplish what we want, which leads me to think what we're doing is needlessly complex16:56
markwasheday: the link is https://code.launchpad.net/~markwash/nova/lp727225/+merge/5275416:56
edaymarkwash: ok, thanks. will look in a minute16:57
mtaylorall of the tools are built around the common source model, which is root of trunk being root of source tree- not root of source tree being two levels down16:57
mtaylorso once we're working from the tarballs, then the tools work again16:57
mtaylorbut we'll have to do lots of custom things to work with the submodule approach _before_ tarballs are cut16:57
edaymtaylor: hmm, I guess. I guess I thought the extra 'cd client/x' would be trivial, but if not, we can certainly break them out16:57
mtayloryeah. it's not trivial (although I have wished in the past that it were, so it's not that I don't understand what you want here)16:58
mtaylorit's _possible_ because anything is possible with enough scripting :)16:58
*** ryker has joined #openstack16:58
edaymtaylor: but as far as the burrow client pkg goes, I really would like to keep at least that one as a separate egg AND in the same repo16:58
edaymtaylor: other languages I'm less concerned about16:59
mtayloris that just because of deps?16:59
edaymtaylor: but it sounds like that would require just as much work as going all out... so not sure what to do16:59
edaymtaylor: yeah16:59
mtaylorbecause if someone is sensible and installs from .debs, then they can get burrowclient without burrow deps (as sep binary packages are actually supported there)17:00
mtaylorI'd argue that trying to squish a model of _actual_ package management onto a pypi-based system which just isn't robust enough for it is going to wind up causing you major headaches17:00
edaymore concerned with PyPI17:00
mtaylorpypi just isn't actually a strong enough package management system17:01
mtaylorit's GREAT - don't get me wrong17:01
mtaylorbut it's not designed to be this flexible17:01
*** Ryan_Lane has quit IRC17:01
mtayloreday: to be honest - I have _NO_ idea why pypi doesn't support multiple eggs for submodules - seems like a major deficiency to me17:01
mtayloreday: as you can't be the first person to want this :)17:01
edayso, should the client module and server modules be in the same pkg? (ie, burrow/client.py and burrow/server.py, ...)? or should they be sep python pkgs too?17:02
mtaylor_I_ think they should just be in the same package - installing a few extra python files just isn't really that big of a deal on modern computers17:02
edaymtaylor: but does that make deb pkging harder? (two modules in same python pkg)?17:03
*** dragondm has quit IRC17:03
mtaylorwith one setup.py? nope17:03
mtaylorthat's a very common model and is handled quite nicely by the tooling17:04
mtayloralthough honestly it would likely split out in to:17:04
mtaylorpython-burrow with all of the things from lib/python17:04
*** iammartian has joined #openstack17:04
mtaylorand then a burrow-server and burrow-client package for any "binaries" that they install into bin/17:04
edaymtaylor: ok... I guess who installs burrow/__init__.py? client, server, or some other base pkg?17:04
*** dragondm has joined #openstack17:05
mtaylorno - the python package itself would just all drop in as a single install in python-burrow17:05
*** Ryan_Lane has joined #openstack17:05
edaymtaylor: ok, so python-burrow, not server or client specific17:05
mtaylorunless I'm missing something as to why they'd need to be split apart17:06
mtaylorbut that's what we do for nova17:06
mtaylorthere is a single python-nova package, and then a bunch of pacakges for each of the different tools that use those python files17:06
edayi suppose that works.. my only concern is a python-burrow dep in the future which requires a bit more work17:06
eday(ie, c/c++ deps, ...)17:06
*** dendrobates is now known as dendro-afk17:06
edayand that;s a bit much for a client that doesn't need it17:06
mtayloryeah - I hear that - if we get there, we can probably sort that out if it gets extremem17:08
edaymtaylor: hmm, also, would it be difficult to have multiple plugin packages? ie, python-burrow, python-burrow-backend-drizzle, python-burrow-frontend-amqp, ...? or again should we just smash it all down in the one pkg?17:08
mtaylorbut honestly, we're just not talking about huge enough dep lists to worry about it for the common case17:08
mtayloryou can totally do multiple plugin pacakges17:08
mtayloreither using pkg_resources for plugin registry - or just installing them into known locations and having your plugin loader scan a dir17:09
edaymtaylor: hmm, ok. right now the plugin loader is the nova one, just a 'import module and getattr class obj'17:10
edayso could live anywhere in your PYTHONPATH17:10
edayI don't need it now, but perhaps in the future where a frontend or backend may not be in the standard lib17:11
justinsbjaypipes: Had a look at the code you asked me to... my tests would be a lot more useful if that merge went in first, as then I could actually test stuff.  But doesn't seem to have attracted any nova-core eyes yet.17:12
justinsbjaypipes: That's https://code.launchpad.net/~blamar/nova/lp732866/+merge/5310117:12
mtayloreday: sorry to be a pain about all of this17:12
edayok, off to reorganize all the burrow dirs now :)17:12
*** mahadev has joined #openstack17:12
edaymtaylor: no worries, your experience is very valuable. i just thought some things would be easier and give us better consistency by having it in one repo :)17:13
*** dendro-afk is now known as dendrobates17:13
jaypipesjustinsb: ya. I wanted your opinion on it b4 I weighed in on it, to be honest.17:13
mtayloreday: yeah ... it would be so nice if that would all work in the way it makes sense in our brains!17:13
jaypipesjustinsb: I didn't see anything bad, just wanted to ask if it would conflict with any of your work, since that looks likely to hit shortly.17:14
justinsbjaypipes: I was debating putting mine to WIP to slow it down until the fix went in17:15
justinsbjaypipes: As mine would be better if his landed first17:15
*** blueadept has joined #openstack17:16
jaypipesjustinsb: ok, cool. sounds good.17:16
*** joearnold has joined #openstack17:19
jaypipessirp_: https://code.launchpad.net/~blamar/nova/lp732866/+merge/53101 approve away if it looks good to you.17:24
*** mgoldmann has joined #openstack17:24
edaymtaylor: so for plugins, would 'python-burrow' pkg contain burrow/backends/__init__.py (interface file), and python-burrow-backends-drizzle contain burrow/backends/drizzle.py (which imports burrow.backends to get interface from __init__)? just want to make sure the layout works for server plugins17:25
sirp_jaypipes: cool, ill check it out17:25
edaymtaylor: as this will be the way to solve the nasty-dep-list-for-just-a-client problem17:26
*** jaypipes is now known as jay-work-on-migr17:26
edayof course, none of this will help pypi, hmm17:27
mtayloreday: probably17:27
mtaylorno. pypi is just going to get you the phone-book - but that's just sort of what you're going to get17:27
jay-work-on-migrsirp_, dprince: there's also this one if you have a minute: https://code.launchpad.net/~jaypipes/glance/bug720816/+merge/5261717:29
devcamcarmtaylor, ttx: thanks for thoughts on the adminclient branch.  its probably time to put it in its own tiny project.  i was going for least disruption originally. :)17:32
mtaylordevcamcar: no prob! I wish it were slightly less work to make a whole new project, actually - I think the number of steps required causes us to balk at times17:33
* mtaylor should make a tool to do that17:33
devcamcarmtaylor: hah, yea - let me know when its ready17:34
*** enigma1 has quit IRC17:37
*** spectorclan_ has quit IRC17:37
*** matclayton has left #openstack17:39
*** MarcMorata has quit IRC17:42
*** daveiw has left #openstack17:44
*** clauden has joined #openstack17:45
*** gasbakid has joined #openstack17:46
*** DigitalFlux has quit IRC17:54
*** shentonfreude has joined #openstack17:54
*** paltman has joined #openstack17:55
*** mahadev has quit IRC18:00
*** mahadev has joined #openstack18:00
*** irahgel has left #openstack18:11
*** spectorclan_ has joined #openstack18:13
*** ramkrsna has joined #openstack18:19
*** ramkrsna has joined #openstack18:19
uvirtbotNew bug: #735629 in nova "Fix leaky S3 ImageService abstraction" [Undecided,New] https://launchpad.net/bugs/73562918:21
zulanyone seen this? http://pastebin.ubuntu.com/580697/18:23
uvirtbotNew bug: #735636 in nova "Standardize BaseImageServices handling of timestamps" [Undecided,New] https://launchpad.net/bugs/73563618:26
jay-work-on-migrsirp_: figured out the solution to the migrate issues...18:27
sirp_jaypipes: ORLY?18:27
jay-work-on-migrsirp_: removing the index=True on the deleted and is_public columns fixes all problems.18:29
sirp_jaypipes: interesting, so that was just left over from the declarative_base style?18:30
jay-work-on-migrsirp_: the bugs only happen on SQLite. Has to do with the "workarounds" that sqlalchemy-migrate does for DROP COLUMN...18:30
*** lionel_ has quit IRC18:30
*** lionel_ has joined #openstack18:31
*** iammartian has quit IRC18:33
uvirtbotNew bug: #735641 in nova "Latest trunk fails." [Undecided,New] https://launchpad.net/bugs/73564118:41
*** JuanPerez has joined #openstack18:42
*** rcc has quit IRC18:43
kpepplezul: see my answer to https://bugs.launchpad.net/nova/+bug/73564118:47
uvirtbotLaunchpad bug 735641 in nova "Latest trunk fails." [Undecided,New]18:47
zulkpepple: thanks18:48
kpepplezul: np18:49
*** maplebed has joined #openstack18:49
*** herki has joined #openstack18:52
*** mrsrikanth has quit IRC18:53
*** MotoMilind has joined #openstack18:53
*** syah has joined #openstack18:55
*** syah has quit IRC19:01
sirp_jaypipes: fyi, i added some more comments to https://code.launchpad.net/~rconradharris/nova/related_images/+merge/5337419:02
*** MotoMilind has quit IRC19:03
*** Pentheus has joined #openstack19:03
edaywestmaas: where did the default deserializer break with the personalities branch?19:05
edaywestmaas: oops, wrong person :)19:05
*** maplebed has quit IRC19:06
*** MarcMorata has joined #openstack19:06
*** MotoMilind has joined #openstack19:07
*** syah has joined #openstack19:08
openstackjenkinsProject nova build #645: SUCCESS in 2 min 1 sec: http://hudson.openstack.org/job/nova/645/19:08
openstackjenkinsTarmac: Fixed lp732866 by catching relevant `exception.NotFound` exception. Tests did not uncover this vulnerability due to "incorrect" FakeAuthManager. I say "incorrect" because potentially different implementations (LDAP or Database driven) of AuthManager might return different errors from `get_user_from_access_key`.19:08
openstackjenkinsAlso, removed all references to 'bacon', 'ham', 'herp', and 'derp' and replaced them with hopefully more helpful terms.19:08
openstackjenkinsLong story short it addresses the immediate issue while throughly ignoring the larger issue, which is correctly testing all implementations of Auth. I find this acceptable as currently the future of auth is in flux.19:08
*** maplebed has joined #openstack19:08
*** dmshelton has quit IRC19:11
*** joearnold has quit IRC19:15
btorchanyone know how I can get the glance 2011.2~bzr87 version ? I have just installed glance and it picked up 2011.2~bzr81 . But I do see a source pkg for 2011.2~bzr8719:16
btorch^^ that's for  mavarick using the ppa:glance-core/trunk repo19:16
uvirtbotbtorch: Error: "^" is not a valid command.19:16
tr3buchetany of the folks from grid dynamics around?19:16
*** dendrobates is now known as dendro-afk19:22
*** CloudChris has quit IRC19:26
*** paltman has quit IRC19:27
*** stewart has quit IRC19:27
*** kashyap has quit IRC19:30
ttxadiantum: ^19:31
ttxBK_man: ^19:31
BK_mantr3buchet: yep19:31
ttxTeam meeting in 90 min. in #openstack-meeting19:31
BK_manothers are coming...19:32
adiantumhello gents19:32
tr3bucheti wanted to talk to someone about libvirt multi-nic19:32
BK_mantr3buchet:  adiantum & reldan19:32
BK_mancan somebody help me to run nova-api? it going to endless cycle... http://paste.openstack.org/show/891/19:34
BK_manI'm on last rev 80719:35
*** hazmat has quit IRC19:43
jarrodweird, what other modules does rxfax require other than spandsp?19:44
jarrodI am getting "Fax processing not successful - result (5) The HDLC carrier did not stop in a timely manner." on one 1.0.7 box, and on another it works perfectly every time19:45
ttxjarrod: warning, your time machine just moved you 20 years into the future, from faxes to cloud.19:47
jarrodi hate fax19:50
jarrodif i didnt have customers id unload the module19:50
*** hazmat has joined #openstack19:51
jarrodman i wonder if im missing a dependency or something19:55
*** brd_from_italy has joined #openstack19:55
*** jay-work-on-migr is now known as jaypipes19:59
*** paltman has joined #openstack20:01
*** dendro-afk is now known as dendrobates20:01
jarrodi am in the wrong freaking channel20:01
*** zenmatt has quit IRC20:01
edayjaypipes: you around?20:04
jaypipeseday: yes20:04
edayjaypipes: if you have a moment, could you check https://code.launchpad.net/~markwash/nova/lp727225/+merge/52754  ? specifically looking for an opinion on overriding the default XML deserilization.20:05
jaypipeseday: yep20:05
*** reldan has quit IRC20:06
edayjaypipes: the current common deserializer doesn't do what this patch needs, and rather than change the common one (possibly breaking other uses), a new one was added for this specific case20:06
eday(new one is nova/api/openstack/servers.ServerCreateRequestXMLDeserializer)20:07
edaymarkwash: ^ any other comments?20:07
*** BK_man has quit IRC20:08
markwasheday, jaypipes: no further comments, there are some notes about it in the merge prop comments20:08
jaypipesmarkwash: ya, reading now...20:09
*** fabiand_ has joined #openstack20:11
jaypipesmarkwash, eday: I'd actually prefer in the code that we use the term injected_files instead of either onset_files or personality...20:11
edayjaypipes: personality is part of the existing API, so can't change that, but I agree something more clear for 'onset' would be good. might be better as a follow up patch (since this doesn't introduce onset)20:12
*** adiantum has quit IRC20:12
jaypipeseday: not what is returned from the API, but how we name the variables in the code. "personality" is a terrible descriptor of these things, but I recognized it's what is in the API.20:13
markwashjaypipes: basically, you want it to be injected_files in the compute layer and calls into that layer20:15
*** glenc_ has joined #openstack20:15
edayjaypipes: sure, but I'm fine with 'personality' var in nova.api.openstack too for consistency. nova.compute.* should be something more sane :)20:15
jaypipesmarkwash: ya, just recommending s/onset_files/injected_files/g :)20:15
_0x44eday: Why can't we change it?20:15
eday_0x44: it's already part of the existing API. we can change it in new versions of the API20:16
*** glenc has quit IRC20:16
*** glenc_ is now known as glenc20:16
*** mgoldmann has quit IRC20:17
markwashjaypipes: I don't mind making that change in the compute api and quota20:18
*** ramkrsna has quit IRC20:18
_0x44eday: I'm kind of in the camp that says that the 1.0 OS API can be thrown away since we're going to throw it away in three months anyway to "upgrade" to 1.1.20:19
* _0x44 might be that camp for all he knows.20:19
*** dprince has quit IRC20:20
eday_0x44: ok, I'll let you have that discussion with folks in the other camp. I'm done with API discussions for a while :)20:20
jaypipesmarkwash: cool. :) I'm putting a note on the review, so no worries...20:20
_0x44eday: Maybe I should join your camp, it's probably safer and has fewer asbestos suits.20:21
*** patrick has joined #openstack20:22
*** patrick is now known as Guest2101420:22
*** paltman has quit IRC20:22
*** joearnold has joined #openstack20:22
eday_0x44: it's way too fragmented to have any useful conversations, so not having any can be just as productive :)20:22
jaypipesHave I ever mentioned how very much I hate XML? :)20:23
sorenjaypipes: Not that I recall. How much do you hate XML, exactly?20:23
sorenYou've piqued my interest.20:23
edayjaypipes: http://twitter.com/#!/samkeen/status/4306889660497920120:25
jaypipessoren: quite a bit.20:25
sorenjaypipes: Wow.20:26
soren"XML is like violence. If it doesn't solve your problem, you're not using enough of it." has to be my all time favourite XML related quote.20:26
jaypipesmarkwash: hey, could you explain why the standard Serializer was not good enough for the <server> element? Was there a bug in it?20:27
jaypipesblamar: you an XML fan?20:28
edayjaypipes: it's the <file path="...">contents</file> element20:28
sorenThe runner up is "Java is a DSL to transform big XML documents into long exception stack traces", but that's more of a swing at Java than XML.20:28
blamarjaypipes, I'm a fan of it's hopefully ultimate demise :)20:28
jaypipesblamar: ah :)20:28
edayjaypipes: it currently will ignore the 'path' attribute and only return the contents20:29
jaypipeseday: why does the regular Serializer puke on that? seems pretty standard to me...20:29
edayjaypipes: it's really a limitation/bug of the current one20:29
jaypipeseday: ah, k. so the current one doesn't handle attrs?20:29
jaypipeseday: only value contents?20:29
edayjaypipes: not for that specific case, it does with other elements20:29
markwashjaypipes: there are multiple things that need to be handled differently20:29
markwashjaypipes: the <file> element label must be dropped20:30
ttxTeam meeting in 30 min. in #openstack-meeting20:30
edayjaypipes: not the hard coded node type checks in the _from_xml_node method20:30
justinsbjaypipes: It's only a matter of time before JSON evolves into XML...20:30
markwashjustinsb: NOOOOO20:30
justinsbSchemas, namespaces etc20:30
justinsbIt's all coming20:30
jaypipesmarkwash: ok, thx.20:31
justinsbThose things weren't all just "committees gone wild" :-)20:31
jaypipesjustinsb: those are all entirely pointless.20:31
justinsbjaypipes: How are we going to do our extensions in JSON?20:31
justinsbjaypipes: While I'm not a fan of extensions in the API, that's why we have namespaces20:32
markwashjaypipes: I think we need to address serialization across the openstack api soon: I think we need to move away from a centralized approach in favor of the open/closed principle20:32
jaypipesjustinsb: uhm, {'extensions': []}20:32
justinsbjaypipes: Not that bit!  The bit where any extension can put their own stuff into any element...20:32
jaypipesjustinsb: XML is a bunch of Java-centric developers over-architecting what should have been a simple definition language.20:32
justinsbjaypipes: I think you're just trying to wind me up now :-)20:33
jaypipesjustinsb: :P20:33
justinsbjaypipes: Though I do agree that attributes are entirely pointless20:33
* blamar gets popcorn20:33
jaypipesjustinsb: thought I'd get you going with that one :)20:33
justinsbjaypipes: But I think that's the result of people having to support their equivalent of the CloudServers API20:33
justinsbjaypipes: Let the wars begin :-)20:34
* blamar readies his "tabs vs. spaces" document20:34
*** imsplitbit has quit IRC20:34
jaypipeshehe, I've got to actually do *work*...20:34
*** pvo has quit IRC20:34
*** maplebed has joined #openstack20:34
* jaypipes goes off to finish review on markwash 20:34
edayomg, 30 active and 20 wip mps. justinsb is probably to blame for about a third of it20:34
devcamcarcreiht, notmyname: my quest to have cyberduck handle different swift endpoints has met an untimely end.  http://trac.cyberduck.ch/ticket/5735#comment:620:34
devcamcarthough i can make it work in general, you have to set the auth endpoint globally, which is lame on cyberducks side.20:35
markwashin all seriousness, those concerned about xml might want to look at how extensibility is added to json in the 1.1 spec jorge emailed out on march 220:35
ttxeday: we need to get him some real day job.20:35
edayperhaps we need a requirement to do 2 reviews for every 1 you propose rule20:36
justinsbeday: That's not a bad rule actually... like a ul / dl ratio20:37
edayjustinsb: for you!20:37
*** pvo has joined #openstack20:37
*** zul has quit IRC20:37
justinsbeday: It would mean I'd have to do a lot more reviews... isn't that the idea?20:37
edayjustinsb: yup :)20:37
edayoh, I read that wrong20:37
edayignore me, but I stand by my proposed rule :)20:38
justinsbeday: You could propose it at the next openstack meeting.  When is that exactly?  I'm confused :-)20:39
*** paltman has joined #openstack20:40
*** Guest21014 has quit IRC20:41
*** bcwaldon has quit IRC20:41
edayjustinsb: 19 minutes20:41
*** jc_smith has joined #openstack20:43
*** edconzel has joined #openstack20:46
edayjaypipes: one alternative for the (de)serializer was to allow specifying custom node methods via the _serialization_metadata member. This could use the current one and override just for 'personality' elements20:46
*** ipersite has joined #openstack20:46
ipersitehi all20:47
ipersiteanybody here can give me some help please?20:48
jaypipesipersite: people will once they know what your issues are ;)20:48
creihtdevcamcar: heh20:49
ipersiteok thanks but nobody answered so I couldn't figure out if there were active people in the channel20:50
creihtmarkwash: do you have a link?20:50
ipersiteok, i tried the nova.sh script20:50
ipersitefrom vishvananda20:50
markwashcreiht: to the mergeprop?20:51
ipersiteand with this script, euca-describe-images doesn't work.20:51
creihtto the json extension spec20:51
markwashcreiht: one sec20:51
ipersiteIt says: UnknownError: An unknown error has occurred. Please try your request again.20:51
*** shentonfreude has quit IRC20:51
pvoipersite: can you paste everything to http://paste.openstack.org/?20:52
markwashcreiht: http://docs.openstack.org/openstack-compute/developer/cs-devguide.pdf20:52
pvoer http://paste.openstack.org ?20:52
pvoit'll help people look at the errors and logs20:52
*** JuanPerez has quit IRC20:52
ipersitewhat do i need to paste^20:52
ipersitewhat do i need to paste?20:52
ipersitecommand outputs, request...20:53
*** allsystemsarego has quit IRC20:53
pvothe logs from the tool or the service that is erroring20:53
*** shentonfreude has joined #openstack20:53
pvoI'm assuming its the api20:53
creihtmarkwash: ok, which section? :)20:53
creihtoh I think I found it20:53
creihtmarkwash: you mean the the /extensions stuff?20:54
pvoipersite: what do the api logs show when you do that?20:55
markwashcreiht: section 3.11, especially around and after page 3720:56
* creiht probably shouldn't comment :/20:57
*** jc_smith has left #openstack20:57
*** reldan has joined #openstack20:57
ttxTeam meeting starting in 2 min. in #openstack-meeting, please join !20:58
justinsbUh oh ... don't tell jaypipes someone put namespaces into his JSON ;-)20:58
ipersitei'm searching for the log file 'cause screen doesn't let me see the full output20:58
pvoipersite: it does. ctrl-a then [20:58
pvolets you page up20:58
*** BK_man has joined #openstack20:59
ipersiteoh thanks20:59
markwashjaypipes: when I i18n a string, do I need to add it to a .pot file?20:59
jaypipesmarkwash: nope. LP does all that for ya.20:59
*** h0cin has quit IRC20:59
markwashjaypipes: so its just that _("such and such")20:59
jaypipesmarkwash: yup. only complexity is when you have >1 variable substitution, but I don't think any of yours were that way?21:00
*** adiantum has joined #openstack21:00
ipersitei tried but nothing happend21:02
*** openfly has left #openstack21:02
ipersitewait... i forgot the AltGr21:03
ipersiteok i updated the paste21:06
ipersitepvo can you check there? http://paste.openstack.org/show/893/21:07
ipersitethis is the correct one21:08
*** ewindisch has joined #openstack21:08
jaypipesblamar: you know if brian waldon is on IRC?21:10
blamarjaypipes, he's out atm21:11
blamarcan I help you with something?21:11
jaypipesblamar: k. :) is he titan?21:11
blamarjaypipes, yup21:11
*** bcwaldon has joined #openstack21:11
jaypipesblamar: no, no big deal. I was just going to say nice work on the version split patch, and recommend s/view/resource...21:11
*** zul has joined #openstack21:12
*** blueadept has quit IRC21:12
blamarjaypipes, hmm, I can see resource making sense. I think the logic was MVC-inspired? although I didn't work with him on that one21:13
ipersitefound a temporary solution... (removing default images)21:13
markwashjaypipes: do you really want me to rename onset files deep down in the virt code? I don't have the ability to do end-to-end tests of this21:15
*** fabiand_ has quit IRC21:17
markwashjaypipes: nm21:17
jaypipesmarkwash: if you want to leave that for a later refactor/discussion, I'm game.21:17
justinsbjaypipes: I owe bazaar an apology.  Your example did work.  Wasn't able to get very far with git-bzr and local repos sadly (not really what it's meant for I guess).21:21
justinsbjaypipes: So I owe you an apology as well!21:21
justinsbjaypipes: I guess I'll give bazaar another chance...21:21
jaypipesjustinsb: no worries mate :P)21:21
jaypipesjustinsb: P wasn't supposed to be in there.. :)21:21
*** alekibango has joined #openstack21:21
*** lionel_ has quit IRC21:22
justinsbjaypipes: It's OK ... I barely know what most of those things really mean anyway!21:22
*** lionel_ has joined #openstack21:23
edayjaypipes: you didn't really comment on the serialization issue, thoughts? :)21:26
jaypipeseday: I didn't comment cuz I'm kinda on the fence about it :) I think it could be refactored in the future to be cleaner, but I don't think it's a huge baddie for markwash's patch...21:27
edayjaypipes: ok21:27
BK_manjaypipes: do we have glance per-commit tarballs like the ones we have for nova?21:29
jaypipesBK_man: you mean the auto-built PPA packages? I think so, yes21:30
BK_manjaypipes: nope. I meant this: nova.openstack.org/tarballs/21:30
*** ppetraki has quit IRC21:31
jaypipesBK_man: did you try http://glance.openstack.org/tarballs/? ;)21:31
*** adiantum_ has joined #openstack21:31
*** ipersite has quit IRC21:31
BK_manjaypipes: thanks for pointing me to that!21:31
* BK_man going to update glance RPMs for Cactus21:32
*** adiantum has quit IRC21:34
*** CloudChris has joined #openstack21:35
markwashjaypipes: about the backslashes in my tests. . the plan was those would escape newlines, and as far as my testing shows, that's just what they're doing21:36
markwashjaypipes: is that unexpected? I'm trying to find if that is documented somewhere so I can make sure I'm not just taking advantage of a side effect21:37
*** bkkrw has joined #openstack21:37
*** adiantum__ has joined #openstack21:38
*** adiantum_ has quit IRC21:38
jaypipesmarkwash: sorry, so you *do* or *don't* want to test with newlines in the XML?21:39
markwashjaypipes: I *don't*21:40
*** hazmat has quit IRC21:40
markwashjaypipes: actually I'm unsure if the api is supposed to handle arbitrary whitespace between elements21:40
jaypipesmarkwash: ah, OK. Yes, sorry, I misunderstood the code and thought you were simply misunderstanding how """ worked. Sorry about that. Might be good to add an additional test case for a piece of XML with newlines in it, then?21:41
edaymarkwash: it should handle ws, since xml/json allow it21:41
jaypipesmarkwash: AFAIK, XML parsers *should* disregard whitespace between elements, but *not* non-whitespace.21:41
markwashjaypipes: okay, in that case, I will remove the backslashes and fix the parser21:42
*** adiantum__ is now known as adiantum21:42
markwashjaypipes: s/parser/deserializer/21:42
jaypipesmarkwash: ok. sorry if I've caused a whole bunch of confusion!21:42
markwashjaypipes: nope, none at all. this still seems like an important catch21:43
*** shentonfreude has quit IRC21:45
uvirtbotNew bug: #735759 in openstack-dashboard "creating a new user in dashboard doesn't create a new user in nova" [Critical,Confirmed] https://launchpad.net/bugs/73575921:46
*** spectorclan_ has quit IRC21:46
*** patcoll has quit IRC21:46
edayttx: on http://wiki.openstack.org/Summit, you say propose common blueprints to openstack-dev. this doesn't exist :)  (openstack-common?)21:48
*** BK_man has quit IRC21:54
*** zenmatt has joined #openstack21:58
edaysandywalsh: any update on zones3 or the zone ideas you were mentioning last week?22:00
*** shentonfreude has joined #openstack22:03
*** grapex has joined #openstack22:03
sandywalsheday, zones3 has a couple of tweaks pending.22:03
sandywalsheday, I was out of town until yesterday, going to try to get notes put together tonight.22:03
sandywalsheday, still needing some reviews on zones3 though :)22:04
edaysandywalsh: ahh, ok22:04
*** miclorb_ has joined #openstack22:04
edaysandywalsh: I reviewed, was waiting on feedback :)22:04
sandywalsheday, ah, ok, I'll get that to you shortly ... catching up. I think I skimmed it and it all seemed reasonable, but stay tuned22:05
sandywalsheday, trying to get zones4 in a state for freeze22:05
edaysandywalsh: might want to switch to WIP22:05
sandywalsheday, good idea22:05
*** littleidea has joined #openstack22:07
*** syah has quit IRC22:08
*** shentonfreude has quit IRC22:09
*** adiantum has quit IRC22:10
*** syah has joined #openstack22:11
*** bkkrw has quit IRC22:11
*** dendrobates is now known as dendro-afk22:12
*** gasbakid has quit IRC22:13
*** pvo has quit IRC22:14
*** adiantum has joined #openstack22:15
*** gondoi has quit IRC22:15
*** shentonfreude has joined #openstack22:15
devcamcaranyone ever had problems with pip taking FOREVER to locate and start downloading stuff?22:17
devcamcarit is stalling out for a good 2 or 3 minutes per item it installs22:18
devcamcardriving me wonko22:18
*** edconzel has quit IRC22:21
sirp_devcamcar: weird, can you strace the process, see what it's hanging on?22:22
*** MarcMorata has quit IRC22:24
ironcamel2eday: you there?22:24
*** Ryan_Lane has quit IRC22:25
edayironcamel2: yes22:25
*** adiantum has quit IRC22:25
*** bkkrw has joined #openstack22:25
ironcamel2eday: thanks for reviewing https://code.launchpad.net/~rackspace-titan/nova/openstack-api-version-split/+merge/5351022:25
ironcamel2i pushed new code per your suggestions, and added comment22:25
edayironcamel2: == Naveed ? :)22:25
edayironcamel2: cool. my main point with the last bit is I don't think the view versioning will be needed in the current form once controllers are versioned (since they can call the specific version directly)22:27
edayironcamel2: so I saw it as solving a problem that won't exist in the next rev22:27
ironcamel2eday: wouldn't a version 1.1 controller, use a version 1.1 view?22:27
ironcamel2we still want to separate view logic from the controller right?22:27
*** paltman has quit IRC22:28
edayironcamel2: yes, it will need to be formatted in v1.1, but that can be a direct call (not check a env version being set). I wasn't aware of any view/controller split22:28
edayironcamel2: was that part of a blueprint somewhere?22:28
ironcamel2eday: we added a views package22:29
*** adiantum has joined #openstack22:29
ironcamel2before the construction of the data model to be serialized was right insided the controller22:30
devcamcarsirp_:  it is dying on a socket connection. i have an extremely normal network config where i am right now.  there aren't any proxies or weirdness for it to go through, just a router.22:30
ironcamel2in _translate_detail_keys() or some such function22:30
edayironcamel2: yup, aware of what is/was there, didn't know there was a push to split them into a new package though (until this branch)22:31
ironcamel2we think it is necessary for sanity as we make 1.0 and 1.1 representations of responses22:31
*** lionel_ has quit IRC22:32
ironcamel2eday: there was no bp specifically for separating out view logic. it is more of an implementation detail. for supporting all the new 1.1 features/changes.22:33
edayironcamel2: ok22:34
*** reldan has quit IRC22:34
edayironcamel2: I guess I was thinking of doing things differently (extending what we already do), but this will work too22:35
devcamcarsirp_: here it is:Could not fetch URL http://www.pycrypto.org/files/pycrypto-2.3.tar.gz (from http://pypi.python.org/simple/pycrypto/): HTTP Error 302: The HTTP server returned a redirect error that would lead to an infinite loop22:35
devcamcarsirp_: but it's doing that for all of them and then eventually working22:35
ironcamel2eday: cool :)22:35
*** lionel_ has joined #openstack22:37
ironcamel2eday: also, sorry for that initial comment in the merge prop. i was being silly, it was an inside team joke22:37
edayironcamel2: I'm not sure we'll need the whole get_view_builder thing once we have versioned controllers, since the controller can call the needed versioned views directly22:37
*** reldan has joined #openstack22:37
ironcamel2eday: yes, i agree that might be unnecessary at that point22:38
edayironcamel2: if we want to go this route, I'll re-review those other parts in more detail then.. stand by22:38
devcamcarsirp_: this hangs: $ curl http://www.pycrypto.org/files/pycrypto-2.3.tar.gz22:39
devcamcarsirp_: but opening url in my browser works instantly22:39
sirp_devcamcar: sorry on a call, which catch up in a sec22:39
devcamcarsirp_: np!22:40
*** adiantum has quit IRC22:42
*** shentonfreude has quit IRC22:42
*** brd_from_italy has quit IRC22:42
*** shentonfreude has joined #openstack22:43
devcamcarsirp_: strange, it's using ipv622:44
*** adiantum has joined #openstack22:47
devcamcarso anyone know how to make pip not using ipv6?22:49
*** bcwaldon has quit IRC22:49
devcamcarsirp_: the fix: disable ipv6 on my wifi adapter =x23:02
devcamcarnever get those 45 mins of my life back!23:02
*** rpedde has quit IRC23:02
kpeppledevcamcar: are you on a mac ?23:03
devcamcarkpepple: yep23:03
kpeppledevcamcar: hmmm .. thx23:04
*** enigma has joined #openstack23:04
edaydevcamcar: well, there's the bigger problem :)23:06
devcamcareday: yea totally23:06
*** enigma has quit IRC23:06
* eday hates on apple for a while23:07
*** bcwaldon has joined #openstack23:08
*** Takinator has joined #openstack23:10
*** MotoMilind has quit IRC23:10
*** KenD has joined #openstack23:11
*** dfg has quit IRC23:11
*** nijaba has quit IRC23:14
*** MotoMilind has joined #openstack23:14
*** nijaba has joined #openstack23:16
*** nijaba has joined #openstack23:16
*** iammartian has joined #openstack23:17
ironcamel2eday: i'm cool with nova.api.openstack.version, unless you can come up with something shorter :)23:18
edayironcamel2: no preference here, just something a bit more descriptive than 'version' :)23:19
ironcamel2eday: openstack.version? api.version?23:20
edayironcamel2: sure :)23:20
*** msassak has quit IRC23:21
*** Takinator has left #openstack23:22
*** CloudChris has quit IRC23:23
*** alekibango has quit IRC23:23
*** alekibango has joined #openstack23:25
*** adiantum has quit IRC23:27
edaymtaylor: so, looking at pypi, I see some projects (like collective.*) define the parent packages (__init__.py) in every package. do you know if the latest version wins with this model?23:27
*** reldan has quit IRC23:34
markwashjaypipes: I made some fixes to the personality file injection merge prop according to your feedback23:36
*** adiantum has joined #openstack23:39
*** nijaba has quit IRC23:41
*** nijaba has joined #openstack23:43
*** nijaba has joined #openstack23:43
*** rnirmal has quit IRC23:44
mtayloreday: yeah - as long as those files are just empty __init__.py files, then the way pypi installs things it won't be a problem23:45
mtayloreday: _debian_ can't do that, because you can't have the same file in two packages23:46
*** ewindisch has quit IRC23:50
ironcamel2eday: just pushed your suggestions to branch of the merge prop23:57
ironcamel2thanks for your suggestions23:58

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