| *** Binbin has joined #openstack-dev | 01:32 | |
| *** zaitcev has quit IRC | 03:06 | |
| *** zaitcev has joined #openstack-dev | 03:16 | |
| *** bcwaldon has quit IRC | 03:27 | |
| *** markwash_ has quit IRC | 03:38 | |
| *** markwash has joined #openstack-dev | 03:38 | |
| *** zaitcev has quit IRC | 03:41 | |
| *** zaitcev has joined #openstack-dev | 03:48 | |
| *** Zangetsue has joined #openstack-dev | 04:39 | |
| *** cloudgroups has joined #openstack-dev | 04:59 | |
| *** cloudgroups has left #openstack-dev | 05:05 | |
| *** markwash has quit IRC | 05:41 | |
| *** zaitcev has quit IRC | 06:35 | |
| *** cloudgroups has joined #openstack-dev | 06:56 | |
| *** ttx has quit IRC | 07:01 | |
| *** ttx has joined #openstack-dev | 07:01 | |
| *** ttx has quit IRC | 07:01 | |
| *** ttx has joined #openstack-dev | 07:01 | |
| *** cloudgroups has quit IRC | 07:08 | |
| *** RobertLaptop has quit IRC | 07:22 | |
| *** RobertLaptop has joined #openstack-dev | 07:23 | |
| *** cloudgroups has joined #openstack-dev | 07:36 | |
| *** cloudgroups has left #openstack-dev | 07:36 | |
| *** Binbin has quit IRC | 08:39 | |
| *** cloudgroups has joined #openstack-dev | 08:45 | |
| *** cloudgroups has left #openstack-dev | 08:45 | |
| *** Binbin has joined #openstack-dev | 08:55 | |
| * ttx tries to reach the bottom of the bug/branchmail | 09:35 | |
| *** agarwalla has joined #openstack-dev | 10:17 | |
| *** Binbin has quit IRC | 10:34 | |
| *** agarwalla has quit IRC | 11:12 | |
| *** nadley_ has joined #openstack-dev | 11:50 | |
| *** nadley_ has left #openstack-dev | 11:50 | |
| *** BK_man has joined #openstack-dev | 12:32 | |
| *** bcwaldon has joined #openstack-dev | 12:57 | |
| *** Zangetsue_ has joined #openstack-dev | 13:10 | |
| *** Zangetsue has quit IRC | 13:11 | |
| *** Zangetsue_ is now known as Zangetsue | 13:11 | |
| *** dprince has joined #openstack-dev | 13:22 | |
| *** BK_man has quit IRC | 13:23 | |
| *** BK_man has joined #openstack-dev | 13:23 | |
| *** Zangetsue_ has joined #openstack-dev | 13:24 | |
| *** Zangetsue has quit IRC | 13:26 | |
| *** Zangetsue_ is now known as Zangetsue | 13:26 | |
| *** Zangetsue_ has joined #openstack-dev | 13:28 | |
| *** Zangetsue has quit IRC | 13:32 | |
| *** Zangetsue_ is now known as Zangetsue | 13:32 | |
| *** mattray has joined #openstack-dev | 14:01 | |
| *** Zangetsue has quit IRC | 14:04 | |
| *** mattray has quit IRC | 14:05 | |
| *** jkoelker has joined #openstack-dev | 14:08 | |
| *** mattray has joined #openstack-dev | 14:12 | |
| *** blamar has quit IRC | 14:16 | |
| *** dragondm has joined #openstack-dev | 14:30 | |
| *** johnpur has joined #openstack-dev | 14:39 | |
| *** ChanServ sets mode: +v johnpur | 14:39 | |
| *** yamahata_lt has quit IRC | 14:41 | |
| *** cp16net has joined #openstack-dev | 14:41 | |
| sandywalsh | hey y'all. I'm bootstrapping a new nova deploy and getting this message http://paste.openstack.org/show/1346/ ... does api need to be running now during the nova-manage steps? | 14:54 |
|---|---|---|
| *** mattray has quit IRC | 14:54 | |
| *** blamar has joined #openstack-dev | 14:58 | |
| *** rnirmal has joined #openstack-dev | 15:00 | |
| markwash1 | nova-core: can you take a look at my merge proposal? https://bugs.launchpad.net/nova/+bug/765782 which is intended to fix lp765782 | 15:02 |
| pyhole | LP Bug #765782: postgres schema needs updating [Status: In Progress, Assignee: Mark Washenberger] https://bugs.launchpad.net/bugs/765782 | 15:02 |
| *** markwash1 is now known as markwash | 15:02 | |
| dprince | sandywalsh: I don't think so. Do you have an openssl.cnf file in /var/lib/nova/CA? | 15:04 |
| sandywalsh | dprince, looking | 15:04 |
| dprince | sandywalsh: It just looks like it is failing to find your openssl.cnf file. | 15:05 |
| sandywalsh | dprince, don't have one anywhere ... is that a new requirement? | 15:05 |
| *** mattray has joined #openstack-dev | 15:06 | |
| sandywalsh | dprince, and this is install from source, not the PPA | 15:07 |
| dprince | sandywalsh: Well. That is where nova creates it in my setup. I'm using Chef. Essentially the /usr/share/pyshared/nova/CA/openssl.cnf.tmpl gets copied into /var/lib/nova/CA. | 15:07 |
| dprince | sandywalsh: Ah. Yep. I'm using PPA packages. | 15:08 |
| dprince | sandywalsh: Doesn't setup.py take care of creating it? | 15:08 |
| sandywalsh | dprince, I shouldn't need to do that should I? | 15:08 |
| sandywalsh | dprince, can you paste your openssl.cnf file so I can see what I'm missing? | 15:09 |
| dprince | sandywalsh: How about you send me your ssh key and I can just let you shell into a working cloud server running the PPA packages? Then you have whatever you need. | 15:10 |
| sandywalsh | dprince, ah, that's ok, I need to resolve this ... shouldn't need to do this to run from source. Something wicked this way has come. | 15:12 |
| dprince | sandywalsh: Yep. That probably shouldn't be a requirement. The VPN stuff is only used for cloud pipe right? | 15:13 |
| dprince | sandywalsh: anyway good luck. | 15:13 |
| sandywalsh | dprince, thx | 15:13 |
| *** markvoelker has joined #openstack-dev | 15:16 | |
| *** markvoelker has quit IRC | 15:21 | |
| *** BK_man has quit IRC | 15:24 | |
| *** markvoelker has joined #openstack-dev | 15:24 | |
| *** BK_man has joined #openstack-dev | 15:29 | |
| *** fyang has joined #openstack-dev | 15:36 | |
| *** blamar has quit IRC | 15:40 | |
| *** blamar has joined #openstack-dev | 15:40 | |
| *** bcwaldon has quit IRC | 15:58 | |
| *** dragondm has quit IRC | 16:04 | |
| dprince | [A[A[A[A[A[A[A[A[A[ADuplicity will do incremental backups of a given directory, encrypt them with your made up password, and dump the backups in a cloud files container. If you need a cloud files account, please talk to your manager. Also, you should only use one container per unique folder you want to backup. | 16:07 |
| *** dprince has quit IRC | 16:08 | |
| *** dprince has joined #openstack-dev | 16:08 | |
| westmaas | +1 | 16:09 |
| markwash | dprince: +1 | 16:09 |
| dprince | oops. ignore that. | 16:09 |
| markwash | dprince: -1 | 16:10 |
| creiht | lol | 16:12 |
| vishy | sandywalsh: it has been that way for a while. It is created the first time you run nova-api | 16:22 |
| vishy | (before you can use nova-manage project zip) | 16:23 |
| vishy | you can use nova-manage project e though | 16:23 |
| vishy | without running it firsh | 16:23 |
| dprince | vishy: So the openssl certs are only used for OpenVPN (cloud pipe) stuff right? | 16:28 |
| vishy | vpn and bundling images | 16:29 |
| vishy | through ec2 | 16:29 |
| dprince | vishy: gotcha. Just wondering how feasible it would be to make the creation of them optional or based on the config. | 16:30 |
| *** jdurgin has joined #openstack-dev | 16:32 | |
| *** bcwaldon has joined #openstack-dev | 16:32 | |
| * nhm works on getting virtualgl working on one of his images. | 16:32 | |
| vishy | dprince: that is a possiblity. We also discussed making a nova-manage command to create the db instead of doing it on api boot, but it means updating all of the install guides everywhere with an additional line | 16:33 |
| *** dragondm has joined #openstack-dev | 16:37 | |
| *** dragondm has quit IRC | 16:51 | |
| *** alex-meade has joined #openstack-dev | 16:54 | |
| sandywalsh | vishy, the problem was when CA moved to nova/CA. --ca_path= needs to be set to point to ./nova/CA when not running the PPA. | 16:55 |
| vishy | ah | 16:56 |
| vishy | i thought --ca_path defaults there? | 16:56 |
| sandywalsh | apparently not. It's $state_path/CA, not sure what $state_path equates to | 16:57 |
| sandywalsh | DEFINE_string('state_path', os.path.join(os.path.dirname(__file__), '../') | 16:58 |
| *** jdurgin has quit IRC | 16:58 | |
| sandywalsh | which is up one too high | 16:58 |
| sandywalsh | vishy, think I should issue a bug for that? | 16:59 |
| vishy | yes | 17:00 |
| vishy | well hold on | 17:00 |
| vishy | no that is right | 17:00 |
| vishy | it creates a CA folder in the topdir | 17:00 |
| vishy | nova/CA holds the templates and such | 17:01 |
| sandywalsh | setup does, you mean? | 17:01 |
| vishy | yes i think so | 17:01 |
| sandywalsh | k, I'll just note it on the wiki page then. Thx | 17:02 |
| vishy | yes, if i run nova-api the CA dir gets created in the top-level dir | 17:02 |
| sandywalsh | hmm, the order of things seems a little quirky (do all the nova-manage stuff, run api, run more nova-manage stuff) | 17:05 |
| vishy | sandywalsh: yes it is a little odd | 17:07 |
| vishy | sandywalsh: thats why creating the CA db "should" be a manage command | 17:07 |
| vishy | sandywalsh: so setting the --ca_path to nova/CA actually works? I would think that would lead to odd issues... | 17:08 |
| nhm | hrm... looks like KVM still isn't capable of pci-passthrough to get at GPUs. | 17:08 |
| pyhole | Unable to fetch Redmine data | 17:08 |
| sandywalsh | vishy, I'm still setting up for testing ... will know shortly. Stay tuned | 17:09 |
| nhm | guess I'll have to look at Xen. | 17:09 |
| vishy | nhm: really? that is too bad | 17:10 |
| nhm | vishy: sure doesn't look that way. The info on the kvm site is from like a year ago when someone was looking at how XEN does it. | 17:11 |
| nhm | vishy: one of the things I've been thinking about doing with openstack is to virtualize GPU resources so people could load OSes with different driver versions for different applications (since CUDA apps appear to be incredibly driver dependant). | 17:14 |
| jk0 | 7/q pyhole | 17:14 |
| vishy | nhm: i thought kvm allowed you to pin the gpu to one vm... | 17:20 |
| *** dragondm has joined #openstack-dev | 17:20 | |
| nhm | vishy: if it does that would be great | 17:20 |
| vishy | nhm: looks like i may be on crack | 17:21 |
| nhm | vishy: join the club. ;) | 17:22 |
| vishy | nhm: lets go write it! | 17:22 |
| nhm | vishy: looks like it's primarily a problem with passing the video bios through. A couple of people over the last year have claimed that they've been working on it.. | 17:27 |
| * vishy wishes I had time to mess with that. I had so much fun writing assembly pixel and vertex shaders for the first generation of programmable gpus | 17:46 | |
| * vishy confuses his first and third person sometimes. | 17:46 | |
| nhm | hehehe | 17:46 |
| nhm | We're going to buy a small GPU cluster here soon. | 17:47 |
| nhm | I wonder how hard it would be to hack up the flavors stuff to be able to request GPU resources along with other resources. | 17:48 |
| vishy | nhm: there is a group working on that specifically | 17:49 |
| nhm | excellent | 17:49 |
| vishy | https://blueprints.launchpad.net/nova/+spec/heterogeneous-gpu-accelerator-support | 17:49 |
| vishy | i think their current version uses flavors/instance_types | 17:50 |
| vishy | but they are waiting on distributed scheduler to land to switch it to using capabilities | 17:50 |
| nhm | If the grant I just helped write comes through we'll have $200k to bootstrap a production openstack cloud here. | 17:50 |
| vishy | nhm: look there is apparently a way to make it work with kvm even | 17:50 |
| nhm | I'm thinking about throwing some GPUs in to make it interesting. | 17:50 |
| vishy | gvirtus...interesting | 17:51 |
| nhm | very interesting! | 17:51 |
| *** sirp has joined #openstack-dev | 17:59 | |
| *** johnpur has quit IRC | 18:02 | |
| *** dragondm has quit IRC | 18:36 | |
| *** dragondm has joined #openstack-dev | 18:44 | |
| *** fyang has left #openstack-dev | 18:48 | |
| *** zaitcev has joined #openstack-dev | 19:11 | |
| *** zaitcev has quit IRC | 19:23 | |
| *** dprince has quit IRC | 19:43 | |
| vishy | are we logging this channel anywhere? | 19:53 |
| *** vishy has left #openstack-dev | 19:53 | |
| *** vishy has joined #openstack-dev | 19:54 | |
| *** zaitcev has joined #openstack-dev | 19:54 | |
| jk0 | pyhole logs, but they're not public | 19:59 |
| jk0 | they *could* be, though | 19:59 |
| vishy | i'm just thinking whatever we have logging in #openstack should be done here | 20:00 |
| bcwaldon | vishy: your input would be greatly appreciated on https://code.launchpad.net/~jshepher/nova/lp781429/+merge/61140 | 20:17 |
| bcwaldon | vishy: thanks, bud | 20:24 |
| soren | Swift question: i'm trying to consolidate the packaging code for Swift. The only functional difference I can find between the two is that one puts swift-account-audit into the swift package. I don't know if doing so is the right thing to do or if not doing so is the right thing to do. Can you help me out? | 20:24 |
| notmyname | gholt: ^? | 20:30 |
| soren | Actually, that may be a red herring. | 20:32 |
| soren | Nope, it's accurate. | 20:35 |
| *** bcwaldon has quit IRC | 20:37 | |
| vishy | soren: any idea who set up the bot in openstack? | 20:48 |
| vishy | soren: it would be great to get jenkins notifications moved here | 20:48 |
| vishy | soren: and logging setup for this channel | 20:48 |
| soren | vishy: antonym, I think. | 20:49 |
| soren | vishy: Err.. | 20:49 |
| soren | vishy: No, wait. | 20:49 |
| soren | vishy: The Jenkins announcements? I can fix those. | 20:49 |
| soren | vishy: I though you were referring to the logging thing. | 20:49 |
| vishy | ah cool... | 20:49 |
| vishy | i'll ping ant about the logging | 20:50 |
| soren | vishy: ...which I whink was antonym. | 20:50 |
| soren | Right. | 20:50 |
| soren | Should we announce things in both places, or juts here? | 20:50 |
| soren | just, even. | 20:50 |
| * soren doesn't care | 20:50 | |
| antonym | http://eavesdrop.openstack.org/irclogs/ better? | 20:56 |
| antonym | was just pointing to #openstack before | 20:56 |
| * soren clarifies: | 20:56 | |
| soren | vishy: Do you think notifications from Jenkins should go to both channels or only to this one? | 20:56 |
| vishy | only this one | 20:57 |
| vishy | they seem only useful to devs | 20:57 |
| *** openstackjenkins has joined #openstack-dev | 20:58 | |
| soren | vishy: Roger that. | 20:59 |
| soren | vishy: Should be in effect now. | 20:59 |
| notmyname | yay. thanks soren | 21:02 |
| openstackjenkins | Project nova build #904: SUCCESS in 2 min 47 sec: http://jenkins.openstack.org/job/nova/904/ | 21:03 |
| openstackjenkins | Tarmac: Migrate quota schema from hardcoded columns to a key-value approach. The hope is that this change would make it easier to change the quota system without future schema changes. It also adds the concept of quotas that are unlimited. | 21:03 |
| soren | \o/ | 21:03 |
| *** mszilagyi has joined #openstack-dev | 21:04 | |
| *** alex-meade has quit IRC | 21:08 | |
| *** ziyadb has joined #openstack-dev | 21:47 | |
| comstud | hm | 22:04 |
| comstud | vishy ? | 22:05 |
| comstud | or anyone that wants to take a quick peak at nova/service.py | 22:06 |
| vishy | comstud? | 22:06 |
| vishy | what is up? | 22:06 |
| comstud | curious if anyone can explain why the carrot consumers are instantiated inside this 'if' on line 97 | 22:06 |
| comstud | this looks like a bug to me | 22:06 |
| comstud | http://paste.openstack.org/show/1349/ | 22:07 |
| comstud | the diff there is what i'd assume it should be | 22:07 |
| comstud | but i didn't know if 'report_interval' being 0 was supposed to disable topic listeners | 22:07 |
| comstud | so I want to check before I submit this as a bug/fix | 22:07 |
| *** ziyadb has quit IRC | 22:08 | |
| comstud | it's minor, but spotted it while thinking about reworking some things | 22:09 |
| gholt | soren: notmyname: swift-account-audit is a utility for an operator to audit an account for problems. swift-account-auditor is a daemon to be run always on an account server, looking for any problems. swift-account-audit should be in the swift package alongside other utilities like st. swift-account-auditor should be in the swift-account package alongside the swift-account-server. | 22:10 |
| *** ziyadb has joined #openstack-dev | 22:13 | |
| vishy | hmm | 22:14 |
| vishy | interesting | 22:14 |
| vishy | does anything have report_interval not set? | 22:15 |
| vishy | maybe that code stops nova-api from listening? | 22:15 |
| vishy | comstud: ^^ | 22:15 |
| comstud | it would if it were set to 0 | 22:15 |
| comstud | i'm not running into this as an issue | 22:15 |
| comstud | but it looks to be an issue if someone were to set report-internval to 0 | 22:16 |
| vishy | although it does seem odd that would be based on report_interval | 22:16 |
| vishy | I'm guessing bug... | 22:16 |
| comstud | yeah, i think it's a bug. but i wanted to check if anyone knew if this was on purpose | 22:16 |
| comstud | i'd think if it were on purpose.. | 22:16 |
| comstud | 'conn1,2,3 = ' | 22:16 |
| comstud | would have also been inside the 'if' | 22:16 |
| comstud | since they would otherwise not be used | 22:17 |
| comstud | I'll open a bug and get a merge prop in for it | 22:17 |
| comstud | i started looking at this because.. i think we're unnecessarily opening too many connections to rabbit | 22:18 |
| comstud | 200 compute node cluster results in 600+ connections to rabbit | 22:19 |
| comstud | all looping and doing a send/recv every .1 sec | 22:20 |
| comstud | (even if there are no events) | 22:21 |
| comstud | that may be not be a real issue.. i'd probably leave it this way and get rid of this .1 second poll and replace it with consumer.wait() and figure out a way to still reconnect on errors | 22:22 |
| comstud | real problem i have is the 50+ builds at once which results in at least 3 connects and teardowns to rabbit.. each. | 22:23 |
| comstud | in API | 22:23 |
| comstud | which is resulting in some 'Socket closed' errors somehow. | 22:23 |
| comstud | so I'm thinking about a connection pool for those | 22:24 |
| comstud | should result in less work for eventlet polling, less context switches, and less load on rabbit. | 22:25 |
| *** ziyadb has left #openstack-dev | 22:32 | |
| *** mattray has quit IRC | 22:42 | |
| comstud | ah. | 22:43 |
| comstud | well, apparently this stuff is inside the 'if' for tests. | 22:43 |
| *** jdurgin has joined #openstack-dev | 22:58 | |
| *** jkoelker has quit IRC | 23:01 | |
| jk0 | heya, can I get one more set of eyes on https://code.launchpad.net/~jk0/nova/lp760342/+merge/61181 please? only two lines | 23:01 |
| *** rnirmal has quit IRC | 23:06 | |
| *** zaitcev has quit IRC | 23:15 | |
| *** cp16net has quit IRC | 23:16 | |
| *** zaitcev has joined #openstack-dev | 23:17 | |
| *** markvoelker has quit IRC | 23:43 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!