| *** edmondsw has quit IRC | 00:03 | |
| *** edmondsw has joined #openstack-powervm | 00:05 | |
| *** edmondsw has quit IRC | 00:08 | |
| *** edmondsw has joined #openstack-powervm | 00:09 | |
| *** edmondsw has quit IRC | 00:13 | |
| *** thorst_afk has joined #openstack-powervm | 00:39 | |
| *** thorst_afk has quit IRC | 00:44 | |
| *** thorst_afk has joined #openstack-powervm | 00:47 | |
| *** thorst_afk has quit IRC | 00:47 | |
| *** thorst_afk has joined #openstack-powervm | 00:49 | |
| *** thorst_afk has quit IRC | 00:49 | |
| *** thorst_afk has joined #openstack-powervm | 01:54 | |
| *** thorst_afk has quit IRC | 01:59 | |
| *** thorst_afk has joined #openstack-powervm | 02:05 | |
| *** thorst_afk has quit IRC | 02:05 | |
| *** esberglu has joined #openstack-powervm | 02:20 | |
| *** esberglu has quit IRC | 02:20 | |
| *** thorst_afk has joined #openstack-powervm | 02:36 | |
| *** thorst_afk has quit IRC | 02:41 | |
| *** esberglu has joined #openstack-powervm | 03:14 | |
| *** esberglu has quit IRC | 03:18 | |
| *** thorst_afk has joined #openstack-powervm | 03:38 | |
| *** thorst_afk has quit IRC | 03:42 | |
| *** chhavi has joined #openstack-powervm | 04:11 | |
| *** thorst_afk has joined #openstack-powervm | 04:38 | |
| *** thorst_afk has quit IRC | 04:43 | |
| *** thorst_afk has joined #openstack-powervm | 05:39 | |
| *** thorst_afk has quit IRC | 05:44 | |
| *** esberglu has joined #openstack-powervm | 05:59 | |
| *** edmondsw has joined #openstack-powervm | 06:02 | |
| *** esberglu has quit IRC | 06:03 | |
| *** edmondsw has quit IRC | 06:06 | |
| *** thorst_afk has joined #openstack-powervm | 06:40 | |
| *** thorst_afk has quit IRC | 06:45 | |
| *** esberglu has joined #openstack-powervm | 06:52 | |
| *** esberglu has quit IRC | 06:53 | |
| *** esberglu has joined #openstack-powervm | 06:53 | |
| *** esberglu has quit IRC | 06:53 | |
| *** esberglu has joined #openstack-powervm | 06:54 | |
| *** esberglu has quit IRC | 06:58 | |
| *** k0da has joined #openstack-powervm | 07:36 | |
| *** thorst_afk has joined #openstack-powervm | 07:41 | |
| *** thorst_afk has quit IRC | 07:45 | |
| *** k0da has quit IRC | 08:09 | |
| *** thorst_afk has joined #openstack-powervm | 08:42 | |
| *** thorst_afk has quit IRC | 08:46 | |
| *** thorst_afk has joined #openstack-powervm | 09:42 | |
| *** thorst_afk has quit IRC | 09:47 | |
| *** chhavi has quit IRC | 10:14 | |
| *** k0da has joined #openstack-powervm | 10:31 | |
| *** thorst_afk has joined #openstack-powervm | 10:43 | |
| *** thorst_afk has quit IRC | 10:48 | |
| *** thorst_afk has joined #openstack-powervm | 11:09 | |
| *** thorst_afk has quit IRC | 11:11 | |
| *** esberglu has joined #openstack-powervm | 11:26 | |
| *** esberglu has quit IRC | 11:26 | |
| *** thorst_afk has joined #openstack-powervm | 12:04 | |
| *** esberglu has joined #openstack-powervm | 12:19 | |
| *** smatzek has joined #openstack-powervm | 12:22 | |
| *** esberglu has quit IRC | 12:23 | |
| *** apearson has joined #openstack-powervm | 12:44 | |
| *** esberglu has joined #openstack-powervm | 13:07 | |
| *** esberglu has quit IRC | 13:07 | |
| *** efried is now known as fried_rice | 13:20 | |
| *** esberglu has joined #openstack-powervm | 13:36 | |
| *** esberglu has quit IRC | 13:37 | |
| *** esberglu has joined #openstack-powervm | 13:39 | |
| fried_rice | esberglu We still running with that tempest patch to ignore connection timeouts? | 13:52 |
|---|---|---|
| esberglu | fried_rice: Yeah | 13:52 |
| fried_rice | mmkay. | 13:52 |
| esberglu | It only ever seems to hit it once at a time, haven't had a chance to look further | 13:53 |
| fried_rice | Doesn't seem like something that should be our problem. | 13:53 |
| fried_rice | I don't really have a problem running with that patch more or less permanently. | 13:53 |
| *** miltonm has joined #openstack-powervm | 14:05 | |
| esberglu | fried_rice: Yeah that was my thinking as well. No hurry to fix it anyways | 14:06 |
| *** k0da has quit IRC | 14:08 | |
| *** edmondsw has joined #openstack-powervm | 14:17 | |
| esberglu | https://blueprints.launchpad.net/nova/+spec/powervm-nova-it-compute-driver | 14:35 |
| esberglu | fried_rice: edmondsw: ^ | 14:36 |
| edmondsw | esberglu awesome... I was going to do that this week | 14:36 |
| edmondsw | you guys keep beating me to things :) | 14:36 |
| esberglu | edmondsw: Now we just have to push a spec to nova-specs right? | 14:37 |
| fried_rice | esberglu Yes | 14:37 |
| edmondsw | yep | 14:37 |
| fried_rice | You can start off by copying the one from pike | 14:37 |
| fried_rice | https://github.com/openstack/nova-specs/blob/master/specs/pike/approved/powervm-nova-compute-driver.rst | 14:37 |
| fried_rice | However, it may be worth correlating with the latest queens template | 14:38 |
| edmondsw | yep, and then we'll need to modify it a bit | 14:38 |
| edmondsw | ++ | 14:38 |
| fried_rice | https://github.com/openstack/nova-specs/blob/master/specs/queens-template.rst | 14:38 |
| edmondsw | I think they'll require that | 14:38 |
| fried_rice | Also note that there's a pending update to that template which you should also incorporate. | 14:38 |
| fried_rice | https://review.openstack.org/#/c/456756/ | 14:39 |
| fried_rice | Not that I think we should have any upgrade impacts. | 14:39 |
| fried_rice | But include the section anyway. | 14:39 |
| edmondsw | esberglu I can do the spec unless you want to | 14:39 |
| esberglu | fried_rice: edmondsw: Have we determined what functionality we are targeting for queens? | 14:40 |
| esberglu | I know we have talked about it, not sure if we had a hard list yet though | 14:41 |
| esberglu | edmondsw: I would like to do it | 14:41 |
| fried_rice | At least config drive and OVS. thorst_afk should chime in on that. | 14:41 |
| edmondsw | yes... PCI passthrough, network (at least SEA, though I'd like OVS as well), config drive, vSCSI | 14:41 |
| fried_rice | PCI passthrough? In tree? | 14:41 |
| thorst_afk | we need enough to deploy a real VM that someone can put workload on | 14:41 |
| thorst_afk | network (just do both), config drive -> must do's | 14:42 |
| thorst_afk | vscsi would be great for a cinder story | 14:42 |
| edmondsw | fried_rice we could do pci just out of tree, but I thought it would be nice to do in both unless there's a reason not to | 14:42 |
| fried_rice | Yeah, I guess there's no reason not to. | 14:42 |
| thorst_afk | PCI passthru would be nice in tree, needed out of tree? | 14:42 |
| thorst_afk | fried_rice edmondsw: reviewer bandwidth | 14:42 |
| thorst_afk | if there is enough bandwidth for all... | 14:42 |
| edmondsw | right, I view the reviewer bandwidth as a reason to do pci IT... we want them involved in that | 14:43 |
| fried_rice | As I dig further into the generic resource provider stuff, it's becoming less probable that we'll be able to do our PCI passthrough that way in q - so we'll need our hacked version probably until r. | 14:43 |
| edmondsw | do you think we'll get objections to that hacked implementation going into q IT? | 14:44 |
| fried_rice | Right, that was my concern. | 14:44 |
| fried_rice | Not necessarily so much objections as no reviewer's gonna want to touch it with a ten-foot pole. | 14:44 |
| edmondsw | well, we can always pull it out later... but easier to put it in the plan for now, and then see how that goes? | 14:44 |
| edmondsw | at least see how the PTG conversations go... could be pulled quickly :) | 14:45 |
| fried_rice | e.g. see Jay's comments on https://review.openstack.org/#/c/496434/ | 14:45 |
| fried_rice | I'm not going to get any traction talking about PowerVM's hacked PCI passthrough implementation at the PTG. | 14:45 |
| fried_rice | Cause it has nothing to do with nova. | 14:45 |
| edmondsw | fried_rice I thought we'd be talking about that hacked up way of doing things as a way to elucidate the issues driving the work that will go into r | 14:46 |
| fried_rice | edmondsw Meh, that's sorta done. | 14:47 |
| edmondsw | cool | 14:47 |
| edmondsw | I did see Jay's comments but haven't had a chance to read them yet... I need to do that today | 14:47 |
| fried_rice | I mean, yeah, it's something I'll talk about, but I don't think it's necessary to have a change set to put in front of everyone. | 14:47 |
| edmondsw | k | 14:48 |
| edmondsw | so you think leave that OOT? | 14:48 |
| edmondsw | I don't think there's a reason we have to have it IT... it was more to help us get things done | 14:49 |
| edmondsw | if it doesn't do that... | 14:49 |
| fried_rice | edmondsw I don't mind porting it in tree, but we should finish it up OOT first. And it has served its purpose as a driver of the in-tree device management revolution. | 14:50 |
| edmondsw | sure | 14:50 |
| edmondsw | esberglu so for the moment anyway don't list PCI passthrough in the spec | 14:50 |
| esberglu | edmondsw: k | 14:51 |
| openstackgerrit | Eric Fried proposed openstack/nova-powervm master: snapshot capability in DiskAdapter (local: False) https://review.openstack.org/466425 | 15:04 |
| openstackgerrit | Eric Fried proposed openstack/nova-powervm master: SSP supports nonstandard image caching https://review.openstack.org/500099 | 15:19 |
| *** tjakobs has joined #openstack-powervm | 15:23 | |
| fried_rice | thorst_afk Is snapshot supported in the localdisk driver? According to the capability bits, it ain't; but according to tjakobs, it works. | 15:23 |
| thorst_afk | I think tjakobs is looking at it with the lens of SDE | 15:24 |
| thorst_afk | and not necessarily with the lens of traditional VIOS? | 15:24 |
| fried_rice | Ah, so supported with SDE but not traditional. | 15:24 |
| fried_rice | thorst_afk Looking at https://bugs.launchpad.net/nova-powervm/+bug/1661395 and https://review.openstack.org/#/c/466425/ btw | 15:25 |
| openstack | Launchpad bug 1661395 in nova-powervm "Snapshot with localdisk not supported" [Low,In progress] - Assigned to Eric Fried (efried) | 15:25 |
| thorst_afk | we should bounce that off seroyer. | 15:25 |
| tjakobs | about a year ago, i put up a fix for supporting SDE, but i had also verified i didn't break anything with an traditional setup https://review.openstack.org/#/c/368157 | 15:25 |
| fried_rice | If that's the case, we may want to detect SDEness and make that capability bit dynamic accordingly. | 15:25 |
| thorst_afk | I think the challenge was that our 'SDE' check was just 'is_mgmt_partition' | 15:25 |
| thorst_afk | rather than something in the VIOS bits itself. | 15:25 |
| thorst_afk | we want to get to capabilities | 15:26 |
| thorst_afk | what is the 'VIOS' capable of. Not, do I happen to be the mgmt partition, so I know its different. | 15:26 |
| thorst_afk | I think that's my main gripe. | 15:27 |
| thorst_afk | we can do it...sure. I just would prefer we do the capabilities approach. | 15:27 |
| fried_rice | So this change https://review.openstack.org/#/c/466425/ will actually regress SDE. | 15:28 |
| fried_rice | So we shouldn't do it. | 15:28 |
| fried_rice | until we have a better answer for that capability bit. | 15:28 |
| thorst_afk | maybe something to ask Hsien about? | 15:29 |
| thorst_afk | tjakobs change looks pretty reasonable IMO, unless we can get a capability bit. | 15:29 |
| thorst_afk | or I wonder if we could hide capability bits in pypowervm until we get it in the REST API. That seems kinda silly tho | 15:30 |
| fried_rice | thorst_afk tjakobs change makes it look like it may be possible to have a nvl-as-vios and *also* traditional VIOSes. | 15:31 |
| fried_rice | And a given VM's disk might be mapped from either one. | 15:31 |
| fried_rice | So we can't say at the outset whether snapshot is supported or not. | 15:31 |
| fried_rice | cause it might be different for different VMs. | 15:31 |
| thorst_afk | but line 246 forces it to just be mgmt | 15:32 |
| thorst_afk | in storage.py | 15:32 |
| thorst_afk | I don't think it says that for a given host, you can have from either | 15:32 |
| fried_rice | only if we happen to find it on the mgmt | 15:32 |
| thorst_afk | its just saying if your host is configured to host local disks through mgmt, you can | 15:32 |
| thorst_afk | and if you're configured for traditional you can't | 15:32 |
| fried_rice | If L250 doesn't pass, we move on to trying to find it on the VIOS. | 15:32 |
| thorst_afk | albeit the code is in the mainline execution path...but I think that's the realization | 15:33 |
| fried_rice | mm, actually, localdisk driver only has one VIOS, nah? | 15:33 |
| fried_rice | yup. | 15:33 |
| thorst_afk | localdisk can only be hosted by one VIOS | 15:33 |
| fried_rice | Okay, so that's solid. | 15:33 |
| thorst_afk | even if there are many VIOSes on the system | 15:33 |
| fried_rice | I can fix that up. | 15:33 |
| thorst_afk | yeah, I think his change (without diving into it deeply) looks like a reasonable path until capability bits | 15:33 |
| thorst_afk | but tjakobs we should instead be pushing on Hsien to get these capabilities expressed in the API | 15:34 |
| fried_rice | tjakobs Under what circumstances does L250 fail, but the rest still succeeds? | 15:34 |
| thorst_afk | (this fix looks nice though because if you are hosting through the mgmt, it looks like it stops trying to map to the mgmt...its a bug fix from that point of view) | 15:34 |
| fried_rice | Is there, like, a disk type where we wouldn't find it via boot_disk_path_for_instance, but we would find it and be able to attach it via connect_instance_disk_to_mgmt?? | 15:35 |
| fried_rice | thorst_afk We're not debating tjakobs change set - it merged almost a year ago :) | 15:35 |
| thorst_afk | lol | 15:36 |
| thorst_afk | whoops | 15:36 |
| tjakobs | efried I don't think so... Trying to remember | 15:36 |
| tjakobs | yeah, thorst_afk you already +2'd it | 15:36 |
| fried_rice | thorst_afk we're trying to figure out what to do about https://review.openstack.org/#/c/466425/ | 15:36 |
| fried_rice | I think I have a fix that we'll be happy with. Sec... | 15:36 |
| thorst_afk | heh | 15:43 |
| thorst_afk | I'm so up to date | 15:43 |
| *** k0da has joined #openstack-powervm | 15:44 | |
| openstackgerrit | Eric Fried proposed openstack/nova-powervm master: SSP supports nonstandard image caching https://review.openstack.org/500099 | 15:52 |
| openstackgerrit | Eric Fried proposed openstack/nova-powervm master: Add snapshot capability in DiskAdapter https://review.openstack.org/466425 | 15:52 |
| fried_rice | tjakobs thorst_afk ^^ | 15:52 |
| fried_rice | thorst_afk Second pair of eyes on 5828 please? | 15:55 |
| fried_rice | wait what the actual f just happened? Merge conflict?? | 15:56 |
| *** apearson has quit IRC | 15:58 | |
| openstackgerrit | Eric Fried proposed openstack/nova-powervm master: Add snapshot capability in DiskAdapter https://review.openstack.org/466425 | 16:01 |
| fried_rice | that was effin bizarre. | 16:01 |
| openstackgerrit | Eric Fried proposed openstack/nova-powervm master: SSP supports nonstandard image caching https://review.openstack.org/500099 | 16:02 |
| *** miltonm has quit IRC | 16:07 | |
| *** miltonm has joined #openstack-powervm | 16:08 | |
| *** esberglu has quit IRC | 16:24 | |
| *** apearson has joined #openstack-powervm | 16:43 | |
| *** esberglu has joined #openstack-powervm | 16:47 | |
| *** edmondsw has quit IRC | 16:50 | |
| *** esberglu has quit IRC | 16:51 | |
| *** edmondsw has joined #openstack-powervm | 16:51 | |
| *** esberglu has joined #openstack-powervm | 17:20 | |
| fried_rice | esberglu Any idea what's going on here https://review.openstack.org/#/c/466425/ https://review.openstack.org/#/c/500099/ ? | 18:05 |
| esberglu | fried_rice: I updated pvm-novalink and there are Internal Server Error 500s all over the place | 18:05 |
| fried_rice | logs are getting cut off | 18:05 |
| fried_rice | so I can't see what's actually happening. | 18:06 |
| esberglu | http://184.172.12.213/18/499718/2/check/nova-out-of-tree-pvm/14c00f7/logs/n-cpu.txt.gz | 18:07 |
| esberglu | fried_rice: Looks like this is the error | 18:11 |
| esberglu | Error 500: javax.servlet.ServletException: java.lang.NoSuchMethodError: com.ibm.hmcl.data.HmclHypervisorInfo.setAixCapable.(Z)V (hmcljni/hmcljni_common.cpp:489) | 18:11 |
| fried_rice | whoah | 18:11 |
| fried_rice | That's a sev-1 to changh | 18:12 |
| fried_rice | edmondsw 9/25 is way too far after the PTG for me to remember anything that happened there. | 18:48 |
| edmondsw | fried_rice a whole week later? | 18:48 |
| fried_rice | It'd be hard enough for me to do after the weekend. | 18:48 |
| fried_rice | I'm mostly joshing you. But I wouldn't be offended if it was closer to the event. | 18:49 |
| edmondsw | lol... well, I guess you'll be stretched :) | 18:49 |
| edmondsw | fried_rice the first part of the week immediately after doesn't work with everyone slammed closing out a PowerVC sprint on Wednesday, Thursday is already booked, and of course Friday doesn't work for the team in ISTL, so 9/25 was earliest I think we can do | 18:51 |
| *** apearson has quit IRC | 18:59 | |
| *** esberglu has quit IRC | 19:04 | |
| *** esberglu has joined #openstack-powervm | 19:05 | |
| *** apearson has joined #openstack-powervm | 19:06 | |
| *** k0da has quit IRC | 19:09 | |
| *** apearson has quit IRC | 19:18 | |
| *** apearson has joined #openstack-powervm | 19:18 | |
| *** k0da has joined #openstack-powervm | 19:26 | |
| *** smatzek has quit IRC | 20:16 | |
| fried_rice | esberglu You could propose a change to move the pike powervm in-tree blueprint from approved to implemented. | 20:27 |
| fried_rice | or I could. Call it. | 20:27 |
| *** tjakobs has quit IRC | 20:38 | |
| *** thorst_afk has quit IRC | 20:46 | |
| esberglu | fried_rice: Just saw this. I'm about to head. I can do it later this evening when I check on CI | 20:48 |
| fried_rice | esberglu It's cool, I should probably be the one to do it, since (I think) I owned it. | 20:49 |
| esberglu | fried_rice: Okay. I'm out for the day have a good weekend | 20:49 |
| *** esberglu has quit IRC | 20:49 | |
| *** esberglu has joined #openstack-powervm | 20:50 | |
| *** esberglu has quit IRC | 20:54 | |
| *** thorst_afk has joined #openstack-powervm | 21:32 | |
| *** thorst_afk has quit IRC | 21:36 | |
| *** thorst_afk has joined #openstack-powervm | 21:48 | |
| *** thorst_afk has quit IRC | 21:50 | |
| *** apearson has quit IRC | 22:06 | |
| *** thorst_afk has joined #openstack-powervm | 22:29 | |
| *** thorst_afk has quit IRC | 22:32 | |
| *** miltonm has quit IRC | 23:18 | |
| *** k0da has quit IRC | 23:47 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!