Thursday, 2018-02-15

*** k0da has quit IRC00:10
*** k0da has joined #openstack-powervm00:25
*** k0da has quit IRC00:36
*** svenkat has joined #openstack-powervm01:44
*** svenkat has quit IRC02:56
*** edmondsw has joined #openstack-powervm07:41
*** edmondsw has quit IRC07:45
*** k0da has joined #openstack-powervm07:56
*** AlexeyAbashkin has joined #openstack-powervm08:08
*** edmondsw has joined #openstack-powervm09:29
*** edmondsw has quit IRC09:34
*** k0da has quit IRC09:53
*** k0da has joined #openstack-powervm10:05
*** k0da has quit IRC10:45
*** svenkat has joined #openstack-powervm12:52
*** svenkat has quit IRC12:59
*** svenkat has joined #openstack-powervm12:59
*** edmondsw has joined #openstack-powervm13:06
*** edmondsw has quit IRC13:10
openstackgerritSridhar Venkat proposed openstack/nova-powervm master: Support maximum capacity for SR-IOV VNIC  https://review.openstack.org/54138713:15
*** edmondsw has joined #openstack-powervm13:24
*** apearson has joined #openstack-powervm13:54
*** esberglu has joined #openstack-powervm13:58
*** apearson has quit IRC14:20
*** apearson has joined #openstack-powervm14:22
efriededmondsw: Nice response from ttx there.  Home run, well done.14:24
*** apearson has quit IRC14:26
edmondswefried tx14:26
efriededmondsw: Also thanks for addressing that debian folder business. I was totally clueless.14:27
edmondswefried we'll see how that goes... I wouldn't call it addressed yet :)14:27
edmondswflipping over to slack so we can pull Julio into that convo14:27
efriedOkay, thanks for even knowing how to respond at all.14:27
*** k0da has joined #openstack-powervm14:28
*** esberglu has quit IRC14:34
*** apearson has joined #openstack-powervm14:38
*** apearson has quit IRC14:45
*** apearson has joined #openstack-powervm14:46
*** tjakobs_ has joined #openstack-powervm15:23
edmondswefried does u-c have to update pypowervm to 1.1.11 before we can propose a version bump in g-r to the same?15:53
edmondswI would think so...15:53
edmondswbut u-c just updates on its own, so we have to wait for that, or should I propose that first, and then propose g-r update?15:53
efriedWe're supposed to wait for the bot to propose the u-c bump, then propose the g-r bump on top of that.15:54
efriedThough I have been told (don't remember by whom, but somebody important) that we don't *have* to wait for the bot to propose the former.15:54
efriedbut they are supposed to be separate patches.15:54
edmondswwill have to update g-r before we can merge the svenkat's nova-powervm change15:55
edmondswsvenkat you ok waiting a few days for that?15:56
svenkatedmondsw: yes… I am fine, it is not needed immediately.15:56
* edmondsw has promoted svenkat to "the svenkat" ;)15:56
edmondswcool, will wait then15:56
svenkat!!15:56
openstacksvenkat: Error: "!" is not a valid command.15:56
*** AlexeyAbashkin has quit IRC16:37
*** k0da has quit IRC16:54
openstackgerritTaylor Jakobson proposed openstack/nova-powervm master: Refactor volume drivers  https://review.openstack.org/53081617:29
*** AlexeyAbashkin has joined #openstack-powervm18:01
*** AlexeyAbashkin has quit IRC18:06
edmondswefried svenkat I was chatting with the requirements folks... u-c changes are still locked down even though they've branched stable/queens19:06
svenkatok…19:06
edmondswthis is so that cycle-trailing projects (which don't branch until after release, within 2 weeks) aren't seeing rocky requirements19:07
edmondswso we'll have to wait a couple weeks for them to open that back up19:07
svenkatok…19:07
edmondswthey say osa has already branched, so just puppet and tripleo left to branch and then they can open it up19:08
edmondswthe 2 weeks after release is an "at the latest" figure19:09
edmondswefried do you know if we support cold migration in the OOT driver?19:37
efriedremind me what "cold migration" really means?19:37
edmondswefried migration while the VM is powered down19:38
efriedRight, but how does that translate to nova-isms?19:38
edmondswI'm not sure19:38
edmondswmdrabe do you know?19:39
edmondswI know we don't support cold migrate in PowerVC, but I'm not sure about the driver itself19:39
mdrabeIt's more or less a rebuild on a new host I always thought19:39
efriededmondsw: From really old notes I have, it's implemented through migrate_disk_and_power_off.  Which I believe we implement.19:40
edmondswefried hmmm... I see that used in the resize flow19:41
edmondswwhich is definitely live19:41
efriedI'm not sure actually.19:41
efriedCold migrate might be "power off and resize to another host"19:41
*** k0da has joined #openstack-powervm19:41
*** k0da has quit IRC20:08
*** AlexeyAbashkin has joined #openstack-powervm20:15
*** AlexeyAbashkin has quit IRC20:19
*** k0da has joined #openstack-powervm20:26
*** k0da has quit IRC20:31
edmondswefried per thorst we do support cold migration out of tree20:36
efriedcoo20:37
edmondswefried I proposed the bp for rocky: https://review.openstack.org/#/c/545111/20:37
edmondswtalking with thorst, I kept it more general as to whether we'd do resize vs/ rebuild and cold vs. live migration20:37
edmondswso we have some flexibility there20:37
edmondswhe is also thinking that we should not bother with the slot stuff as we move this IT20:38
edmondswsince that is an unsolved problem in nova in general (boo)20:38
edmondswso skin it more globally later20:38
efriededmondsw: Gonna put this comment here rather than in the spec:20:40
efriedYou'll get easier traction if you resist the temptation to fix spelling, grammar, sentence structure, and other things not directly related to the content.20:41
efriedi.e. smallest viable diff (which shall henceforth be abbreviated SVD)20:41
edmondswyeah... it's just such a strong urge... ;)20:42
efriedoh, I know.20:42
edmondswI didn't think the little I did of that here made much difference... at least I hope not20:42
efriedno, but you want mriedem to say "here's the diff, it's small, fast approving"20:43
efriedIf the diff doesn't fit on one page...20:43
*** k0da has joined #openstack-powervm20:43
edmondswefried it can't... the use cases, proposal, and workitems have to be different on multiple lines because we were so specific in the last BP20:43
edmondswplus you have to replace Queens with Rocky everywhere...20:44
efriededmondsw: Do we really want localdisk?  Why?20:45
edmondswefried because ssp sucks?20:45
efriedwhoah, whoah, what?20:45
efriedSSP is awesome20:45
edmondswnot if you talk to customers20:45
efriedVIOS VG is better??20:45
efriedYou got a way to thin clone a VG?20:46
efriedor do you like your deploys taking minutes instead of seconds?20:46
edmondswI don't know either way... I just know a lot of customers hate SSP20:46
efriedIt's a minor PITA to set up, but once it's running...20:46
efriedAnyway, what customers are we concerned about using in-tree?20:46
edmondsw(if you can keep it running)20:46
efriedJust sayin, if we're talking priorities, I would put that pretty low.20:46
edmondswI asked thorst about it and he thought it would be a good one, so I put it in20:47
edmondswit's also easier to use with PoCs, right?20:48
efriedI suppose.20:51
edmondswefried at the last PTG I threw something on the agenda to walk through what we did in Pike and what we planned to do in Queens.20:55
edmondswBut I won't be at this PTG so can't do something similar.20:55
edmondswcan you cover that, or think it's not necessary?20:55
efriedNo opinion.20:56
*** k0da has quit IRC21:16
*** k0da has joined #openstack-powervm21:32
*** openstack has joined #openstack-powervm21:45
*** ChanServ sets mode: +o openstack21:45
*** svenkat has quit IRC21:56
*** tjakobs_ has quit IRC22:23
*** svenkat has joined #openstack-powervm22:29
*** tjakobs_ has joined #openstack-powervm22:47
efriededmondsw: I'm responding.22:51
efriedWell, at least to the first one.22:52
*** tjakobs_ has quit IRC22:59
*** apearson has quit IRC23:17
*** svenkat has quit IRC23:29
*** tjakobs_ has joined #openstack-powervm23:46

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!