| *** thorst has joined #openstack-powervm | 00:53 | |
| *** thorst has quit IRC | 00:54 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 00:55 | |
| *** AlexeyAbashkin has quit IRC | 00:59 | |
| *** jwcroppe has joined #openstack-powervm | 01:32 | |
| *** jwcroppe has quit IRC | 01:54 | |
| *** thorst has joined #openstack-powervm | 02:17 | |
| *** thorst has quit IRC | 02:18 | |
| *** thorst has joined #openstack-powervm | 02:33 | |
| *** thorst has quit IRC | 02:38 | |
| *** jwcroppe has joined #openstack-powervm | 03:57 | |
| *** thorst has joined #openstack-powervm | 04:39 | |
| *** thorst has quit IRC | 04:44 | |
| *** jwcroppe has quit IRC | 05:59 | |
| *** thorst has joined #openstack-powervm | 06:40 | |
| *** thorst has quit IRC | 06:44 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 07:37 | |
| *** thorst has joined #openstack-powervm | 07:41 | |
| *** thorst has quit IRC | 07:45 | |
| *** thorst has joined #openstack-powervm | 09:42 | |
| *** thorst has quit IRC | 09:48 | |
| -openstackstatus- NOTICE: The CI system will be offline starting at 11:00 UTC (in just under an hour) for Zuul v3 rollout: http://lists.openstack.org/pipermail/openstack-dev/2017-October/123337.html | 10:08 | |
| *** AlexeyAbashkin has quit IRC | 11:00 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 11:01 | |
| *** efried has joined #openstack-powervm | 11:32 | |
| *** edmondsw has joined #openstack-powervm | 12:04 | |
| *** thorst has joined #openstack-powervm | 12:06 | |
| *** thorst_ has joined #openstack-powervm | 12:38 | |
| *** esberglu has quit IRC | 12:38 | |
| *** thorst has quit IRC | 12:39 | |
| *** esberglu has joined #openstack-powervm | 12:57 | |
| -openstackstatus- NOTICE: Due to unrelated emergencies, the Zuul v3 rollout has not started yet; stay tuned for further updates | 13:06 | |
| esberglu | Has anyone stacked with OVS lately? Looking for an OVS local.conf to start with, no big deal if not | 13:24 |
|---|---|---|
| *** jwcroppe has joined #openstack-powervm | 14:07 | |
| efried | not I | 14:10 |
| *** tjakobs has joined #openstack-powervm | 15:09 | |
| *** jwcroppe has quit IRC | 15:15 | |
| *** jwcroppe has joined #openstack-powervm | 15:16 | |
| *** kylek3h has joined #openstack-powervm | 15:20 | |
| *** jwcroppe has quit IRC | 15:20 | |
| *** tjakobs has quit IRC | 15:38 | |
| *** tjakobs has joined #openstack-powervm | 15:41 | |
| *** tjakobs has quit IRC | 15:50 | |
| *** AlexeyAbashkin has quit IRC | 15:53 | |
| *** kylek3h has quit IRC | 16:01 | |
| *** kylek3h has joined #openstack-powervm | 16:02 | |
| *** kylek3h has quit IRC | 16:06 | |
| *** kylek3h has joined #openstack-powervm | 16:09 | |
| *** kylek3h has quit IRC | 16:09 | |
| *** kylek3h has joined #openstack-powervm | 16:15 | |
| *** kylek3h has quit IRC | 16:25 | |
| *** kylek3h has joined #openstack-powervm | 16:25 | |
| *** kylek3h has quit IRC | 16:29 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 17:21 | |
| *** AlexeyAbashkin has quit IRC | 17:26 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 17:29 | |
| *** AlexeyAbashkin has quit IRC | 17:34 | |
| *** kylek3h has joined #openstack-powervm | 17:59 | |
| *** jwcroppe has joined #openstack-powervm | 18:10 | |
| *** kylek3h has quit IRC | 18:12 | |
| *** efried is now known as efried_nomnom | 18:17 | |
| *** efried_nomnom is now known as efried | 18:58 | |
| *** jwcroppe has quit IRC | 19:31 | |
| *** edmondsw_ has joined #openstack-powervm | 20:08 | |
| *** edmondsw has quit IRC | 20:08 | |
| *** edmondsw_ is now known as edmondsw | 20:12 | |
| *** thorst_ has quit IRC | 20:55 | |
| *** thorst has joined #openstack-powervm | 20:58 | |
| *** esberglu has quit IRC | 21:03 | |
| *** thorst has quit IRC | 21:03 | |
| *** edmondsw has quit IRC | 21:07 | |
| *** edmondsw has joined #openstack-powervm | 21:14 | |
| *** edmondsw has quit IRC | 21:19 | |
| *** esberglu has joined #openstack-powervm | 21:34 | |
| *** openstackgerrit has joined #openstack-powervm | 21:39 | |
| openstackgerrit | Matt Rabe proposed openstack/nova-powervm master: Persist the max virtio slots number when resizing instances https://review.openstack.org/511343 | 21:39 |
| openstackgerrit | Matt Rabe proposed openstack/nova-powervm master: Persist the max virtio slots number when resizing instances https://review.openstack.org/511343 | 21:39 |
| mdrabe | efried: ^ | 21:39 |
| *** thorst has joined #openstack-powervm | 21:40 | |
| efried | mdrabe Oh, ew, ew, ew. Special-casing that one field? We *must* be able to fix this in pypowervm. | 21:40 |
| mdrabe | It's ultimately nova-powervm's fault though | 21:41 |
| efried | mdrabe Oh, ew, ew, ew, we were *already* special-casing that one field. | 21:41 |
| openstackgerrit | Matt Rabe proposed openstack/nova-powervm master: Persist the max virtio slots number when resizing instances https://review.openstack.org/511343 | 21:42 |
| efried | mdrabe But not that ickily. | 21:42 |
| mdrabe | I stared at it in pypowervm for the past hour | 21:42 |
| efried | mdrabe Okay, I'm gonna need you to convince me that we absolutely can't fix this in pypowervm. | 21:42 |
| efried | Cause I don't think this is nova-powervm's fault. | 21:42 |
| mdrabe | Problem is that pypowervm doesn't have context to know if it's a resize or create | 21:43 |
| efried | I don't think it's unreasonable for nova-powervm to use DefaultStandardize and assume that any already-set fields otherwise not mentioned will be untouched. | 21:43 |
| mdrabe | But the DefaultStandardize object doesn't know if it's gonna be used for creating a wrapper or updating one | 21:44 |
| mdrabe | So you have to tell it "hey use these values" | 21:44 |
| efried | Then we're gonna have this same problem with DEF_PROC_UNIT_FACT = 0.5 | 21:46 |
| efried | DEF_MAX_SLOT = 64 | 21:46 |
| efried | DEF_UNCAPPED_WT = 64 | 21:46 |
| efried | DEF_SPP = 0 | 21:46 |
| efried | DEF_AVAIL_PRI = 127 | 21:46 |
| efried | DEF_SRR = 'false' | 21:46 |
| efried | DEF_LPAR_METRIC = False | 21:46 |
| mdrabe | I'm pretty darn sure we don't have those problems but I'm struggling to figure out why | 21:49 |
| efried | mdrabe Well, I can't imagine a scenario where we would ever want to change those values. | 21:49 |
| mdrabe | Ican | 21:50 |
| efried | So I bet we have the problem, but would never have noticed it. | 21:50 |
| mdrabe | Oh ah okay | 21:51 |
| mdrabe | no I see why it doesn't happen for those parameters | 21:51 |
| mdrabe | god this is confusing | 21:51 |
| efried | proc units factor and uncapped weight we won't change; those are always coming from the conf. | 21:51 |
| efried | I would think we could write a teensy test program to exercise this on a real system. | 21:51 |
| mdrabe | Those other params aren't always defaulted because they come from _format_flavor and are built into that attrs dictionary | 21:51 |
| efried | whee | 21:52 |
| mdrabe | And the values from the attrs dictionary in turn gets set in the DefaultStandardizer by the LPARBuilder | 21:52 |
| mdrabe | https://github.com/powervm/pypowervm/blob/develop/pypowervm/utils/lpar_builder.py#L794 | 21:52 |
| mdrabe | So actually... | 21:53 |
| efried | So if we set max slots in format_flavor instead of init... | 21:54 |
| mdrabe | yea | 21:54 |
| efried | Is max slots something we care to be able to configure via the flavor? | 21:54 |
| efried | Not that we have to | 21:54 |
| mdrabe | Right it's like 1 step short of just exposing it in the flavor | 21:55 |
| efried | it would just be prettier if we didn't have a hardcoded non-flavory thing in _format_flavor | 21:55 |
| mdrabe | I feel like somebody asked me about it before | 21:55 |
| efried | But with appropriate commenting, I'm good with it. | 21:55 |
| mdrabe | I'll just add it as an extra spec in the flavor | 21:56 |
| mdrabe | Give it the full seal | 21:56 |
| efried | But even so... | 21:56 |
| efried | If they omit it from the flavor, it'll have the same behavior. | 21:57 |
| mdrabe | Yep | 21:58 |
| mdrabe | I was assuming there was some magic that populated unchanged attributes into the instance.flavor | 21:59 |
| efried | What happens if you say attrs['max_slots'] = None ? Does it leave the cursed thing alone, or does it actually override it? | 21:59 |
| efried | mdrabe Oh, pretty sure no freakin way. | 21:59 |
| mdrabe | When you say attrs['max_slots'] = None | 22:00 |
| mdrabe | Do you mean, what if the element was missing from the XML entirely? | 22:01 |
| efried | If the element is missing from the flavor extra specs and we include it in the loop there, we'll be not fixing the bug. | 22:02 |
| efried | If we hardcode a line that says attrs['max_slots'] = flavor.get('max_slots', None) - i.e. default it to None (rather than absent) in attrs if unspecified in the flavor... | 22:03 |
| efried | ...then either lpar_builder will leave whatever value is already in the wrapper; or it'll actually "set" it to None, and I don't know what the latter would mean. | 22:03 |
| mdrabe | Yea and now I don't understand how this isn't a problem for those other attributes | 22:05 |
| efried | mdrabe As I said, I bet it is. | 22:06 |
| mdrabe | I think in our case we populate flavors with most of those fields, so we wouldn't see it | 22:06 |
| efried | And we just never tried to change 'em. | 22:06 |
| efried | or that | 22:06 |
| mdrabe | Yea this'll probably need a greater fix, I'll stare at it more tomorrow with fresh eyes | 22:07 |
| *** edmondsw has joined #openstack-powervm | 22:25 | |
| *** edmondsw has quit IRC | 22:30 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 22:47 | |
| *** edmondsw has joined #openstack-powervm | 22:48 | |
| *** AlexeyAbashkin has quit IRC | 22:52 | |
| *** edmondsw has quit IRC | 22:54 | |
| *** thorst has quit IRC | 23:33 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!