| *** edmondsw has joined #openstack-powervm | 00:09 | |
| *** edmondsw has quit IRC | 00:14 | |
| *** apearson has joined #openstack-powervm | 00:23 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 01:50 | |
| *** tonyb has joined #openstack-powervm | 01:50 | |
| *** AlexeyAbashkin has quit IRC | 01:54 | |
| *** edmondsw has joined #openstack-powervm | 01:57 | |
| *** edmondsw has quit IRC | 02:02 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 02:50 | |
| *** AlexeyAbashkin has quit IRC | 02:55 | |
| *** apearson has quit IRC | 03:34 | |
| *** edmondsw has joined #openstack-powervm | 03:45 | |
| *** edmondsw has quit IRC | 03:50 | |
| *** chhagarw has joined #openstack-powervm | 04:07 | |
| *** edmondsw has joined #openstack-powervm | 05:33 | |
| *** edmondsw has quit IRC | 05:38 | |
| *** edmondsw has joined #openstack-powervm | 07:21 | |
| *** edmondsw has quit IRC | 07:26 | |
| *** edmondsw has joined #openstack-powervm | 09:10 | |
| *** edmondsw has quit IRC | 09:14 | |
| *** edmondsw has joined #openstack-powervm | 10:58 | |
| *** edmondsw has quit IRC | 11:03 | |
| *** edmondsw has joined #openstack-powervm | 12:04 | |
| openstackgerrit | Chhavi Agarwal proposed openstack/nova-powervm master: WIP: Having iSCSI Initiator locks per VIOS https://review.openstack.org/557800 | 12:53 |
|---|---|---|
| *** apearson has joined #openstack-powervm | 13:00 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 13:01 | |
| *** AlexeyAbashkin has quit IRC | 13:10 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 13:10 | |
| *** apearson has quit IRC | 13:12 | |
| *** AlexeyAbashkin has quit IRC | 13:17 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 13:17 | |
| *** apearson has joined #openstack-powervm | 13:18 | |
| *** AlexeyAbashkin has quit IRC | 13:34 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 13:37 | |
| *** esberglu has joined #openstack-powervm | 13:39 | |
| *** chhagarw has quit IRC | 13:48 | |
| *** apearson has quit IRC | 13:50 | |
| *** tjakobs has joined #openstack-powervm | 13:51 | |
| *** apearson has joined #openstack-powervm | 13:53 | |
| esberglu | edmondsw: efried: Redeploying CI now since networking issues were resolved | 14:19 |
| edmondsw | ack | 14:19 |
| esberglu | edmondsw: https://review.openstack.org/#/c/554688/ | 14:44 |
| esberglu | Okay if I set max to 1 and min to 0.05 there? | 14:44 |
| edmondsw | max 1 is fine. I'm not sure about min... 0.05 is the current min, but I'd kinda like to leave that up to validation at the lower level | 14:45 |
| edmondsw | min is actually 0.1 on some systems, and 0.05 on others, and who knows if the future will include a system where the min is something else | 14:45 |
| edmondsw | I guess we could set min 0.05 and then change it later if we need to | 14:46 |
| esberglu | edmondsw: Would be consistent with https://github.com/powervm/pypowervm/blob/develop/pypowervm/utils/lpar_builder.py#L211-L214 | 14:46 |
| edmondsw | yeah, ok, go for it | 14:47 |
| edmondsw | if we add min/max, I assume oslo.config does the validation and we don't need the validation that you added anymore? | 14:47 |
| edmondsw | esberglu ^ | 14:48 |
| esberglu | edmondsw: Yep | 14:48 |
| edmondsw | cool | 14:48 |
| *** AlexeyAbashkin has quit IRC | 14:54 | |
| *** tjakobs has quit IRC | 15:36 | |
| *** tjakobs has joined #openstack-powervm | 15:39 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 16:18 | |
| esberglu | edmondsw: New localdisk patch is up and has been live tested | 16:21 |
| openstackgerrit | Chhavi Agarwal proposed openstack/nova-powervm master: WIP: Having iSCSI Initiator locks per VIOS https://review.openstack.org/557800 | 16:21 |
| esberglu | efried: ^^ | 16:22 |
| esberglu | No localdisk snapshot live testing yet though | 16:23 |
| openstackgerrit | Chhavi Agarwal proposed openstack/nova-powervm master: WIP: Having iSCSI Initiator locks per VIOS https://review.openstack.org/557800 | 16:24 |
| efried | esberglu: These in-tree driver patches are in a series - that's on purpose, right? | 16:26 |
| edmondsw | efried yes, he recently changed them to be in a series | 16:32 |
| efried | k | 16:33 |
| edmondsw | I think he's at lunch now | 16:33 |
| -openstackstatus- NOTICE: zuul was restarted to update to the latest code; please recheck any changes uploaded within the past 10 minutes | 16:52 | |
| efried | esberglu, edmondsw: Left a review on localdisk. The rest are +2. | 16:59 |
| edmondsw | efried tx. I need to look at localdisk again myself | 16:59 |
| efried | edmondsw: Looking at it with fresh eyes revealed some stuffs. | 17:00 |
| edmondsw | no doubt | 17:00 |
| *** AlexeyAbashkin has quit IRC | 17:02 | |
| *** AlexeyAbashkin has joined #openstack-powervm | 17:09 | |
| *** apearson has quit IRC | 17:32 | |
| openstackgerrit | Chhavi Agarwal proposed openstack/nova-powervm master: WIP: Having iSCSI Initiator locks per VIOS https://review.openstack.org/557800 | 17:38 |
| *** chhagarw has joined #openstack-powervm | 17:42 | |
| chhagarw | @here | 17:43 |
| edmondsw | chhagarw I was looking at the test failure for ps14 of ^ and have a question | 17:43 |
| edmondsw | where are 'fake-mini' and 'host1' coming from? | 17:43 |
| edmondsw | and how... since there can only ever be one host here | 17:43 |
| edmondsw | http://logs.openstack.org/00/557800/14/check/openstack-tox-py27/33e9264/testr_results.html.gz | 17:44 |
| chhagarw | 'fake-mini' is the host name | 17:45 |
| chhagarw | expected_output = { | 17:47 |
| chhagarw | 'fake-mini_1300C76F-9814-4A4D-B1F0-5B69352A7DEA': 'test_initiator', | 17:47 |
| chhagarw | 'fake-mini_7DBBE705-E4C4-4458-8223-3EBE07015CA9': 'test_initiator' | 17:47 |
| chhagarw | } | 17:47 |
| edmondsw | chhagarw I understand that... that's not what I mean | 17:49 |
| edmondsw | where did the test get that fake-mini is the host name? | 17:49 |
| edmondsw | I did a grep and can't find fake-mini anywhere | 17:49 |
| edmondsw | and not only fake-mini, but the link I pasted shows that it also thinks there is a host named 'host1' | 17:49 |
| edmondsw | where is that coming from... and why 2 different hosts? | 17:50 |
| *** apearson has joined #openstack-powervm | 17:51 | |
| edmondsw | chhagarw ? | 17:52 |
| chhagarw | yes it is picking from the wrong feed file | 17:52 |
| edmondsw | chhagarw is it though? I don't see either 'host1' or 'fake-mini' in any feed files | 17:54 |
| chhagarw | where are the feed files located ? | 17:58 |
| *** AlexeyAbashkin has quit IRC | 18:01 | |
| *** apearson has quit IRC | 18:02 | |
| edmondsw | chhagarw I don't think this has anything to do with feed files | 18:03 |
| *** apearson has joined #openstack-powervm | 18:03 | |
| edmondsw | hostname isn't in feed files | 18:05 |
| chhagarw | mock tests compares the data based on the feed file mocked, in the above test i am expecting the host name as fake-mini, which is getting returned as host1 | 18:05 |
| edmondsw | hostname should come from CONF.host | 18:05 |
| edmondsw | chhagarw not quite... you're getting 2 hosts somehow | 18:05 |
| edmondsw | fake-mini AND host1 | 18:06 |
| edmondsw | why are you expecting fake-mini? | 18:06 |
| chhagarw | host is CONF.host | 18:06 |
| edmondsw | do you set CONF.host to fake-mini somewhere? I'm not seeing that | 18:06 |
| chhagarw | no i didn't, the UT passed with the new submit | 18:09 |
| *** apearson has quit IRC | 18:09 | |
| *** apearson has joined #openstack-powervm | 18:09 | |
| edmondsw | chhagarw I'm more concerned with understanding what's going on than just whether we got lucky and the UT passed this time | 18:10 |
| edmondsw | it needs to pass all the time | 18:10 |
| edmondsw | and we need to understand what it's oing | 18:10 |
| edmondsw | doing | 18:10 |
| esberglu | efried: New conf group/option goes in features category for renos? | 18:18 |
| esberglu | All the other conf related stuff is in upgrade, but I haven't found one for a new group yet, just moving existing groups | 18:18 |
| efried | esberglu: I don't actually know; just echoing what takashin said. Ask mriedem in -nova | 18:18 |
| chhagarw | edmondsw: I can explain, can you point me to the location of feed files | 18:19 |
| edmondsw | esberglu I would assume that's a feature, yes | 18:19 |
| edmondsw | but doesn't hurt to check in nova | 18:20 |
| edmondsw | chhagarw you already know the location of feed files.. you pinged it on slack | 18:20 |
| edmondsw | but again... this has nothing to do with feed files | 18:20 |
| edmondsw | host does not come from feed files, it comes from CONF.host | 18:20 |
| chhagarw | i did not changed conf file, | 18:22 |
| chhagarw | somewhere it might be hardcoded | 18:23 |
| chhagarw | got it its picking from conf fixtures. | 18:24 |
| chhagarw | .tox/py27/src/nova/nova/tests/unit/conf_fixture.py | 18:25 |
| chhagarw | have to change the setup | 18:25 |
| efried | FYI, use self.flags to change conf values | 18:25 |
| *** AlexeyAbashkin has joined #openstack-powervm | 18:28 | |
| chhagarw | yes | 18:29 |
| edmondsw | ah, fake-mini comes from nova... that makes sense | 18:29 |
| edmondsw | efried any idea what happened here? http://logs.openstack.org/00/557800/14/check/openstack-tox-py27/33e9264/testr_results.html.gz | 18:30 |
| edmondsw | oh, nm, I think I know... | 18:31 |
| efried | looking... | 18:31 |
| efried | okay, good. Not looking... | 18:31 |
| chhagarw | test_driver we are setting the flag to host1 | 18:31 |
| chhagarw | self.flags(host='host1', my_ip='127.0.0.1') | 18:31 |
| edmondsw | an earlier test must have used CONF.host=host1 | 18:31 |
| chhagarw | i suspect it must be coming from here | 18:31 |
| edmondsw | and caused _ISCSI_INITIATORS global to get set for that | 18:31 |
| edmondsw | chhagarw so I know how to fix this | 18:31 |
| edmondsw | chhagarw stop including <host>_ in the key for _ISCSI_INITIATORS | 18:32 |
| edmondsw | there is only one host value, so it doesn't need to be included | 18:32 |
| edmondsw | and it will prevent us from getting random gate failures if tests run in different orders | 18:32 |
| chhagarw | we can just have vios_uuid as well | 18:32 |
| edmondsw | right, just key by vios uuid | 18:32 |
| chhagarw | they will be unique | 18:33 |
| chhagarw | yeah | 18:33 |
| chhagarw | also do i need to add self.flags in the setup | 18:33 |
| chhagarw | it should not be needed ? | 18:33 |
| chhagarw | or better to add in the tet setup ? | 18:34 |
| edmondsw | not needed | 18:36 |
| edmondsw | chhagarw see the comments I just posted | 18:38 |
| chhagarw | ok | 18:39 |
| esberglu | edmondsw: See nova pls | 18:40 |
| *** chhagarw has quit IRC | 18:45 | |
| *** chhagarw has joined #openstack-powervm | 18:58 | |
| *** AlexeyAbashkin has quit IRC | 19:09 | |
| chhagarw | edmondsw: replied to comments | 19:25 |
| edmondsw | chhagarw: replied again | 19:31 |
| edmondsw | I haven't see a new patch set yet | 19:31 |
| chhagarw | no patch, some queries and clarification on your comments | 19:32 |
| edmondsw | hope I covered it | 19:33 |
| chhagarw | yeah thanks | 19:35 |
| chhagarw | edmondsw: https://github.com/openstack/nova-powervm/blob/stable/queens/nova_powervm/tests/virt/powervm/volume/test_iscsi.py#L198 | 20:04 |
| chhagarw | so if we do the proper handling of iSCSIDiscoveryFailed and return None, None this test will change and return the proper exception as VolumeAttachFailed | 20:05 |
| chhagarw | p_exc.VolumeAttachFailed | 20:05 |
| chhagarw | MultipleExceptionsInFeedTask exc was previously thrown since iscsdiscovery exceptions gets overridden with none type iterator exception | 20:06 |
| edmondsw | chhagarw sounds good | 20:11 |
| edmondsw | chhagarw does it also fix the method above that? | 20:12 |
| edmondsw | I see it's also expecting MultipleExceptionsInFeedTask | 20:12 |
| chhagarw | yes | 20:12 |
| edmondsw | that's what I thought... good | 20:12 |
| openstackgerrit | Chhavi Agarwal proposed openstack/nova-powervm master: WIP: Having iSCSI Initiator locks per VIOS https://review.openstack.org/557800 | 20:18 |
| openstackgerrit | Chhavi Agarwal proposed openstack/nova-powervm master: WIP: Having iSCSI Initiator locks per VIOS https://review.openstack.org/557800 | 20:27 |
| *** esberglu has quit IRC | 20:27 | |
| *** esberglu has joined #openstack-powervm | 20:28 | |
| *** esberglu_ has joined #openstack-powervm | 20:30 | |
| *** esbergl__ has joined #openstack-powervm | 20:31 | |
| *** esbergl__ is now known as esberglu__ | 20:32 | |
| *** esberglu has quit IRC | 20:33 | |
| *** esberglu_ has quit IRC | 20:34 | |
| *** esberglu__ is now known as esberglu | 20:37 | |
| *** chhagarw has quit IRC | 20:45 | |
| efried | edmondsw: Do you think I should rip out the emulate business from VIOS-based storage types, per seroyer's comment, or just leave it alone? | 20:59 |
| edmondsw | efried oh I missed that | 21:05 |
| edmondsw | which ones would you rip it out of? | 21:05 |
| edmondsw | I think VDisk is used for both? FileIO and RBD are only Linux. Not sure about PV and LU | 21:07 |
| esberglu | edmondsw: Want to make sure I have this proc_units_factor stuff straight | 21:07 |
| esberglu | It will use the powervm:proc_units extra spec if specified | 21:08 |
| esberglu | Otherwise will use the conf value if specified, with default 0.1 if not specified | 21:08 |
| edmondsw | esberglu powervm:proc_units is how many proc_units you want | 21:09 |
| edmondsw | I believe powervm:min_proc_units and max are the min/max that you can resize it to later | 21:10 |
| edmondsw | mdrabe is that right? | 21:10 |
| edmondsw | esberglu so yes | 21:10 |
| efried | edmondsw: I think at least LU is a VIOS-only thing, because SSP-only. | 21:11 |
| mdrabe | Yep that's it | 21:11 |
| edmondsw | efried today, and that may change, or... ? | 21:11 |
| edmondsw | I guess if it did change we could always add at that point | 21:12 |
| efried | shrug. Fact is, even for the places we've put it in, there's a bunch of code paths and permutations where it would be ignored/irrelevant. | 21:12 |
| efried | So having one more doesn't bother me a lot. | 21:12 |
| edmondsw | I'm not sure I really have a strong opinion on this | 21:12 |
| efried | Especially if it means a tad more consistency in the interfaces. Though that ship has sailed long ago. | 21:12 |
| edmondsw | yeah | 21:13 |
| edmondsw | I'm fine merging as-is | 21:13 |
| efried | k | 21:13 |
| *** edmondsw has quit IRC | 21:22 | |
| *** edmondsw has joined #openstack-powervm | 21:24 | |
| *** edmondsw has quit IRC | 21:29 | |
| *** tjakobs has quit IRC | 21:43 | |
| efried | esberglu: Is it not called proc_units_factor ? | 22:19 |
| efried | I thought proc_units was something different (number of proc units) | 22:20 |
| efried | yeah | 22:20 |
| esberglu | efried: proc_units_factor doesn't exist in nova. I'm confused again now, just asked about this above | 22:22 |
| efried | esberglu: When you say "in nova", where are you looking? | 22:22 |
| efried | Oh, I guess we must be doing the extra specs processing in the nova side of the code. | 22:22 |
| efried | So we probably have to port in the proc_units_factor processing from OOT for this to be sane. | 22:23 |
| efried | oh. | 22:24 |
| efried | Which we've done. | 22:24 |
| efried | We just did it via **kwargs | 22:24 |
| efried | So yeah, other than the typo and the possible reno enhancement, what you've got is fine. | 22:25 |
| *** apearson has quit IRC | 22:26 | |
| efried | esberglu: Sorry, no | 22:27 |
| esberglu | efried: I've really got to run, can you post whatever comments and I'll check it out tonight? | 22:27 |
| efried | yah | 22:27 |
| esberglu | tx | 22:27 |
| *** esberglu has quit IRC | 22:27 | |
| *** esberglu has joined #openstack-powervm | 22:28 | |
| *** esberglu_ has joined #openstack-powervm | 22:29 | |
| *** esberglu has quit IRC | 22:33 | |
| *** esberglu_ has quit IRC | 22:34 | |
| *** esberglu has joined #openstack-powervm | 22:50 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!