Friday, 2017-08-11

*** thorst has quit IRC00:09
*** thorst has joined #openstack-powervm00:10
*** thorst has quit IRC00:11
*** edmondsw has joined #openstack-powervm00:17
*** edmondsw has quit IRC00:22
*** smatzek has joined #openstack-powervm00:27
*** esberglu has joined #openstack-powervm00:58
*** esberglu has quit IRC01:02
*** thorst has joined #openstack-powervm01:11
*** thorst has quit IRC01:17
*** edmondsw has joined #openstack-powervm02:06
*** edmondsw has quit IRC02:10
*** thorst has joined #openstack-powervm02:12
*** thorst has quit IRC02:17
*** thorst has joined #openstack-powervm03:13
*** smatzek has quit IRC03:20
*** thorst has quit IRC03:21
*** edmondsw has joined #openstack-powervm03:54
*** edmondsw has quit IRC03:58
*** chhavi has joined #openstack-powervm04:00
*** thorst has joined #openstack-powervm04:17
*** thorst has quit IRC04:22
*** thorst has joined #openstack-powervm05:17
*** thorst has quit IRC05:22
*** edmondsw has joined #openstack-powervm05:42
*** edmondsw has quit IRC05:46
*** thorst has joined #openstack-powervm06:18
*** thorst has quit IRC06:23
*** thorst has joined #openstack-powervm07:19
*** thorst has quit IRC07:23
*** chhavi has quit IRC09:10
*** thorst has joined #openstack-powervm09:21
*** thorst has quit IRC09:25
*** chhavi has joined #openstack-powervm10:00
*** thorst has joined #openstack-powervm10:22
*** thorst has quit IRC10:26
*** edmondsw has joined #openstack-powervm11:10
*** thorst has joined #openstack-powervm11:22
*** thorst has quit IRC11:27
*** svenkat has joined #openstack-powervm11:53
*** svenkat has quit IRC11:57
*** svenkat has joined #openstack-powervm11:58
*** thorst has joined #openstack-powervm12:08
*** adi_____ has quit IRC12:19
*** adi_____ has joined #openstack-powervm12:19
*** efried is now known as fried_rice12:23
*** kylek3h has joined #openstack-powervm12:38
*** dwayne has quit IRC13:05
*** esberglu has joined #openstack-powervm13:33
edmondswthorst, know of any reason we have enabled to neutron qos plugin in our CI?14:21
thorstbecause we can14:21
thorstthough it only 'works' for OVS14:21
thorstfor SEA it 'works', but doesn't actually apply rules14:21
edmondswso its kinda pointless, right?14:21
thorstit was more of a 'why not'.14:21
edmondswthere's a bug that we have to workaround because of that plugin14:22
edmondswwe're talking about removing it, so we don't have to have that workaround14:22
edmondswunless you object14:22
thorstI'm not wedded to that being in the CI14:22
edmondswesberglu fried_rice ^14:23
fried_riceMake it so, #1.14:23
*** k0da has joined #openstack-powervm14:44
*** openstackgerrit has joined #openstack-powervm14:55
openstackgerritEric Berglund proposed openstack/nova-powervm master: DNM: ci check
*** k0da has quit IRC15:09
esberglufried_rice: edmondsw: On it15:12
fried_riceWe also need to propose updates to our tox configs for pike.  I'll get on that.15:12
*** tjakobs has joined #openstack-powervm15:12
openstackgerritEric Fried proposed openstack/nova-powervm master: Update tox.ini settings for pike
openstackgerritEric Fried proposed openstack/networking-powervm master: Update tox.ini settings for pike
esbergluedmondsw: What workaround are you referring to for qos?15:19
edmondswesberglu creating /etc/neutron15:20
openstackgerritEric Fried proposed openstack/ceilometer-powervm master: Update tox.ini settings for pike
esbergluOh disabling qos fixes that? Nice15:20
edmondswesberglu I think it should... haven't tried it15:21
fried_riceesberglu edmondsw I have, it works.15:22
edmondswat least it will roll back the thing that's currently hitting it15:22
edmondswok, cool15:22
esbergluedmondsw: I'll do a CI run with it before merging15:22
fried_riceesberglu I did this:15:22
fried_rice# enable_service n-novnc neutron neutron-api neutron-qos pvm-q-sea-agt pvm-q-sriov-agt15:22
fried_riceenable_service n-novnc neutron neutron-api pvm-q-sea-agt pvm-q-sriov-agt15:22
fried_riceand this:15:22
*** k0da has joined #openstack-powervm15:22
fried_rice# extension_drivers=port_security,qos15:23
*** dwayne has joined #openstack-powervm15:31
*** k0da has quit IRC15:59
*** k0da has joined #openstack-powervm16:13
*** dwayne has quit IRC16:15
*** dwayne has joined #openstack-powervm16:25
mdrabefried_rice Do you remember off the top of your head if the slot map is supposed to honor the slot *pairs* between source and destination VIOSes for NPIV RR?16:34
mdrabei.e. vios 1 has client slots 7 and 8, vios 2 has slots 9 and 10, so should 7 and 8 stay together on a vios on the destination?16:35
mdrabeWait crap it's not the VIOSes that are important with NPIV it's the FC ports (fabrics) I think.16:42
mdrabethorst do you remember this at all ^ ?16:42
*** k0da has quit IRC16:42
thorstmdrabe: yes, it must honor pairs16:45
thorstand its a huge PITA to do that16:45
mdrabethorst: pairs between VIOSes or pairs between FC ports?16:45
thorst(A1, B1), (A2, B2) on source can not flip to (A1, B2), (A2, B1) no target16:46
thorstFC ports on a given vios16:46
mdrabethorst: Is this OK then?
*** apearson has joined #openstack-powervm17:00
thorstmdrabe: nope17:00
thorstthat flipped17:00
thorstit shouldn't have flipped17:00
thorstthat indicates maybe a loss of the slot store?17:00
mdrabeIt honored the port pairs though17:01
mdrabeAnd I don't think we lost the slot store since we have the same client slot numbers17:01
mdrabethorst: I'm hungry and gonna get something to eat, I'll dive when I get back.17:02
*** mdrabe is now known as mdrabe_afk17:02
thorst7/9 were on the same vios initially.  Now they're on separate vioses?17:03
*** apearson has quit IRC18:09
*** esberglu has quit IRC18:16
*** esberglu has joined #openstack-powervm18:16
*** esberglu has quit IRC18:21
*** kylek3h has quit IRC18:48
edmondswfried_rice so these pike tox changes... we make them on master, branch pike, and then revert them on master? Not branch pike and then make them on pike?18:57
mdrabe_afkthorst: yea that's correct18:58
*** mdrabe_afk is now known as mdrabe18:58
thorstmdrabe: that's the scenario we try to avoid.18:58
mdrabethorst: I think we've always been doing it this way18:58
edmondswfried_rice I would think we'd want to branch pike and then make these pike tox changes on the stable/pike branch instead of master. And that's what we apparently did back in April19:07
mdrabethorst: Shyama made the change awhile ago to honor slot peers on vioses for migration. The bug is
openstackLaunchpad bug 1623889 in nova-powervm "live migration of instance using npiv volumes and multiple vios does not take the client slot into account" [Undecided,Fix released] - Assigned to Shyama (shyvenug)19:24
mdrabeSo that was done to make sure the MPIO config was replicated on the target, the question I have then is do we care about this during evacuation?19:25
edmondswmdrabe why would care about it on LPM and not on evac?19:34
mdrabeedmondsw because the guest is effectively rebooted on evac19:34
mdrabeSo I *think* the MPIO config gets reset19:34
mdrabeI'm sending a note out about it19:35
edmondswoh, you mean SRR when you say evac19:35
mdrabeWell it's actually evac evac in this case19:36
edmondswthe term evac gets used for too many things...19:37
mdrabeYes it's... complicated19:37
*** fried_rice has quit IRC19:41
*** k0da has joined #openstack-powervm19:49
*** fried_rice has joined #openstack-powervm19:53
*** thorst has quit IRC20:52
*** chhavi has quit IRC20:53
openstackgerritMatthew Edmonds proposed openstack/nova-powervm master: ISCSI volume attachment fixes
*** edmondsw has quit IRC21:01
fried_riceHm, that's bizarre; I totally created my local branches against stable/pike.21:02
fried_riceI think I know what I did wrong - I said git review instead of git review stable/pike.  But I'm surprised it let me do it at all in that case.21:04
*** k0da has quit IRC21:04
*** dwayne has quit IRC21:09
*** tjakobs has quit IRC21:42
*** thorst has joined #openstack-powervm21:44
*** thorst has quit IRC21:48
*** svenkat has quit IRC21:51
*** dwayne has joined #openstack-powervm22:03
*** esberglu has joined #openstack-powervm23:17
*** esberglu has quit IRC23:21

Generated by 2.15.3 by Marius Gedminas - find it at!