Monday, 2017-01-30

*** thorst_ has joined #openstack-powervm00:00
*** thorst_ has quit IRC00:04
*** BorD_ has quit IRC00:08
*** thorst_ has joined #openstack-powervm00:13
*** thorst_ has quit IRC00:13
*** Jay2 has quit IRC00:21
*** Jay2 has joined #openstack-powervm00:22
*** Jay2 has quit IRC00:26
*** BorD_ has joined #openstack-powervm00:38
*** Jay2 has joined #openstack-powervm00:57
*** Jay3 has joined #openstack-powervm01:42
*** Jay2 has quit IRC01:43
*** thorst_ has joined #openstack-powervm01:44
*** thorst_ has quit IRC01:49
*** thorst_ has joined #openstack-powervm02:09
*** thorst_ has quit IRC02:12
*** thorst_ has joined #openstack-powervm02:57
*** thorst_ has quit IRC02:57
*** thorst_ has joined #openstack-powervm03:07
*** thorst_ has quit IRC03:07
*** Jay2 has joined #openstack-powervm03:48
*** Jay3 has quit IRC03:49
*** Jay2 has quit IRC03:52
*** Jay2 has joined #openstack-powervm03:53
*** Jay3 has joined #openstack-powervm03:59
*** Jay2 has quit IRC04:00
*** dwayne has quit IRC04:15
*** thorst_ has joined #openstack-powervm05:49
*** Jay2 has joined #openstack-powervm05:50
*** Jay3 has quit IRC05:52
*** thorst_ has quit IRC05:55
*** tjakobs has joined #openstack-powervm06:41
*** thorst_ has joined #openstack-powervm07:51
*** thorst_ has quit IRC07:56
*** tjakobs has quit IRC08:21
*** Jay3 has joined #openstack-powervm08:21
*** Jay2 has quit IRC08:23
*** k0da has joined #openstack-powervm08:29
*** Jay1 has joined #openstack-powervm09:50
*** Jay3 has quit IRC09:51
*** thorst_ has joined #openstack-powervm09:52
*** thorst_ has quit IRC09:57
*** Jay2 has joined #openstack-powervm11:08
*** Jay1 has quit IRC11:11
*** Jay2 has quit IRC11:31
*** Jay1 has joined #openstack-powervm11:31
*** Jay1 has quit IRC11:35
*** thorst_ has joined #openstack-powervm11:53
*** thorst_ has quit IRC11:57
*** thorst_ has joined #openstack-powervm12:41
*** edmondsw has joined #openstack-powervm13:07
efriedthorst_ What's new?13:15
efriedIma be trucking along on the OVS and SSP change sets, just wanted to make sure there wasn't something else that needed more urgent attention.13:19
thorst_nothing that i know of.  I'm getting the system test team going now, and they're going to try an 'in tree' devstack just to get familiar with how to set up a devstack with pending patch sets.13:20
efrieddig.  Are they on IRC? (Loaded question)13:23
thorst_they should be.  nilesh is one.  Haven't seen jay on yet13:23
thorst_but that's how communication is to be handled...13:23
*** tblakes has joined #openstack-powervm13:31
*** mdrabe has joined #openstack-powervm13:36
*** tlian has joined #openstack-powervm13:52
*** nbante has joined #openstack-powervm13:57
*** kjw3 has quit IRC14:04
*** nbante has quit IRC14:37
*** kriskend has joined #openstack-powervm14:53
*** dwayne has joined #openstack-powervm14:54
*** burgerk has joined #openstack-powervm15:07
*** apearson has joined #openstack-powervm15:27
*** tjakobs has joined #openstack-powervm15:46
efriedthorst_ There was one question I had on the ovs-vif patch16:02
thorst_?  in comments or here?16:02
efriedWell, either would do.
efriedThe question is whether we really need to introduce these conf options at all.16:03
efriedI get the sense there's a desire to be conservative with the number of conf options we propose.  If that's not a thing, I won't worry about it.16:03
thorst_maybe not for pvm_vswitch_for_novalink_io16:03
thorst_but definitely for use_rmc_mgmt_vif16:03
thorst_we want that there for sure16:04
thorst_we turn that off in our CI actually16:04
thorst_it was leading to slow activation times when we were deploying real VMs (can't remember why off hand)16:04
thorst_ahh, no it was undercloud we had it turned off16:11
thorst_I think it made that weird pypowervm patch we have work better16:12
*** esberglu has joined #openstack-powervm16:14
*** k0da has quit IRC16:23
*** apearson has quit IRC16:31
*** apearson has joined #openstack-powervm16:33
*** nbante has joined #openstack-powervm16:52
*** apearson has quit IRC17:09
*** apearson has joined #openstack-powervm17:09
*** apearson has quit IRC17:21
*** apearson has joined #openstack-powervm17:50
*** nbante has quit IRC18:07
*** k0da has joined #openstack-powervm19:04
*** burgerk has quit IRC19:30
*** burgerk has joined #openstack-powervm19:32
*** kriskend has quit IRC21:25
*** kriskend has joined #openstack-powervm21:26
efriedthorst_ There's a bit of incongruity in the vif unplugging stack.  The low-level methods (the vif driver-specific ones that do the real work) return stuff; but the module-level method doesn't kick them back up the stack; and the consuming Tasks don't use 'em.21:30
efriedI'm okay with that last thing, but it seems to me like the module-level unplug method should return whatever the low-level methods return.21:30
efriedOr we should rip the return values out of the low-level methods.21:30
efried( btw)21:31
thorst_yeah, I don't think it should return up the stuff21:32
thorst_because depending on what you plug/unplug that stuff is wildly different21:33
thorst_vNIC vs. VF vs. VEA21:33
efriedOkay, I'll rip the return values out of the low-level methods.21:36
thorst_well, we did use them to do slot store stuff21:36
thorst_but the issue in tree is we're not doing any slot store initially?21:36
efriedEven OOT, the slot storing is done locally, and the module-level unplug isn't returning the value.21:37
thorst_I think it could be removed off hand21:37
efriedOkay, I'll make sure, and rip it.21:37
efriedOne other thing.21:37
efriedPvmLioVifDriver.plug is doing 'ip link set up'21:38
efriedAnd then PvmOvsVifDriver.plug is doing it again.21:38
efriedIs that on purpose?21:38
efriedBoth are doing get_trunk_dev_name too.  Is that related?21:38
thorst_when you DLPAR in an adapter on the NovaLink it is not in the up state automatically21:38
thorst_yeah, same deal...just was harder to refactor out (looked grosser) than to leave the duplication in21:39
thorst_it was a diminishing return, also in readability I think21:39
efriedSoooo, harmless to do it twice?21:39
thorst_(if I remember right)21:39
thorst_o yeah, harmless21:39
thorst_I guess if we're doing it twice, we could remove the second call of link up21:39
efriedDoes the trunk_dev_name change?21:39
efried...within one instance of a vif driver class?21:40
thorst_within one instance of a vif21:40
efriedOkay, so no good making the trunk_dev_name an instance var so we don't have to re-grab it.21:40
thorst_well, I think it is within an instance of the driver as written today21:41
thorst_but meh...its simple string replacement21:41
thorst_your call21:41
efriedoh, yeah, I thought it was a REST calll.21:42
thorst_nope, input into a rest call21:43
efriedthorst_  does it matter if I get_trunk_dev_name before or after the crt_p2p_cna call?21:44
efriedDoesn't look like it.21:44
thorst_its input into crt_p2p_cna21:44
thorst_so get it before that.21:44
efriedBut no need to re-get it after.21:44
efriedSo yeah, those two lines appear redundant to me.21:44
thorst_nah, you're just telling the REST API, here is what to set the device name when you create it.21:44
*** apearson has quit IRC22:06
*** kriskend has quit IRC22:13
*** kriskend has joined #openstack-powervm22:17
openstackgerritEric Fried proposed openstack/nova-powervm: In-tree backports (2)
*** tblakes has quit IRC22:24
*** apearson has joined #openstack-powervm22:29
*** apearson has quit IRC22:31
*** edmondsw has quit IRC22:34
*** thorst_ has quit IRC22:36
*** apearson has joined #openstack-powervm22:44
*** kriskend has quit IRC22:58
openstackgerritTaylor Jakobson proposed openstack/nova-powervm: Add support for File I/O Driver
*** thorst_ has joined #openstack-powervm23:06
*** tjakobs has quit IRC23:10
*** thorst_ has quit IRC23:10
*** mdrabe has quit IRC23:11
*** apearson has quit IRC23:18
*** k0da has quit IRC23:35

Generated by 2.14.0 by Marius Gedminas - find it at!