Wednesday, 2019-01-02

*** catmando has joined #openstack-powervm14:53
*** catmando_ has joined #openstack-powervm14:56
*** efried has joined #openstack-powervm15:02
catmandohey all]15:07
catmandobeen having a look at powervc 1.4.215:08
catmandoi'm trying to find the "Fibre Channel tape pluggable driver" option in the interface somewhere: any ideas?15:08
efriedcatmando: You may want to wait for mdrabe or edmondsw to show up.15:08
*** edmondsw has joined #openstack-powervm15:09
catmandoi shall15:09
catmandono rush, thank you15:09
edmondswcatmando I'm waiting for eavesdrop to catch up, but I heard you might have a question for me?15:12
catmandopowervc 1.4.2: i'm trying to find the "Fibre Channel tape pluggable driver" option in the interface somewhere: any ideas?15:13
catmandoedmondsw: ^15:17
edmondswwell that's weird... it should have that info in the Knowledge Center but I'm not finding it15:18
edmondswgive me a minute15:18
edmondswI assume you're asking where the CLI command is to configure tapes?15:20
edmondswI believe it's in /opt/ibm/powervc/bin15:20
edmondswsomething like powervc-tape-register15:20
edmondswcatmando ^15:20
catmandoedmondsw: aha, let me have a look at that15:21
catmandoi was hoping that there would be something like the port load balancing15:22
edmondswthat registers an FC tape driver, so I think the port configuration stuff would work the same as other FC drivers15:25
edmondswbut I'm not sure about that15:27
edmondswpowervc-tape-register -n DISPLAY_NAME -p TARGET_PORTS [-c CAPACITY]15:27
edmondswor something like that15:27
edmondswcatmando I believe chhagarw designed/wrote that, so I've pinged her to get on IRC if/when she's available.15:35
edmondswand pinged our docs folks about why it's not in the KC15:35
*** chhagarw has joined #openstack-powervm15:40
chhagarwcatmando: Seems, you have a question for FC Tape pluggable driver ?15:45
chhagarwHope you are able to find powervc-tape-register, this CLI can be used to register the FC Storage target ports to PowerVC.15:46
catmandochhagarw: yes, i am looking at that now16:32
catmandobut there's no docs and I thought there might be a frontend (similar to the one for port groups)16:35
catmandochhagarw: the cli docs state: "specifies the comma separated target ports of the FC tape storage device."; does this mean the "target" as in the port on the vios?16:42
chhagarwcatmando: we do not have it documented in KC currently, we will check on that.16:42
catmandoi assume it does not mean the actual target port on the remote side16:42
chhagarwCLI, requires target ports list, which are the FC Tape device target ports,16:43
chhagarwonce its registered using CLI, it will get listed in the PowerVC GUI as the storage device.16:44
catmandoso, the wwpn of the remote side, as opposed to the wwpn on the vios...16:44
chhagarwYou can consider, FC Tape as a pluggable storage device, so while registering user needs to provide the remote storage target ports, not the vios wwpns.16:47
chhagarwonce registered these target ports wwpns will be used to zone the NPIV VM wwpns.16:47
catmandothat makes sense16:48
catmandochhagarw: i'll have a play around with it in our lab and get back to you when i see you online if anything seems odd. many thanks for the help and for getting this into powervc16:50
chhagarwchhagarw: sure, will look forward for your response.16:51
*** catmando_ has quit IRC17:04
*** chhagarw has quit IRC17:31
openstackgerritMatthew Edmonds proposed openstack/networking-powervm master: Update tests for policy-in-code
edmondswefried ^ that should fix UTs for networking-powervm, which are currently failing17:39
openstackgerritMatthew Edmonds proposed openstack/networking-powervm master: Update tests for policy-in-code
edmondswhad to fix a mistake in the commit message...17:40
efriededmondsw: Seem to remember we're not supposed to import things from neutron. Doesn't neutronlib expose a base test case?17:50
edmondswI don't know... I can look17:50
edmondswthat wouldn't be new, but we could certainly clean it up while we're in here if you're right17:51
edmondswneutron has a, which means it is not exposed17:52
edmondswI mean neutron-lib17:52
edmondswthe most correct thing to do may be to create our own BaseTestCase class in networking-powervm, but I don't know that I want to bite that off right now17:57
edmondswthat's what networking-hyperv does17:57
efriedthere's a hacking rule that we have on our ignore list.17:58
efried# N530 direct neutron imports not allowed17:58
edmondswyeah, I'm not disagreeing with you... just saying a) I'm not introducing that and b) let's worry about that later in another commit17:59
edmondswall I'm doing here is removing the middleman17:59
efriedI understand. Cleanest thing would probably be to switch to18:01
efriedclass BasePVMTestCase(neutron.tests.base.BaseWhatever):18:01
efried    pass18:01
efriedso we don't have to switch those other things again when we figure out what we're supposed to be doing.18:01
edmondswthat's easy enough18:01
efriedfyi, asked in -neutron and the answer is that a neutron-lib base test case is on the to-do list; for now import from neutron. But yeah, I think doing the empty local class for now - possibly with a TODO about ^ - is the right thing.18:03
openstackgerritAndreas Jaeger proposed openstack/networking-powervm master: Use template for lower-constraints
openstackgerritMatthew Edmonds proposed openstack/networking-powervm master: Update tests for policy-in-code
edmondswefried ^18:39
openstackgerritEric Fried proposed openstack/networking-powervm master: Use template for lower-constraints
edmondswefried while you're looking at that lc one, there's an identical change in nova-powervm that needs your +218:48
edmondswefried and when you get a chance, these:
openstackgerritMerged openstack/networking-powervm master: Update tests for policy-in-code
openstackgerritMerged openstack/nova-powervm master: Use template for lower-constraints
openstackgerritMerged openstack/networking-powervm master: Use template for lower-constraints
efriededmondsw: So before the py36 job would fail?19:25
edmondswit would fail locally, but it was somehow working for zuul.... I'm not sure how19:26
efriedugh. I'm looking for some justification for these changes to be combined into one.19:32
edmondswthat's the justification... neither makes sense without the other19:33
edmondswif I just changed the UT, what's the point if the reason for the change was for py36 local UTs and we don't have that enabled in our tox.ini19:33
edmondswand if I just do the tox.ini change, what's the point if I still can't run the UTs locally because I'm missing the UT change19:34
edmondswbut I am concerned, now that you mention it, that the py36 zuul job hasn't been failing all this time...19:34
efriedI'm +2 on those three changes19:34
efriedIIUC, the failure would be intermittent/unpredictable based on random hash ordering or something similar.19:34
efriedSaid hashing/ordering is affected by things, including env vars like PYTHONHASHSEED.19:35
efriedWhich for all we know might be (un)set on the zuul server and vice versa when we run locally.19:35
efriedin any case, changing it to sorted() so it's predictable is goodness.19:35
edmondswthat helps... tx19:36
edmondswinfra may have set PYTHONHASHSEED earlier for some other reason and it has masked this for us19:36
openstackgerritMerged openstack/networking-powervm master: More py3.x support
openstackgerritMerged openstack/nova-powervm master: More py3.x support
openstackgerritMerged openstack/ceilometer-powervm master: More py3.x support
*** efried has quit IRC21:21
*** efried has joined #openstack-powervm21:29
*** efried has quit IRC22:39
*** catmando has quit IRC23:08

Generated by 2.15.3 by Marius Gedminas - find it at!