Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-04-25
10:52:02 stephenfin What do you mean?
10:52:12 bauzas there is a contextmanager
10:52:50 stephenfin bauzas, jaypipes: As an aside, here's my DB schema for the cell0 table http://paste.openstack.org/show/719902/
10:52:54 stephenfin bauzas: Let me check
10:53:08 bauzas stephenfin: I'll need to go out
10:53:10 bauzas for eating
10:53:18 stephenfin bauzas: No problem. Let's pick this up after lunch
10:53:19 bauzas but lemme find some explanations
10:57:20 openstackgerrit Kashyap Chamarthy proposed openstack/nova master: libvirt: Lift the restriction of choices for `cpu_model_extra_flags` https://review.openstack.org/563926
10:57:21 openstackgerrit Kashyap Chamarthy proposed openstack/nova master: Do-Not-Merge: Test 1GB Huge Pages https://review.openstack.org/564161
10:59:38 bauzas stephenfin: in https://docs.openstack.org/nova/latest/user/cellsv2-layout.html#multiple-cells you can see the top-level superconductor using multiple MQs
11:09:09 kashyap sean-k-mooney: Respun the patch above, if you get a minute to ACK it today...
11:20:54 bhagyashris DISK_GB used from the Compute Node resource provider. Expected that it should take the DISK_GB from the Shared Resource Provider and VCPU, MEMORY_MB from Compute Node resource provider.
11:20:54 bhagyashris kashyap: Hi, I am trying to write the functional test for https://review.openstack.org/#/c/560444/ https://review.openstack.org/#/c/560459/ which related to libvirt change so can you guide me how should i write the test for the same. Basically this change is fixing related to scenario: As on current master, if user configured the shared resource provider along with the Compute Node resource providers on libvirt and boot the instance then the
11:27:46 bhagyashris kashyap: I have checked the existing functional test cases on master under /nova/tests/functional/libvirt/test_pci_sriov_servers.py but there they are directly building the instance using the post call but in my case i will need to resource provider to build the instance
11:28:32 bhagyashris kashyap: so can you guide?
11:29:49 openstackgerrit Balazs Gibizer proposed openstack/nova master: Transform instance.exists notification https://review.openstack.org/403660
12:05:38 openstackgerrit wanghongtao proposed openstack/nova master: Fix the metadata re to match the unicode https://review.openstack.org/536236
12:09:36 kashyap bhagyashris: Was AFK, reading the scrollback now
12:10:26 bhagyashris kashyap: ok
12:11:06 kashyap bhagyashris: I haven't had a chance yet to use Resource Providers, so wouldn't really have great advice. Still, reading the change
12:11:47 kashyap bhagyashris: I think you wanted to address Eric's feedback:
12:11:51 kashyap "...or even better, don't mock - create an actual ProviderTree instance and assert its actual contents on the other side of the update_provider_tree call."
12:14:50 bhagyashris kashyap: no no Actually I have address the unit test cases related comments
12:15:27 bhagyashris kashyap: now i am talking about the fuctional test case part
12:18:14 openstackgerrit Chris Dent proposed openstack/nova master: Optional separate database for placement API https://review.openstack.org/362766
12:18:15 openstackgerrit Chris Dent proposed openstack/nova master: WIP: Add PLACEMENT_DB_ENABLED=True to the nova-next job https://review.openstack.org/564067
12:18:35 kashyap bhagyashris: Ah, right
12:19:06 openstackgerrit Chris Dent proposed openstack/nova master: Isolate placement database config https://review.openstack.org/541435
12:19:20 kashyap bhagyashris: So you don't have a precedent for functional test that uses Resource Provider to build an instance
12:19:43 openstackgerrit Chris Dent proposed openstack/nova master: WIP: Ensure that os-traits sync is attempted only at start of process https://review.openstack.org/553857
12:24:15 bhagyashris kashyap: yeah as of now I haven't seen the functional like the resource providers are created and that are used to build the instance
12:24:57 kashyap bhagyashris: I'm afraid, I'm not really familiar with that area, so I can't help without extensive trial-and-error myself. Perhaps ask on the mailing list with [nova] tag?
12:25:45 bhagyashris kashyap: ok
12:27:00 kashyap bhagyashris: Since people are spread across time zones, you might get delayed (which is normal) responses on IRC. That's why I suggested the list.
12:28:35 bhagyashris kashyap: ok. currently i am struggling to write that so asking on IRC for quick response
12:28:49 kashyap Yeah, a combination is fine. bhagyashris How about this:
12:29:13 kashyap bhagyashris: Post what you have based on how far you've come, and then point that to a Resource Providers expert (maybe jaypipes) for more guidance?
12:30:16 kashyap Also, Zuul is still failing on your https://review.openstack.org/#/c/560459/.
12:30:31 bhagyashris kashyap: ok waiting for others will see
12:31:24 bhagyashris kashyap: yeah struggling for the same also randomly the tempest test are falling
12:32:58 bhagyashris kashyap: and all are failing with the error as No valid host is found. I have also checked the logs and at n-sch side it's showing that the Placement is returning allocation_candidates
12:33:05 kashyap I see
12:33:58 bhagyashris kashyap: I am not getting why the n-sch is not returning the allocation_candicates as my change is at nova compute side
12:34:29 bhagyashris kashyap: and on master evrthing is fine only on patch patches it's failing
12:35:08 bhagyashris kashyap: any idea/
12:35:10 bhagyashris ?
12:35:58 openstackgerrit Balazs Gibizer proposed openstack/nova master: Escalate UUID validation warning to error in test https://review.openstack.org/540386
12:35:59 openstackgerrit Balazs Gibizer proposed openstack/nova master: Fix invalid UUIDs in test https://review.openstack.org/564190
12:37:25 kashyap bhagyashris: Not without delving too deep. Please go through the traces systematically to find the root cause. I'm a bit buried with a couple of urgent things
12:37:54 bhagyashris kashyap: ok np thank you :)
12:38:34 bhagyashris jaypipes: Hi,
12:47:35 openstackgerrit Martin Midolesov proposed openstack/nova master: Validating video ram against the max video ram allowed https://review.openstack.org/564193
13:14:26 openstackgerrit Balazs Gibizer proposed openstack/nova master: Enhance service restart in functional env https://review.openstack.org/512552
13:14:28 openstackgerrit Balazs Gibizer proposed openstack/nova master: cleanup evacuated instances not on hypervisor https://review.openstack.org/512623
13:14:28 openstack bug 1724172 in OpenStack Compute (nova) "Allocation of an evacuated instance is not cleaned on the source host if instance is not defined on the hypervisor" [Undecided,In progress] https://launchpad.net/bugs/1724172 - Assigned to Balazs Gibizer (balazs-gibizer)
13:14:28 openstackgerrit Balazs Gibizer proposed openstack/nova master: Reproduce bug 1724172 in the functional test env https://review.openstack.org/512553
13:29:29 openstackgerrit Sylvain Bauza proposed openstack/nova master: libvirt: mdevs returning parent and vendor PCI info https://review.openstack.org/562304
13:29:30 openstackgerrit Sylvain Bauza proposed openstack/nova master: libvirt: Configure multiple vGPU types https://review.openstack.org/564214
13:29:31 openstackgerrit Sylvain Bauza proposed openstack/nova master: WIP: libvirt: Get inventories per pGPU https://review.openstack.org/564215
13:38:05 mnaser super unrelated but i feel like this might be something devs have had to deal with
13:38:21 mnaser is there some sort of 'tcpdump for rabbitmq' toolset available anywhere (or part of rabbitmq)
13:43:44 stephenfin bauzas: OK, looked at that and it does seem like it's targeted. See line 35 from the paste paste.openstack.org/show/719901/
13:43:50 stephenfin and compare to https://github.com/stephenfin/nova/blob/bp/numa-aware-vswitches/nova/conductor/manager.py#L1172-L1175
13:46:41 stephenfin But wait - the *cell1* table wasn't migrated. Well that's annoying. I would have expected 'nova-manage sync' to do all cell tables unless I explicitly said otherwise http://paste.openstack.org/show/719912/
13:46:56 openstackgerrit Hamdy Khader proposed openstack/nova master: Adding NVMEoF for libvirt driver https://review.openstack.org/482640
13:47:08 stephenfin dansmith: Any reason 'nova-manage db sync' doesn't sync all cell DBs at once?
13:47:35 dansmith stephenfin: yeah, because that would potentially lock down all of them at once (or nearly) and maybe you're upgrading one cell before another
13:48:12 dansmith db sync is/can be run at runtime before upgrading code and has a significant impact on the performance of the db while running
13:48:32 dansmith we have or had a --all-cells flag proposed for it
13:48:47 stephenfin Right, and the non upgraded cell would be using old code so you wouldn't see backtraces like I did above
13:49:14 stephenfin (my next question was going to be "wouldn't the older schema be invalid for newer code")
13:53:53 mriedem stephenfin: if you want to brush this off https://review.openstack.org/#/c/420973/
13:55:07 stephenfin Um, so how do I upgrade the schema for the cell1 DB? It's an all-in-one deployment so cell0 and cell1 DB are on the same node
13:55:44 stephenfin The docs say " If --local_cell is specified, then only the main database in the current cell is upgraded." but my main database for an all in one is...both of them?
13:55:47 stephenfin mriedem: I can do that
13:55:54 dansmith use the config that the cell1 services use
13:58:09 stephenfin So change '[database] connection' in nova.conf to the cell1 DB?
13:58:10 stephenfin Wait
13:58:24 stephenfin There's a '--config-file'. That should show up when I do 'nova-manage db sync --help'
13:58:39 stephenfin #usability
13:59:27 mriedem stephenfin: is this devstack?
13:59:33 stephenfin mriedem: It is
13:59:35 mriedem if so, by default, even though it's single node, it's not all in one
13:59:41 mriedem see under /etc/nova
13:59:45 mriedem there should be multiple config files
13:59:58 mriedem nova_cell1.conf is the cell1 conf
14:00:00 stephenfin Sorry, I'm using the wrong terminology so. I meant everything is deployed onto one machine
14:00:02 mriedem nova-cpu.conf is for nova-compute
14:00:16 mriedem nova.conf is used for the controller services and n-super-cond
14:00:24 stephenfin Yup, just ran this and things are looking good -> $ nova-manage --config-file /etc/nova/nova_cell1.conf db sync
14:06:11 bhagyashris efried: Hi, I am trying to write the functional test for https://review.openstack.org/#/c/560444/ https://review.openstack.org/#/c/560459/ but have some question about that: As of on current master there no as such functional test that is creating the RP providers and used to boot the instance or to build the instance
14:06:53 openstackgerrit Balazs Gibizer proposed openstack/nova master: Send soft_delete from context manager https://review.openstack.org/476459
14:10:31 efried bhagyashris: You may have to piece it together from other bits.
14:11:58 efried bhagyashris: Did you figure out the error from yesterday?
14:13:04 bhagyashris efried: or I can say there is no frame work in fictional test that creates the resource provider and that providers will be used to build the instances
14:14:45 bhagyashris efried: and one more doubt is how that created provider will be pick by sch at the time of build instance
14:15:44 bhagyashris efried: in current functional test cases the create server are created on the single host
14:17:09 efried bhagyashris: You're trying to make a functional test that does exactly what you did in your live environment. That is, you create the provider, mark it as a sharing provider, give it DISK_GB inventory, and associate it via aggregate with the compute node provider. Then the purpose of the patch we're testing is that we fixed it exactly so the scheduler knows to pick that sharing provider for the DISK_GB.
14:18:36 fishbone__ hello all; I am receiving a strange error when launching windows instances: ERROR cloudbaseinit.init pywintypes.com_error: (-2147352567, 'Exception occurred.', (0, 'Session', 'Access is denied. ', None, 0, -2147024891), None)

Earlier   Later