| *** lixiaoy1 has joined #openstack-cinder | 00:32 | |
| *** lixiaoy1 has quit IRC | 01:02 | |
| *** lbragstad has quit IRC | 01:07 | |
| *** hedvig_01 has quit IRC | 01:25 | |
| *** takamatsu has quit IRC | 01:37 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 01:40 | |
| *** Dinesh_Bhor has quit IRC | 01:56 | |
| openstackgerrit | zhongshengping proposed openstack/cinder master: Change openstack-dev to openstack-discuss https://review.openstack.org/621823 | 01:59 |
|---|---|---|
| openstackgerrit | Yingxin Cheng proposed openstack/cinder master: [wip] Cinder RSD Driver https://review.openstack.org/621465 | 02:17 |
| *** bhagyashris_ has joined #openstack-cinder | 02:17 | |
| *** lixiaoy1 has joined #openstack-cinder | 02:17 | |
| openstackgerrit | Brin Zhang proposed openstack/cinder master: Change openstack-dev to openstack-discuss https://review.openstack.org/621841 | 02:21 |
| *** itlinux has joined #openstack-cinder | 02:23 | |
| openstackgerrit | Brin Zhang proposed openstack/cinder-specs master: Change openstack-dev to openstack-discuss https://review.openstack.org/621850 | 02:30 |
| *** itlinux has quit IRC | 02:33 | |
| *** itlinux has joined #openstack-cinder | 02:34 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 02:36 | |
| *** itlinux has quit IRC | 02:36 | |
| *** mhen has quit IRC | 02:37 | |
| *** mhen has joined #openstack-cinder | 02:41 | |
| *** psachin has joined #openstack-cinder | 02:48 | |
| openstackgerrit | Jason SUN proposed openstack/python-cinderclient master: Change openstack-dev to openstack-discuss https://review.openstack.org/621882 | 02:51 |
| *** imacdonn has quit IRC | 02:52 | |
| *** imacdonn has joined #openstack-cinder | 02:52 | |
| *** jmlowe has joined #openstack-cinder | 03:22 | |
| *** dklyle has joined #openstack-cinder | 03:23 | |
| *** david-lyle has quit IRC | 03:25 | |
| *** brinzhang has joined #openstack-cinder | 03:25 | |
| *** dave-mccowan has quit IRC | 03:53 | |
| *** mriedem_away has quit IRC | 03:58 | |
| *** whoami-rajat has joined #openstack-cinder | 04:12 | |
| *** itlinux has joined #openstack-cinder | 04:14 | |
| *** Nel1x has quit IRC | 04:15 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Tests : notifier problem with backup https://review.openstack.org/618670 | 04:34 |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of api-ref samples https://review.openstack.org/619676 | 04:37 |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of volumes api-ref samples https://review.openstack.org/621478 | 04:39 |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: Automate generation of API versions api-ref samples https://review.openstack.org/620033 | 04:40 |
| *** wanghao has joined #openstack-cinder | 05:47 | |
| *** Dinesh_Bhor has quit IRC | 05:50 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 05:54 | |
| *** gcb_ has joined #openstack-cinder | 06:06 | |
| *** gcb_ has quit IRC | 06:11 | |
| *** Liang__ has joined #openstack-cinder | 06:44 | |
| *** Luzi has joined #openstack-cinder | 06:46 | |
| *** takamatsu has joined #openstack-cinder | 06:53 | |
| openstackgerrit | Yingxin Cheng proposed openstack/cinder master: [wip] Cinder RSD Driver https://review.openstack.org/621465 | 07:05 |
| *** brinzhang has quit IRC | 07:09 | |
| *** brinzhang has joined #openstack-cinder | 07:09 | |
| *** pcaruana has joined #openstack-cinder | 07:10 | |
| openstackgerrit | Boxiang Zhu proposed openstack/cinder master: Change a function name of swift backup https://review.openstack.org/622039 | 07:12 |
| *** arne_wiebalck has joined #openstack-cinder | 07:29 | |
| *** dpawlik has joined #openstack-cinder | 07:29 | |
| *** rcernin has quit IRC | 07:38 | |
| *** rosmaita has quit IRC | 08:00 | |
| *** rosmaita has joined #openstack-cinder | 08:01 | |
| openstackgerrit | 98k proposed openstack/python-brick-cinderclient-ext master: Change openstack-dev to openstack-discuss https://review.openstack.org/622085 | 08:03 |
| *** takamatsu has quit IRC | 08:15 | |
| *** gkadam has joined #openstack-cinder | 08:22 | |
| *** ianychoi has quit IRC | 08:30 | |
| *** ianychoi has joined #openstack-cinder | 08:30 | |
| *** trident has quit IRC | 08:31 | |
| *** trident has joined #openstack-cinder | 08:32 | |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume update/del policy https://review.openstack.org/617543 | 08:35 |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume action policy https://review.openstack.org/617544 | 08:35 |
| *** hoonetorg has quit IRC | 08:45 | |
| *** takamatsu has joined #openstack-cinder | 08:46 | |
| *** lixiaoy1 has quit IRC | 08:53 | |
| *** hoonetorg has joined #openstack-cinder | 08:58 | |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume metadata policy https://review.openstack.org/619977 | 09:11 |
| openstackgerrit | Yikun Jiang proposed openstack/cinder master: Add policy test for volume extend/retype/update_ro policy https://review.openstack.org/619978 | 09:11 |
| *** takamatsu has quit IRC | 09:14 | |
| *** takamatsu has joined #openstack-cinder | 09:16 | |
| lyarwood | https://review.openstack.org/#/q/topic:bug/1786327+status:open - is ready for additional stable core review now if we have any online this morning | 09:19 |
| *** lbragstad has joined #openstack-cinder | 09:24 | |
| *** mszwed has joined #openstack-cinder | 09:33 | |
| mszwed | smcginnis, jungleboyj to summerize yesterdays conversation: do we need to enable all the tests, or current set of tests is enough? | 09:36 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinderlib functional tests https://review.openstack.org/620671 | 09:40 |
| *** e0ne has joined #openstack-cinder | 09:44 | |
| *** takamatsu has quit IRC | 09:49 | |
| *** takamatsu has joined #openstack-cinder | 09:50 | |
| *** bhagyashris_ has quit IRC | 09:54 | |
| *** takamatsu has quit IRC | 10:13 | |
| *** erlon has joined #openstack-cinder | 10:25 | |
| gmann | jungleboyj: i sent email to you regarding the cinder health check on nov 6th. it will be great if you can respond to that. thanks. | 10:29 |
| *** luizbag has joined #openstack-cinder | 10:41 | |
| *** ganso has joined #openstack-cinder | 10:45 | |
| *** kholkina has joined #openstack-cinder | 10:48 | |
| *** takamatsu has joined #openstack-cinder | 10:50 | |
| *** Dinesh_Bhor has quit IRC | 10:52 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 10:53 | |
| *** Dinesh_Bhor has quit IRC | 11:00 | |
| *** takamatsu has quit IRC | 11:04 | |
| *** takamatsu has joined #openstack-cinder | 11:05 | |
| *** luizbag has quit IRC | 11:08 | |
| *** luizbag has joined #openstack-cinder | 11:22 | |
| *** e0ne has quit IRC | 11:35 | |
| *** e0ne has joined #openstack-cinder | 11:38 | |
| *** lumir_ has joined #openstack-cinder | 11:53 | |
| *** wanghao has quit IRC | 12:05 | |
| *** dave-mccowan has joined #openstack-cinder | 12:13 | |
| openstackgerrit | Merged openstack/cinder-specs master: Change openstack-dev to openstack-discuss https://review.openstack.org/621850 | 12:13 |
| *** lbragstad has quit IRC | 12:21 | |
| *** lbragstad has joined #openstack-cinder | 12:23 | |
| *** lbragstad has quit IRC | 12:23 | |
| *** lbragstad has joined #openstack-cinder | 12:24 | |
| *** lbragsta_ has joined #openstack-cinder | 12:26 | |
| *** lbragsta_ has quit IRC | 12:31 | |
| *** lbragstad has quit IRC | 12:31 | |
| *** lbragstad has joined #openstack-cinder | 12:38 | |
| *** tpsilva has joined #openstack-cinder | 12:44 | |
| *** brinzhang has quit IRC | 12:46 | |
| openstackgerrit | Merged openstack/cinder master: Make lower constraints CI works well https://review.openstack.org/614943 | 12:48 |
| *** abishop has joined #openstack-cinder | 12:49 | |
| *** psachin has quit IRC | 12:59 | |
| *** udesale has joined #openstack-cinder | 13:00 | |
| *** arne_wiebalck_ has joined #openstack-cinder | 13:01 | |
| *** MrRon has joined #openstack-cinder | 13:04 | |
| *** MrRon has left #openstack-cinder | 13:05 | |
| *** Liang__ has quit IRC | 13:13 | |
| *** Liang__ has joined #openstack-cinder | 13:14 | |
| geguileo | erlon: I see some tests have run, but I can't see the logs. Page doesn't load in the link from gerrit, and replacing with the IP you told me gives me a File Not Found | 13:24 |
| openstackgerrit | Eric Harney proposed openstack/os-brick master: Fix NFS "already mounted" detection https://review.openstack.org/609543 | 13:38 |
| *** psachin has joined #openstack-cinder | 13:38 | |
| erlon | geguileo, great :( | 13:46 |
| erlon | we will work on that, and let you know | 13:46 |
| erlon | sorry about it | 13:46 |
| geguileo | erlon: I may be doing it wrong since I'm running on a foggy brain today XD | 13:46 |
| geguileo | erlon: thanks | 13:46 |
| *** eharney has joined #openstack-cinder | 13:47 | |
| *** psachin has quit IRC | 13:51 | |
| *** arne_wiebalck_ has quit IRC | 13:53 | |
| openstackgerrit | Merged openstack/cinder master: Change openstack-dev to openstack-discuss https://review.openstack.org/621823 | 13:53 |
| openstackgerrit | Merged openstack/python-cinderclient master: Change openstack-dev to openstack-discuss https://review.openstack.org/621882 | 13:55 |
| *** finix has joined #openstack-cinder | 14:01 | |
| *** finixlei has joined #openstack-cinder | 14:01 | |
| *** finixlei has quit IRC | 14:01 | |
| *** udesale has quit IRC | 14:08 | |
| *** Liang__ has quit IRC | 14:10 | |
| openstackgerrit | Jay Rubenstein proposed openstack/cinder master: SF: Handle qos values on extend volume https://review.openstack.org/606959 | 14:10 |
| *** irclogbot_0 has quit IRC | 14:15 | |
| *** trident has quit IRC | 14:18 | |
| *** trident has joined #openstack-cinder | 14:21 | |
| *** mriedem has joined #openstack-cinder | 14:33 | |
| *** zul has quit IRC | 14:37 | |
| *** zul has joined #openstack-cinder | 14:41 | |
| *** udesale has joined #openstack-cinder | 14:53 | |
| *** irclogbot_0 has joined #openstack-cinder | 14:54 | |
| *** eharney has quit IRC | 14:56 | |
| *** zul has quit IRC | 14:58 | |
| *** lixiaoy1 has joined #openstack-cinder | 15:01 | |
| *** dpawlik has quit IRC | 15:04 | |
| jungleboyj | gmann: Sorry for not seeing that e-mail. It went to my review inbox which has been horribly backed up. I see it now and will respond. | 15:05 |
| jungleboyj | mszwed: For which CI are you talking? | 15:05 |
| openstackgerrit | Keith Berger proposed openstack/os-brick stable/pike: Fixing FC scanning https://review.openstack.org/622348 | 15:06 |
| *** mriedem is now known as mriedem_afk | 15:16 | |
| *** itlinux has quit IRC | 15:18 | |
| *** gnufied has quit IRC | 15:22 | |
| *** gnufied has joined #openstack-cinder | 15:24 | |
| *** jmlowe has quit IRC | 15:26 | |
| *** Luzi has quit IRC | 15:26 | |
| *** gnufied has quit IRC | 15:30 | |
| *** finix has quit IRC | 15:38 | |
| *** gnufied has joined #openstack-cinder | 15:40 | |
| *** zul has joined #openstack-cinder | 15:40 | |
| *** dpawlik has joined #openstack-cinder | 15:40 | |
| *** dpawlik has quit IRC | 15:45 | |
| *** jmlowe has joined #openstack-cinder | 15:50 | |
| *** jmlowe has quit IRC | 15:57 | |
| *** jmlowe has joined #openstack-cinder | 15:57 | |
| *** lixiaoy1 has quit IRC | 16:01 | |
| openstackgerrit | Yingxin Cheng proposed openstack/cinder master: [wip] Cinder RSD Driver https://review.openstack.org/621465 | 16:02 |
| *** lemko has joined #openstack-cinder | 16:03 | |
| lemko | Hi. Detaching volumes is extremely slow and often fails with multipath for me. I'm using queens. When I inspect the process which is stucked, it is "multipath -f MULTIPATH". | 16:04 |
| lemko | Attaching volumes is fine. | 16:05 |
| lemko | But detaching them is all the time a problem. | 16:05 |
| lemko | What could be the reason of this? | 16:05 |
| lemko | nova-compute from what I understand is in charge of the volume part on the compute and is doing multipath -f ... | 16:06 |
| *** pcaruana has quit IRC | 16:12 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: PowerMax driver - rebranding VMAX to PowerMax https://review.openstack.org/620659 | 16:17 |
| *** itlinux has joined #openstack-cinder | 16:18 | |
| *** eharney has joined #openstack-cinder | 16:24 | |
| *** gkadam has quit IRC | 16:28 | |
| *** itlinux has quit IRC | 16:29 | |
| *** itlinux has joined #openstack-cinder | 16:36 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX Driver - VMAX OS Upgrade Bug https://review.openstack.org/598961 | 16:36 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinderlib functional tests https://review.openstack.org/620671 | 16:41 |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX Driver - Failover Unisphere Support https://review.openstack.org/570401 | 16:42 |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: PowerMax driver - rebranding VMAX to PowerMax https://review.openstack.org/620659 | 16:48 |
| *** mriedem_afk is now known as mriedem | 16:50 | |
| openstackgerrit | Merged openstack/cinder stable/rocky: lvm: Avoid premature calls to terminate_connection for muiltiattach vols https://review.openstack.org/618472 | 16:57 |
| *** hedvig_01 has joined #openstack-cinder | 16:59 | |
| mszwed | jungleboyj: Mellanmox CI, SPDK driver tests | 17:03 |
| mszwed | *Mellanox | 17:04 |
| jungleboyj | mszwed: Well, I don't know that we came to a conclusion. | 17:06 |
| jungleboyj | It seems that we are inconsistent across the CIs that are running. | 17:06 |
| *** takamatsu has quit IRC | 17:06 | |
| jungleboyj | I have added discussing the issue to our mid-cycle meet-up. There are many issues around CI that need to be addressed. | 17:06 |
| mszwed | ok | 17:06 |
| mszwed | Could you remind me when it takes place? | 17:07 |
| mszwed | ok, I see it now: February 4th, 2019 | 17:09 |
| mszwed | (proposed) | 17:10 |
| hedvig_01 | Hi, | 17:10 |
| *** pcaruana has joined #openstack-cinder | 17:12 | |
| hedvig_01 | The test_volume_boot_pattern test Creates a volume snapshot while the instance is running and Boot an additional instance from the new snapshot based volume. Is IO quiescing is taken care of? | 17:12 |
| *** e0ne has quit IRC | 17:13 | |
| *** luizbag has quit IRC | 17:13 | |
| hedvig_01 | I picked it up from the description of the test case. I'm looking for information about the IO quiescing. | 17:13 |
| hedvig_01 | ref: https://github.com/openstack/tempest/blob/master/tempest/scenario/test_volume_boot_pattern.py | 17:14 |
| mszwed | jungleboyj: this meet-up is actually after code freez for new drivers. Won't it be too late to get SPDK drivers to Stein release? We will probably need some decision before that, so that SPDK drivers can be merged before freeze. | 17:14 |
| openstackgerrit | Filippo Inzaghi proposed openstack/os-brick master: Change openstack-dev to openstack-discuss https://review.openstack.org/622441 | 17:15 |
| *** dpawlik has joined #openstack-cinder | 17:15 | |
| jungleboyj | mszwed: Oh, good point. I guess we need to talk about this sooner then. | 17:19 |
| jungleboyj | Can you add it to the agenda for tomorrow's weekly meeting? | 17:19 |
| eharney | hedvig_01: that test doesn't perform quiescing in general since that test just does a standard cinder snapshot which is only performed on the storage backend | 17:20 |
| openstackgerrit | Nate Johnston proposed openstack/cinder master: Replace neutron-grenade job with grenade-py3 https://review.openstack.org/622448 | 17:21 |
| hedvig_01 | @eharney : I agree it just does a standard cinder snapshot. However, it's a running instance and IOs are performed as a part of the same testcase while taking snapshot. The filesystems has to be quiesced before this snapshot operation right? | 17:24 |
| hedvig_01 | I believe KVM provides that capability. | 17:25 |
| jungleboyj | mszwed: I added it to tomorrow's meeting. Can you attend? | 17:30 |
| eharney | hedvig_01: KVM provides the capability but we don't, in general, leverage it in Cinder, at least from the Cinder snapshot API. There is a way to do it via Nova instance snapshots of Cinder volumes, I think. | 17:30 |
| hedvig_01 | @eharney: I believe the test is failing for us because IOs are not quiesced before taking a snapshot that result in an inconsistent filesystem. Do you have any suggestions for the same ? | 17:32 |
| hedvig_01 | This is the only test out of 253 tests which fails at times for hedvig. | 17:33 |
| hedvig_01 | We pass all the other tests without adding any regex to skip tests | 17:33 |
| hedvig_01 | We as a storage vendor do not take care of quiescing and let the hypervisor quiesce it for us | 17:34 |
| *** francois has quit IRC | 17:34 | |
| eharney | hedvig_01: why would lack of quiescing cause a failure in that test? | 17:35 |
| hedvig_01 | @eharney: Write content , Create a volume snapshot while the instance is running , Boot an additional instance from the new snapshot based volume , Check written content in the instance booted from snapshot | 17:36 |
| hedvig_01 | Checking the written content might lead in incorrect data since IO was not quiesced | 17:37 |
| *** jmlowe has quit IRC | 17:38 | |
| *** jdillaman has quit IRC | 17:43 | |
| jungleboyj | mszwed: I don't know that there is a lot of conversation here given the wiki: https://wiki.openstack.org/wiki/Cinder/tested-3rdParty-drivers#Volume_.26_Connector_Drivers | 17:43 |
| *** luizbag has joined #openstack-cinder | 17:44 | |
| *** pcaruana has quit IRC | 17:56 | |
| *** e0ne has joined #openstack-cinder | 18:13 | |
| *** jdillaman has joined #openstack-cinder | 18:18 | |
| _hemna | jungleboyj: is zuul always going to puke on these backports? https://review.openstack.org/#/c/622348/ | 18:21 |
| _hemna | it's a cherry pick, should pass | 18:22 |
| jungleboyj | _hemna: No, that should pass. | 18:23 |
| _hemna | I issued a recheck, but it doesn't look like it got picked up | 18:24 |
| *** e0ne has quit IRC | 18:30 | |
| imacdonn | recheck is in progress | 18:34 |
| erlon | hedvig_01, this test is known to fail | 18:34 |
| erlon | hedvig_01, IIRC its not only in the 'mandatory' list of tests to be run by CIs. We only run tempest.api.volume.* | 18:35 |
| hedvig_01 | @erlon: I didn't get you. Are you implying it's known to fail and we skip it? tempest.api.volume.* comprises of test_volume_boot_pattern | 18:39 |
| mszwed | jungleboyj: thanks, i should be able to make it to the meeting | 18:39 |
| imacdonn | _hemna: ICYMI: your recheck is in progress .. it hasn't failed so-far ;) | 18:39 |
| _hemna | ok, I just hadn't seen a reset of the vote on the review yet | 18:40 |
| imacdonn | ah. I don't know if zuul is supposed to do that .. I generally watch the zuul.o.o dashboard, and expect it to update the review when it's finished | 18:41 |
| *** dhineshb has joined #openstack-cinder | 18:43 | |
| eharney | test_volume_boot_pattern should not be skipped | 18:48 |
| erlon | hedvig_01, yep, was just messing with a bug we hit in the past | 18:49 |
| erlon | hedvig_01, currently it is passing in most of other CIs | 18:49 |
| imacdonn | hedvig_01: do you know that I/O is the problem, or is it a guess? Does your backend refuse to make the snapshot if the volume us busy? | 18:50 |
| *** jmlowe has joined #openstack-cinder | 18:50 | |
| erlon | hedvig_01, http://logs.openstack.org/17/614217/2/check/tempest-full/10c2181/testr_results.html.gz | 18:50 |
| hedvig_01 | @imacdonn: We do not refuse to make a snapshot if the volume is busy. If the IO's are not quiesced before it reaches backend, there will be incorrect data. All of the data might not be flushed before the snapshot was taken. A clone from this snapshot will lead to incorrect data. | 18:52 |
| eharney | Cinder drivers have to be able to create snapshots for volumes that are busy | 18:53 |
| hedvig_01 | As a storage backend, we expect the hypervisor to take care of quiescing data. | 18:53 |
| imacdonn | hedvig_01: OK .. but I don't think the tempest test does any verification of the snapshot content, so that's probably not why it's failing the test.... ? | 18:53 |
| eharney | hedvig_01: ^ this is an incorrect expectation for Cinder drivers | 18:53 |
| hedvig_01 | Write content , Create a volume snapshot while the instance is running , Boot an additional instance from the new snapshot based volume , Check written content in the instance booted from snapshot | 18:53 |
| hedvig_01 | This is the test description | 18:53 |
| hedvig_01 | the data is being verified from the clone | 18:54 |
| imacdonn | oh, I see .. hmm | 18:54 |
| _hemna | I believe you still have to have agents inside the VM for the volumes/io to be quiesced, and we don't require that | 18:54 |
| *** luizbag has quit IRC | 18:55 | |
| hedvig_01 | @eharney: ^ this is an incorrect expectation for Cinder drivers. Are you talking about quiescing? | 18:55 |
| eharney | not only do we not require it, we don't generally use them even if you do have them | 18:55 |
| _hemna | there is no guarantee from a cinder perspective that io is quiesced on the vm | 18:55 |
| imacdonn | I think hedvig_01's point is that the snapshot should be taken via nova, not directly in cinder .. so that nova can try to quiesce the instance | 18:55 |
| eharney | imacdonn: and it can do that if you use nova instance snapshots, iirc. but not cinder volume snapshots | 18:55 |
| hedvig_01 | Yes | 18:55 |
| _hemna | imacdonn: nova can't unless there is an agent installed inside the vm afaik | 18:55 |
| hedvig_01 | The subsequent clone created from that snapshot has an inconsistent filesystem if not quiesced by nova | 18:56 |
| eharney | hedvig_01: it's not inconsistent. it's crash-consistent. | 18:57 |
| hedvig_01 | https://blueprints.launchpad.net/cinder/+spec/quiesced-snapshots-with-qemu-guest-agent I ran into this | 18:57 |
| eharney | test_volume_boot_pattern assumes that crash consistency is enough. and it should be... | 18:58 |
| hedvig_01 | Since we are verifying content, we need app consistent snapshots | 19:00 |
| eharney | i don't think you do. | 19:00 |
| hedvig_01 | crash consistent snapshots does not take care of data in memory or any pending I/O operations. | 19:00 |
| hedvig_01 | We are doing IO's in this test and then snapshotting | 19:01 |
| eharney | which is how cinder volume drivers are used when deployed | 19:01 |
| *** N3l1x has joined #openstack-cinder | 19:01 | |
| imacdonn | it does do a sync | 19:02 |
| eharney | it may be the case that test_volume_boot_pattern is cheating on some assumptions around caching on the attachment, i'm not sure how it configures attachments in that test | 19:02 |
| imacdonn | https://github.com/openstack/tempest/blob/master/tempest/scenario/manager.py#L661 | 19:02 |
| hedvig_01 | how are you handling the quiescing of data? unless you have the agent running on the instance | 19:02 |
| eharney | cinder snapshots guarantee a point-in-time snapshot | 19:03 |
| eharney | they do not guarantee any kind of quiescing | 19:03 |
| *** e0ne has joined #openstack-cinder | 19:03 | |
| eharney | they typically do not even have a mechanism to ask for quiescing, since that needs to be driven by the hypervisor, and cinder snapshots don't go through the hypervisor | 19:04 |
| hedvig_01 | okay. then this test shouldnt be verifying the content | 19:05 |
| eharney | i disagree | 19:05 |
| hedvig_01 | Content can be verified only when we ensure that all the data is flushed | 19:05 |
| hedvig_01 | how can storage show the data which is not flushed at the time of snapshot? | 19:05 |
| eharney | you can verify that without having a quiescing mechanism, whether this test does, i'm not sure | 19:05 |
| imacdonn | hedvig_01: have you determined what happened when it fails? Does the VM fail to boot? Is the filesystem corrupt? Is the timestamp file missing ? | 19:06 |
| hedvig_01 | I didnt get you . --- you can verify that without having a quiescing mechanism, whether this test does, i'm not sure | 19:06 |
| hedvig_01 | Is the timestamp file missing | 19:06 |
| hedvig_01 | yes, timestamp file is missing | 19:06 |
| imacdonn | hmm, strange ... must be some sort of caching .. the 'sync' should make sure that I/O has been flushed from the VM OS perspective | 19:07 |
| *** itlinux has quit IRC | 19:07 | |
| eharney | there is usually caching on the nova side, writeback by default, this test looks like it might be assuming that there is no caching there, but like i said, i'm not sure if it makes any attempt to configure that etc | 19:08 |
| *** erlon has quit IRC | 19:09 | |
| imacdonn | if that was the issue, it seems like it'd be affecting lots of other cinder drivers ? | 19:09 |
| eharney | yep | 19:09 |
| hedvig_01 | Yes, Unless we ensure that it does all of that, we cannot expect this to pass | 19:09 |
| eharney | in all of the millions and millions of test runs we have run for years | 19:09 |
| hedvig_01 | and It's not just hedvig. I've seen the reports of serveral other vendors skipping it using regex. I had raised this issue last week. | 19:10 |
| eharney | hedvig_01: how often is this test failing for you exactly? | 19:10 |
| eharney | yeah, other vendors shouldn't be turning off tests like this, but i don't know how to chase that problem efficiently right now | 19:10 |
| *** itlinux has joined #openstack-cinder | 19:10 | |
| eharney | part of this comes down to vendors just deciding "yes, i want a functioning cinder driver so i'll run tests" | 19:11 |
| hedvig_01 | it fails 90% of the time when I run through jenkins . On my local devstack setup, I run it in loop. It passes at times it fails at times | 19:12 |
| imacdonn | _hemna: Your recheck failed too, but looks like for a different reason this time .. urgh | 19:12 |
| eharney | does your driver do anything different enough from other drivers to explain why it would fail 90% of the time? that smells more like a bug in the driver... | 19:13 |
| hedvig_01 | @eharney: I agree. That's the reason we are not skipping any tests. I'm glad openstack has such a robust setup. It makes things easy for us as well. But this is one issue we cannot do anything about. We rely on the hypervisor to take care of this | 19:13 |
| hedvig_01 | No. why it's not a bug is because the data could have been flushed. you cannot predict | 19:14 |
| _hemna | bah | 19:14 |
| _hemna | once more with feeling? | 19:14 |
| hedvig_01 | 90% is just my random number. if I check my jenkins setup, I see ony 7 to 8 pass so far | 19:14 |
| imacdonn | I guess so :) | 19:15 |
| hedvig_01 | @_hemna: We need to ensure we are covered for all corner case scenarios for a valid test. | 19:16 |
| *** itlinux has quit IRC | 19:23 | |
| *** jmlowe has quit IRC | 19:23 | |
| *** dpawlik has quit IRC | 19:30 | |
| *** dpawlik has joined #openstack-cinder | 19:31 | |
| hedvig_01 | Based on the above conversation, is it okay to skip this test? | 19:35 |
| *** itlinux has joined #openstack-cinder | 19:53 | |
| *** erlon has joined #openstack-cinder | 19:53 | |
| *** e0ne has quit IRC | 19:58 | |
| *** jmlowe has joined #openstack-cinder | 20:23 | |
| *** jmlowe has quit IRC | 20:33 | |
| *** hoonetorg has quit IRC | 20:39 | |
| *** takamatsu has joined #openstack-cinder | 20:44 | |
| *** eharney has quit IRC | 20:53 | |
| *** hoonetorg has joined #openstack-cinder | 20:57 | |
| *** dpawlik has quit IRC | 21:01 | |
| *** eharney has joined #openstack-cinder | 21:04 | |
| *** jmlowe has joined #openstack-cinder | 21:12 | |
| *** ganso has quit IRC | 21:28 | |
| *** tpsilva has quit IRC | 21:32 | |
| openstackgerrit | Dave McCowan proposed openstack/cinder master: Prevent logging of encryption key value https://review.openstack.org/622606 | 21:37 |
| *** eharney has quit IRC | 21:40 | |
| *** abishop has quit IRC | 22:09 | |
| openstackgerrit | Matt Riedemann proposed openstack/python-cinderclient master: WIP: Add support for microversion 3.56 (volume re-image) https://review.openstack.org/622614 | 22:22 |
| *** takamatsu has quit IRC | 22:29 | |
| *** erlon has quit IRC | 22:33 | |
| *** dave-mccowan has quit IRC | 22:33 | |
| *** rcernin has joined #openstack-cinder | 22:34 | |
| openstackgerrit | Matt Riedemann proposed openstack/python-cinderclient master: Cleanup the home page https://review.openstack.org/622617 | 22:34 |
| openstackgerrit | Matt Riedemann proposed openstack/python-cinderclient master: WIP: Add support for microversion 3.56 (volume re-image) https://review.openstack.org/622614 | 22:34 |
| *** mriedem has quit IRC | 22:44 | |
| *** itlinux has quit IRC | 22:45 | |
| *** rcernin_ has joined #openstack-cinder | 22:45 | |
| *** rcernin has quit IRC | 22:45 | |
| *** jmlowe has quit IRC | 22:46 | |
| *** jmlowe has joined #openstack-cinder | 22:48 | |
| *** N3l1x has quit IRC | 22:49 | |
| openstackgerrit | Nate Johnston proposed openstack/cinder master: Replace neutron-grenade job with grenade-py3 https://review.openstack.org/622448 | 23:06 |
| *** rcernin_ has quit IRC | 23:12 | |
| *** etp has quit IRC | 23:12 | |
| *** etp has joined #openstack-cinder | 23:13 | |
| *** rcernin has joined #openstack-cinder | 23:13 | |
| *** udesale has quit IRC | 23:40 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!