*** mtanino has quit IRC | 00:23 | |
*** catintheroof has quit IRC | 00:29 | |
*** crushil has quit IRC | 01:57 | |
*** dsariel has quit IRC | 02:09 | |
*** gouthamr has quit IRC | 02:45 | |
*** gouthamr has joined #openstack-manila | 02:46 | |
*** crushil has joined #openstack-manila | 02:50 | |
*** crushil has quit IRC | 02:56 | |
*** crushil has joined #openstack-manila | 02:58 | |
*** waj334_ has joined #openstack-manila | 03:04 | |
*** dmellado_ has joined #openstack-manila | 03:06 | |
*** draynium_ has joined #openstack-manila | 03:06 | |
*** dmellado has quit IRC | 03:07 | |
*** waj334 has quit IRC | 03:07 | |
*** smcginnis has quit IRC | 03:07 | |
*** crushil has quit IRC | 03:07 | |
*** bswartz has quit IRC | 03:07 | |
*** gcb has quit IRC | 03:07 | |
*** draynium has quit IRC | 03:07 | |
*** smcginnis_ has joined #openstack-manila | 03:07 | |
*** waj334_ is now known as waj334 | 03:07 | |
*** smcginnis_ is now known as smcginnis | 03:07 | |
*** crushil has joined #openstack-manila | 03:08 | |
*** gcb has joined #openstack-manila | 03:08 | |
openstackgerrit | TommyLike proposed openstack/manila master: Fix: [LVM] Can't extend share when snapshots exist https://review.openstack.org/431315 | 03:14 |
---|---|---|
*** crushil has quit IRC | 03:15 | |
*** gouthamr has quit IRC | 04:01 | |
openstackgerrit | TommyLike proposed openstack/manila master: Fix: [LVM] Can't extend share when snapshots exist https://review.openstack.org/431315 | 04:49 |
openstackgerrit | TommyLike proposed openstack/manila master: Fix: [LVM] Can't extend share when snapshots exist https://review.openstack.org/431315 | 04:59 |
openstackgerrit | Danny Al-Gaaf proposed openstack/manila master: Adds manila-manage 'db purge' command to man page https://review.openstack.org/433676 | 05:11 |
openstackgerrit | Ghanshyam Mann proposed openstack/manila master: Fix to use correct config options for network_for_ssh https://review.openstack.org/434070 | 05:13 |
*** faizy has joined #openstack-manila | 05:28 | |
*** sandanar has joined #openstack-manila | 05:36 | |
*** wiebalck has joined #openstack-manila | 06:05 | |
openstackgerrit | Danny Al-Gaaf proposed openstack/manila master: doc: verify all rst files https://review.openstack.org/434086 | 06:08 |
*** lpetrut has joined #openstack-manila | 06:09 | |
*** pcaruana has joined #openstack-manila | 06:20 | |
openstackgerrit | TommyLike proposed openstack/manila master: Add shrink share in LVM driver https://review.openstack.org/433489 | 06:21 |
*** dsariel has joined #openstack-manila | 06:27 | |
*** wiebalck has quit IRC | 06:57 | |
openstackgerrit | TommyLike proposed openstack/manila master: Fix: LVM driver doesn't resize lv while is required https://review.openstack.org/433559 | 07:07 |
*** kaisers has joined #openstack-manila | 07:12 | |
*** lpetrut has quit IRC | 07:22 | |
*** jprovazn has joined #openstack-manila | 07:25 | |
*** tommylikehu_ has joined #openstack-manila | 07:30 | |
*** tommylikehu_ has quit IRC | 07:35 | |
*** lpetrut has joined #openstack-manila | 07:37 | |
*** rraja has joined #openstack-manila | 07:38 | |
*** lpetrut has quit IRC | 07:48 | |
*** tuanluong has joined #openstack-manila | 08:35 | |
openstackgerrit | dengshaolin proposed openstack/python-manilaclient master: Raise TypeError in manilaclient/common/httpclient.py https://review.openstack.org/434140 | 08:39 |
vponomaryov | bswartz: https://github.com/openstack/manila/blob/4cfdbea4/manila_tempest_tests/tests/api/base.py#L799-L826 | 08:42 |
vponomaryov | bswartz: it is used for dynamic creation of share types in lots of places for share replication/migration and other kind of tests... | 08:42 |
openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Upgrade share group snapshot members DB model https://review.openstack.org/433854 | 08:51 |
openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Upgrade share group snapshot members DB model https://review.openstack.org/433854 | 08:55 |
*** akapil has joined #openstack-manila | 09:02 | |
*** lpetrut has joined #openstack-manila | 09:17 | |
*** lpetrut has quit IRC | 09:26 | |
*** lpetrut has joined #openstack-manila | 09:28 | |
*** gcb has quit IRC | 09:31 | |
*** gcb has joined #openstack-manila | 09:33 | |
*** dsariel has quit IRC | 09:43 | |
*** dsariel has joined #openstack-manila | 09:54 | |
*** tuanluong has quit IRC | 10:14 | |
*** wiebalck has joined #openstack-manila | 10:22 | |
*** ociuhandu has joined #openstack-manila | 10:28 | |
*** ganso has joined #openstack-manila | 10:47 | |
*** wiebalck has quit IRC | 11:22 | |
*** yumiriam has joined #openstack-manila | 11:24 | |
*** crushil has joined #openstack-manila | 11:25 | |
*** tommylikehu_ has joined #openstack-manila | 11:40 | |
*** ociuhandu has quit IRC | 11:42 | |
*** gcb has quit IRC | 11:48 | |
*** rraja has quit IRC | 11:50 | |
*** ociuhandu has joined #openstack-manila | 11:55 | |
*** crushil has quit IRC | 12:06 | |
*** catintheroof has joined #openstack-manila | 12:36 | |
*** faizy has quit IRC | 12:39 | |
*** tommylikehu_ has quit IRC | 12:47 | |
*** ociuhandu has quit IRC | 12:48 | |
openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Upgrade share group snapshot members DB model https://review.openstack.org/433854 | 12:48 |
*** tommylikehu_ has joined #openstack-manila | 12:49 | |
*** tommylikehu_ has quit IRC | 12:51 | |
*** tommylikehu_ has joined #openstack-manila | 12:51 | |
*** gouthamr has joined #openstack-manila | 13:13 | |
*** faizy has joined #openstack-manila | 13:15 | |
*** gcb has joined #openstack-manila | 13:23 | |
*** eharney has joined #openstack-manila | 13:26 | |
openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Upgrade share group snapshot members DB model https://review.openstack.org/433854 | 13:30 |
*** bswartz has joined #openstack-manila | 13:38 | |
openstackgerrit | Jan Provaznik proposed openstack/manila-specs master: Add spec for user messages https://review.openstack.org/434277 | 13:40 |
*** xyang_ has joined #openstack-manila | 13:41 | |
*** dustins has joined #openstack-manila | 13:49 | |
openstackgerrit | Merged openstack/manila master: [Tempest] Refactor api/tests/admin/test_share_servers module https://review.openstack.org/433003 | 13:51 |
*** ociuhandu has joined #openstack-manila | 13:51 | |
*** tongli has joined #openstack-manila | 13:54 | |
*** xyang_ has quit IRC | 13:59 | |
*** xyang_ has joined #openstack-manila | 13:59 | |
*** xyang_ has quit IRC | 14:02 | |
*** crushil has joined #openstack-manila | 14:03 | |
vponomaryov | markstur: >>> https://review.openstack.org/#/c/425904/ <<< | 14:10 |
*** porrua has joined #openstack-manila | 14:15 | |
*** xyang_ has joined #openstack-manila | 14:18 | |
openstackgerrit | Merged openstack/manila master: Add Share Migration devref docs https://review.openstack.org/375771 | 14:18 |
openstackgerrit | Merged openstack/manila master: Improve HNAS driver coverage https://review.openstack.org/433610 | 14:24 |
*** dmellado_ is now known as dmellado | 14:27 | |
*** chlong has quit IRC | 14:37 | |
*** tongli has quit IRC | 14:40 | |
ganso | gouthamr: ping | 14:47 |
*** gcb has quit IRC | 14:49 | |
*** gcb has joined #openstack-manila | 14:49 | |
gouthamr | ganso: pong | 14:49 |
ganso | gouthamr: Hi Goutham. are you saying that this patch https://review.openstack.org/#/c/433613/ cannot be backported after Ocata release on Feb 23rd ? | 14:50 |
*** ociuhandu has quit IRC | 14:51 | |
*** dustins has quit IRC | 14:51 | |
bswartz | ganso: yikes! | 14:57 |
bswartz | you forgot to update the version before the branch? | 14:57 |
ganso | bswartz: yes | 14:57 |
bswartz | ganso: this seems sketchy to me | 14:59 |
bswartz | I don't agree that a reno would make this any better | 14:59 |
ganso | bswartz: indeed it wouldn't | 14:59 |
bswartz | ganso: can't we phrase it as a bug? | 15:00 |
bswartz | the fix is not trivial because it does impact user-observable functionality | 15:00 |
ganso | bswartz: we could, stating that it is shown the same version as newton with the same functionality | 15:00 |
bswartz | and the problem you're solving is that someone forgot to bump a version number | 15:00 |
bswartz | it feels like a bug | 15:00 |
ganso | bswartz: ok, if considered as a bug can be solved as a bug | 15:01 |
ganso | bswartz: can it be backported? | 15:01 |
bswartz | yeah I don't see why not | 15:01 |
ganso | bswartz: oh ok then, gouthamr's statement made me think it could not | 15:01 |
gouthamr | ganso: it's possible to backport imo... it's not a trivialfix | 15:03 |
bswartz | ganso: the trick will be when the fix merges compared to when we cut rc2 | 15:03 |
bswartz | if we can get a backport completed before rc2 then I think it's fine | 15:04 |
bswartz | otherwise, we'd want to wait until after official release to do the backport | 15:04 |
ganso | bswartz: oh it can be included in RC2? I was assuming this one could only be backported after the release on Feb 23rd | 15:04 |
bswartz | ganso: we would never do an RC2 just for such a small bug | 15:05 |
bswartz | but because we have a critical bug, other low-risk fixes can get combined with that fix for rc2 | 15:05 |
ganso | bswartz: cool I thought RC2 was restricted to include only the critical bugfix | 15:07 |
bswartz | ganso: we could do it that way too | 15:08 |
bswartz | IMO if we're doing the tag anyways, there's no reason not to include obviously low risk changes along with it | 15:08 |
*** cknight has joined #openstack-manila | 15:09 | |
*** zhonghua2 has joined #openstack-manila | 15:10 | |
bswartz | I'm also thinking of the windows driver fix | 15:10 |
bswartz | in fact we should get those bugs targetted so we don't forget | 15:11 |
bswartz | anyone have a link to the windows driver bug? | 15:11 |
*** zhonghua has quit IRC | 15:12 | |
bswartz | nm | 15:14 |
openstackgerrit | Merged openstack/manila master: Fix to use correct config options for network_for_ssh https://review.openstack.org/434070 | 15:14 |
openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: Update HNAS driver version history https://review.openstack.org/433613 | 15:14 |
*** JoseMello has joined #openstack-manila | 15:15 | |
*** mtanino has joined #openstack-manila | 15:16 | |
*** dustins has joined #openstack-manila | 15:36 | |
*** dustins_ has joined #openstack-manila | 15:42 | |
*** dustins has quit IRC | 15:45 | |
*** chlong has joined #openstack-manila | 16:04 | |
markstur | gouthamr: ping | 16:07 |
gouthamr | markstur: pong | 16:10 |
markstur | gouthamr: Those reno patches that kind of mess up the previous release info. Do we merge those as is or wait for some fix? They don't look right as-is. | 16:12 |
gouthamr | markstur: bswartz thinks we should merge and investigate | 16:14 |
bswartz | markstur: I don't care either way -- fixing broken renos isn't a priority for me -- that infrastructure appears to be fragile and owned by someone else | 16:17 |
*** dschoenb__ has joined #openstack-manila | 16:24 | |
openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Only return share host for admins https://review.openstack.org/433716 | 16:25 |
markstur | gouthamr: OK. Merge first and ask questions later. Is that how the saying goes? | 16:26 |
markstur | but it makes sense to me if we don't have some fix in the works that should go in first | 16:26 |
*** dustins_ has quit IRC | 16:27 | |
*** faizy has quit IRC | 16:28 | |
*** tommylikehu_ has quit IRC | 16:29 | |
*** crushil has quit IRC | 16:29 | |
*** crushil has joined #openstack-manila | 16:29 | |
openstackgerrit | Merged openstack/manila-ui master: Update reno for stable/ocata https://review.openstack.org/425907 | 16:32 |
*** sandanar has quit IRC | 16:36 | |
*** arnewiebalck_ has joined #openstack-manila | 16:38 | |
*** arnewiebalck has quit IRC | 16:40 | |
*** arnewiebalck_ is now known as arnewiebalck | 16:40 | |
*** akapil_ has joined #openstack-manila | 16:55 | |
*** shausy has joined #openstack-manila | 16:59 | |
*** akapil has quit IRC | 16:59 | |
*** akapil_ has quit IRC | 16:59 | |
*** crushil has quit IRC | 17:01 | |
*** crushil has joined #openstack-manila | 17:02 | |
openstackgerrit | Merged openstack/python-manilaclient master: Update reno for stable/ocata https://review.openstack.org/425904 | 17:07 |
*** faizy has joined #openstack-manila | 17:09 | |
*** jprovazn has quit IRC | 17:12 | |
*** dustins_ has joined #openstack-manila | 17:14 | |
*** dschoenb__ has quit IRC | 17:16 | |
*** lpetrut has quit IRC | 17:19 | |
*** shausy has quit IRC | 17:24 | |
*** shausy has joined #openstack-manila | 17:24 | |
*** chlong has quit IRC | 17:24 | |
*** porrua has quit IRC | 17:25 | |
*** dsariel has quit IRC | 17:29 | |
bswartz | ganso: ping | 17:34 |
ganso | bswartz: pong | 17:34 |
bswartz | ganso: I'm seeing migration tests fail on my system -- do migration tests required anything exltra? | 17:34 |
bswartz | extra* | 17:34 |
ganso | bswartz: which exactly? I just found a bug in migration | 17:35 |
bswartz | like do I need nova or neutron or cinder for migration tests to pass? | 17:35 |
bswartz | testing with manila lvm driver | 17:35 |
ganso | bswartz: a single test failed or all of them? | 17:35 |
bswartz | 3 tests | 17:35 |
ganso | bswartz: which ones? | 17:35 |
bswartz | manila_tempest_tests.tests.api.admin.test_migration.MigrationCancelNFSTest.test_migration_cancel_1_True[backend,positive] | 17:36 |
bswartz | manila_tempest_tests.tests.api.admin.test_migration.MigrationTwoPhaseNFSTest.test_migration_2phase_1_True[backend,positive] | 17:36 |
bswartz | manila_tempest_tests.tests.api.admin.test_migration.MigrationWithShareExtendingNFSTest.test_extend_on_migrated_share_1_True[backend,positive] | 17:36 |
ganso | bswartz: did test_migrate_share_to_no_snapshot_capability_share_type pass? | 17:36 |
bswartz | {0} manila_tempest_tests.tests.api.admin.test_migration.MigrationWithDifferentSnapshotSupportNFSTest.test_migrate_share_to_snapshot_capability_share_type_1_True [40.915820s] ... ok | 17:37 |
ganso | bswartz: o_O | 17:38 |
ganso | bswartz: well, gotta debug to find out the problem | 17:39 |
ganso | bswartz: if test_migrate_share_to_no_snapshot_capability_share_type passed, then it is probably working | 17:39 |
bswartz | ganso: more detail here: https://paste.fedoraproject.org/559026/71803561/ | 17:39 |
ganso | bswartz: it wouldn't have if configuration was wrong | 17:39 |
bswartz | gah! bad line wrapping is bad | 17:39 |
ganso | bswartz: and test_migrate_share_to_no_snapshot_capability_share_type is the one that has the bug I found lol | 17:39 |
ganso | bswartz: so it should not be passing at all | 17:39 |
bswartz | oh fpaste allows you to unwrap lines | 17:40 |
bswartz | +1 fpaste | 17:40 |
ganso | bswartz: gotta check the screen logs | 17:40 |
bswartz | ganso that's what I was afraid of | 17:40 |
bswartz | ganso: this tell you anything? https://paste.fedoraproject.org/559028/80530148/ | 17:42 |
bswartz | what is 50.50.50.50? | 17:42 |
*** porrua has joined #openstack-manila | 17:42 | |
bswartz | and what is 51.51.51.51? | 17:43 |
bswartz | are these IPs coming from tempest or devstack? | 17:43 |
ganso | bswartz: those are fake access rules | 17:44 |
ganso | bswartz: migration test attempts to migrate a share which has access rules | 17:44 |
ganso | bswartz: and the rules should be migrated as well | 17:44 |
ganso | bswartz: looking at that trace | 17:44 |
ganso | bswartz: migration_start is trying to cast the rules to readonly | 17:44 |
*** faizy has quit IRC | 17:45 | |
bswartz | fwiw all the exports on my test system are to 51.51.51.51 | 17:46 |
bswartz | anyone remember the command to 1 a single tempest test? | 17:46 |
bswartz | vponomaryov: ^ | 17:46 |
bswartz | to run a single tempest test | 17:46 |
ganso | bswartz: tox -eall-plugin manila_tempest_tests.tests.api.admin.test_migration -- --concurrency=1 | tee test2.log | 17:47 |
ganso | bswartz: it is invoking LVM driver update_access with parameters add_rules=[] and delete_rules=[] | 17:47 |
ganso | bswartz: which makes the driver remove all rules and add them again in RO | 17:47 |
ganso | bswartz: so, it is trying to remove 50.50.50.50, and it failed | 17:48 |
ganso | bswartz: it could have failed because it wasn't there | 17:48 |
bswartz | ganso: where is the flakiness most likely? | 17:49 |
bswartz | test error, manila error, or environment error? | 17:49 |
ganso | bswartz: I am thinking why it wouldnt be there... like some sort of concurrency | 17:49 |
bswartz | I ran with concurrency=1 | 17:49 |
ganso | bswartz: ok, no concurrency | 17:50 |
ganso | bswartz: well, I have no idea then how the rule was not there, or why the command just failed | 17:50 |
ganso | bswartz: it does not say why it failed | 17:50 |
ganso | bswartz: ProcessExecutionError: Unexpected error while running command. | 17:51 |
bswartz | ikr | 17:51 |
bswartz | that's pretty crappy error reporting | 17:51 |
*** crushil has quit IRC | 17:51 | |
ganso | bswartz: 6 times the same error | 17:51 |
ganso | bswartz: I guess it can't be a coincidence | 17:51 |
bswartz | yet another reason to rewrite the access helper for lvm | 17:51 |
*** crushil has joined #openstack-manila | 17:51 | |
ganso | bswartz: 3, in fact | 17:52 |
ganso | bswartz: messages are repeated | 17:52 |
bswartz | ganso: in the gate, how many backends do we run with? | 17:52 |
ganso | bswartz: 2 | 17:52 |
bswartz | and the migration always goes between the backends right? | 17:53 |
bswartz | we never migration back to the same pool? | 17:53 |
bswartz | migrate* | 17:53 |
ganso | bswartz: currently, no | 17:53 |
ganso | bswartz: Yogesh may have added a test that migrates to the same pool | 17:53 |
ganso | bswartz: I am not sure | 17:53 |
bswartz | https://paste.fedoraproject.org/559038/18135114/ | 17:56 |
bswartz | that's all the commands run when I did the single tempest test | 17:56 |
*** crushil has quit IRC | 17:56 | |
*** crushil has joined #openstack-manila | 17:56 | |
bswartz | there are 2 errors 3 seconds apart | 17:57 |
bswartz | geez these logs are hard to read | 17:58 |
ganso | bswartz: still no way to guess the reason for that | 18:02 |
ganso | bswartz: if you could add a log line before the execute | 18:03 |
ganso | bswartz: that prints the content of /etc/exports | 18:03 |
ganso | bswartz: we could figure out if the rule was there | 18:03 |
bswartz | yeah | 18:03 |
bswartz | still this isn't happening on the gate which points to something about my system | 18:04 |
ganso | bswartz: yes | 18:04 |
bswartz | gah | 18:06 |
bswartz | the LVM driver has no debug logic whatsoever | 18:06 |
*** lpetrut has joined #openstack-manila | 18:07 | |
bswartz | ganso: added the debug and ran again | 18:11 |
*** shausy has quit IRC | 18:11 | |
bswartz | that export did exist at the time the exportfs -u command was invoked | 18:11 |
*** shausy has joined #openstack-manila | 18:12 | |
bswartz | ganso: I worry there's a timing issue here, despite the lack on concurrency | 18:14 |
bswartz | lack of concurrency | 18:14 |
*** xyang_ has quit IRC | 18:14 | |
ganso | bswartz: lol we need concurrency for it to work, let's remove tooz and locks :P | 18:14 |
bswartz | I wonder if it's possible that commands are getting sent too fast | 18:14 |
bswartz | nooooo | 18:14 |
*** JoseMello has quit IRC | 18:15 | |
ganso | bswartz: if the rule was there, why did the command fail? | 18:15 |
ganso | bswartz: maybe you should add -v to the exportfs command | 18:15 |
ganso | bswartz: exportfs -vu | 18:15 |
bswartz | what would that do? | 18:15 |
ganso | bswartz: verbose | 18:16 |
ganso | bswartz: maybe improve error message | 18:16 |
bswartz | trying... | 18:16 |
bswartz | okay the exportfs error might be a red herring | 18:18 |
bswartz | it looks like the failure occurs before then, and that particular error is the cleanup logic failing to cleanup | 18:18 |
ganso | bswartz: but it is in migration_start | 18:19 |
ganso | bswartz: it could be that the test adds the rule | 18:19 |
ganso | bswartz: but does not wait for the rule to be added | 18:19 |
bswartz | oh wait | 18:19 |
ganso | bswartz: so it invokes migration_start | 18:19 |
bswartz | https://paste.fedoraproject.org/559044/87182783/ | 18:19 |
bswartz | exportfs -vu does cause a message to get printed | 18:20 |
bswartz | however the message indicates success while the exit code indicates failure | 18:20 |
bswartz | ganso: are there any config options I need to be setting related to LVM and migration? | 18:21 |
ganso | bswartz: the data_node_access_ip | 18:21 |
bswartz | I'm not running devstack-gate and I'm not using the pre/post hooks | 18:22 |
bswartz | I'm just running bare devstack and tempest with my own config | 18:22 |
ganso | bswartz: but it is not even reaching the Data service part | 18:22 |
bswartz | okay that value is getting set correctly by default | 18:23 |
bswartz | oh no I just had a really dark thought | 18:25 |
bswartz | I'm not running the normal ubuntu nfs-utils, I'm running my own compiled versions | 18:26 |
bswartz | what if there's a bug in my exportfs binary? | 18:26 |
* bswartz smh | 18:26 | |
*** shausy has quit IRC | 18:27 | |
ganso | bswartz: that's a possibility | 18:27 |
bswartz | testing that theory will require rebuilding my base image | 18:28 |
*** shausy has joined #openstack-manila | 18:28 | |
bswartz | rebuilding now... | 18:29 |
*** shausy has quit IRC | 18:36 | |
bswartz | 20 minutes to rebuild... | 18:51 |
bswartz | restacking now | 18:51 |
openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: Fix migration_success before completing https://review.openstack.org/434462 | 18:51 |
*** yumiriam has quit IRC | 19:03 | |
*** pcaruana has quit IRC | 19:03 | |
openstackgerrit | Rodrigo Barbieri proposed openstack/manila master: Fix migration_success before completing https://review.openstack.org/434462 | 19:06 |
bswartz | ganso: unbelievable! I reverted my nfs-utils to the ubuntu version and the test failures went away | 19:13 |
ganso | bswartz: that's very good! :D | 19:15 |
ganso | bswartz: and bad at the same time | 19:15 |
bswartz | it's good for this bug, bad for future work on ipv6 | 19:15 |
ganso | bswartz: as that nfs-utils was going to be required for ipv6, right? | 19:15 |
bswartz | ganso: yes | 19:15 |
bswartz | ganso: the ubuntu guy are unlikely to give us a decent nfs-utils until zesty+1 | 19:17 |
ganso | bswartz: yep | 19:18 |
ganso | bswartz: as they haven't so far | 19:18 |
bswartz | ganso: well debian stretch has a newer version, and stretch just became "stable" last week | 19:19 |
bswartz | so ubuntu should start pulling packages from there for zesty+1 | 19:19 |
ganso | bswartz: but does this newer include what we need for ipv6? | 19:20 |
bswartz | ganso: yes by a lot | 19:20 |
bswartz | however it's unknown what regressions the new package might cause | 19:20 |
bswartz | here we just found 1 hidden regression caused by me compiling my own version | 19:21 |
*** chlong has joined #openstack-manila | 19:36 | |
*** xyang_ has joined #openstack-manila | 19:40 | |
*** chlong has quit IRC | 19:51 | |
*** xyang_ has quit IRC | 19:54 | |
*** wiebalck has joined #openstack-manila | 20:02 | |
*** xyang_ has joined #openstack-manila | 20:04 | |
*** xyang_ has quit IRC | 20:11 | |
*** akapil has joined #openstack-manila | 20:14 | |
*** xyang_ has joined #openstack-manila | 20:17 | |
*** akapil has quit IRC | 20:19 | |
ganso | bswartz, gouthamr: maybe we can include this one in RC2 as well? https://review.openstack.org/#/c/434462/ | 20:25 |
*** fishbone has joined #openstack-manila | 20:29 | |
*** wiebalck has quit IRC | 20:31 | |
*** timcl has joined #openstack-manila | 20:37 | |
*** xyang_ has quit IRC | 20:39 | |
*** xyang_ has joined #openstack-manila | 20:39 | |
bswartz | ganso: that's a big change that hasn't even been reviewed and merged in master | 20:40 |
bswartz | let me read the bug report | 20:40 |
bswartz | okay can you explain where exactly this can affect users negatively? | 20:41 |
ganso | bswartz: the change is not very big | 20:42 |
ganso | bswartz: and users are unlikely to notice if no errors occur | 20:42 |
bswartz | ganso: am I looking at the wrong change? | 20:42 |
bswartz | I see 44 lines in manager.py | 20:43 |
ganso | bswartz: but if an error occurs after the task_state is set to migration_success, and before it is actually completed, it will be an inconsistent state | 20:43 |
ganso | bswartz: yes 44 lines, not very big | 20:43 |
bswartz | 6 separate edits | 20:43 |
bswartz | my point is that it's not trivially safe -- this change affects logic | 20:44 |
ganso | bswartz: also, there is a slightly big window of chance the user can invoke migration_complete twice causing problems | 20:44 |
ganso | bswartz: well, it is ok to backport after Ocata release | 20:44 |
bswartz | upon closer inspection it looks like you copy/pasted some code around | 20:45 |
ganso | bswartz: yes, I moved code, basically | 20:45 |
bswartz | Let's focus on merging in master | 20:45 |
bswartz | and revisit the decision to backport when we're getting ready for RC2 | 20:46 |
bswartz | I'd like risk assessments from others too | 20:46 |
ganso | bswartz: RC2 is tomorrow, right? | 20:46 |
bswartz | it's whenever we get the critical bug fixed | 20:47 |
ganso | bswartz: ok | 20:47 |
bswartz | due to the vagaries of the time-based release process it's possible for RC2 to not happen at all | 20:47 |
bswartz | but I'm hopeful we'll get it done tomorrow | 20:47 |
-openstackstatus- NOTICE: We're currently battling an increase in log volume which isn't leaving sufficient space for new jobs to upload logs and results in POST_FAILURE in those cases; recheck if necessary but keep spurious rebasing and rechecking to a minimum until we're in the clear. | 20:56 | |
*** xyang_ has quit IRC | 21:12 | |
*** xyang_ has joined #openstack-manila | 21:18 | |
*** zhonghua2 has quit IRC | 21:19 | |
*** zhonghua has joined #openstack-manila | 21:20 | |
*** porrua has quit IRC | 21:20 | |
*** dustins_ has quit IRC | 21:29 | |
openstackgerrit | Ben Swartzlander proposed openstack/manila-test-image master: Add fake sudo script https://review.openstack.org/434527 | 21:35 |
*** xyang_ has quit IRC | 21:36 | |
*** xyang_ has joined #openstack-manila | 21:37 | |
*** xyang_ has quit IRC | 21:40 | |
*** lpetrut has quit IRC | 21:41 | |
*** xyang_ has joined #openstack-manila | 21:42 | |
*** cknight has quit IRC | 21:46 | |
gouthamr | ganso: still around? | 21:49 |
ganso | gouthamr: yes | 21:50 |
gouthamr | ganso: hey! about https://review.openstack.org/#/c/434462/ | 21:50 |
*** eharney_ has joined #openstack-manila | 21:50 | |
*** eharney has quit IRC | 21:50 | |
gouthamr | ganso: i still have the same concern i may have spoken to you about before the code merged in ocata/newton (don't remember) | 21:50 |
*** eharney_ is now known as eharney | 21:50 | |
gouthamr | ganso: why can't we update the instance and the share in the same update.. | 21:51 |
gouthamr | ganso: i.e, within the same sqlalchemy session | 21:51 |
gouthamr | ganso: i know share_instance_update currently doesn't do that.. but it feels like that's the fix we need | 21:51 |
*** crushil has quit IRC | 21:51 | |
ganso | gouthamr: the first share instance update that sets the destination instance to available should be before attempting to delete the source instance, else the user will see the status "inactive" when listing the shares. Then later, when everything is complete, a share update is needed to set the task_state to migration_success | 21:52 |
ganso | gouthamr: so the timing of the updates is different, how can it be combined here? | 21:52 |
gouthamr | ganso: do it in the second update.. | 21:57 |
gouthamr | ganso: currently, you have three round trips to the database.. | 21:57 |
gouthamr | ganso: we can delete the source instance, get the list of updates we need based on the share type and update the destination instance and the share in one transaction | 22:01 |
gouthamr | ganso: for a brief unit of time, the user might see "migrating_to" -> however, i think that's alright | 22:02 |
ganso | gouthamr: ok I'll look into optimizing it tomorrow | 22:06 |
ganso | gouthamr: if we are okay with the user seeing "migrating_to", that could work | 22:06 |
gouthamr | ganso: sure.. this might be something to punt to "share migration pike improvements" if we want to not touch the database APIs... | 22:07 |
ganso | gouthamr: it could be backported | 22:08 |
ganso | gouthamr: I guess | 22:08 |
ganso | gouthamr: looks simple enough | 22:08 |
ganso | gouthamr: I mean, for this patch, reduce a DB call | 22:08 |
ganso | gouthamr: for Pike, add a method that updates share_instance and share in one call | 22:08 |
gouthamr | ganso: yes... looks simple and we update share access rules the same way | 22:08 |
ganso | gouthamr: that would reduce one more | 22:08 |
gouthamr | ganso: only the ceph driver exercises that path.. but you can update an instance access mapping and share access mapping in one transaction | 22:09 |
ganso | gouthamr: if the user starts seeing migrating => migrating_to => available instead of migrating => available like it is today, then it is an APIImpact, right? | 22:10 |
gouthamr | ganso: it likely is.. | 22:10 |
*** gouthamr has quit IRC | 22:22 | |
*** eharney has quit IRC | 22:49 | |
*** gouthamr has joined #openstack-manila | 22:53 | |
*** catintheroof has quit IRC | 23:00 | |
*** fishbone has quit IRC | 23:03 | |
*** tommylikehu_ has joined #openstack-manila | 23:13 | |
*** fishbone has joined #openstack-manila | 23:15 | |
*** tommylikehu_ has quit IRC | 23:20 | |
openstackgerrit | Merged openstack/manila master: Update HNAS driver version history https://review.openstack.org/433613 | 23:28 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!