| opendevreview | Goutham Pacha Ravi proposed openstack/manila stable/wallaby: Fix ipaddress issues in the infinidat driver https://review.opendev.org/c/openstack/manila/+/799136 | 00:08 |
|---|---|---|
| opendevreview | Merged openstack/manila master: Fix ipaddress issues in the infinidat driver https://review.opendev.org/c/openstack/manila/+/799107 | 00:54 |
| opendevreview | Takashi Kajinami proposed openstack/manila master: Migrate all quota parameters to [quota] section. https://review.opendev.org/c/openstack/manila/+/799130 | 01:46 |
| opendevreview | Goutham Pacha Ravi proposed openstack/manila stable/wallaby: TrivialFix: Fix the filter name in config helper https://review.opendev.org/c/openstack/manila/+/799137 | 04:54 |
| opendevreview | Takashi Kajinami proposed openstack/manila master: Migrate all quota parameters to [quota] section. https://review.opendev.org/c/openstack/manila/+/799130 | 05:49 |
| *** dviroel|out is now known as dviroel | 11:34 | |
| opendevreview | Merged openstack/manila stable/wallaby: Fix ipaddress issues in the infinidat driver https://review.opendev.org/c/openstack/manila/+/799136 | 12:02 |
| opendevreview | Tom Barron proposed openstack/manila stable/wallaby: Adds a clear error message to operator-set limits https://review.opendev.org/c/openstack/manila/+/798882 | 12:14 |
| opendevreview | Tom Barron proposed openstack/manila stable/wallaby: Add Ceph version check https://review.opendev.org/c/openstack/manila/+/797955 | 12:14 |
| felipe_rodrigues | hi guys, I am getting an error while running the test `test_add_multiple_share_replicas` [1], I dont have enough pools for creating 2 replicas. In fact, I would expect that the test is skipped in that scenario. There is a skip condition in the test [2]. The problem is: the skip is looking if I have 3 pools using the REPLICATION DOMAIN (I have) , but the test is creating the replicas using the AVAILABILITY ZONE. | 12:21 |
| felipe_rodrigues | Should the skip be checking the number of pools in the replication domain within the target az ?! | 12:21 |
| felipe_rodrigues | [1] https://github.com/openstack/manila-tempest-plugin/blob/master/manila_tempest_tests/tests/api/test_replication.py#L238 | 12:21 |
| felipe_rodrigues | [2] https://github.com/openstack/manila-tempest-plugin/blob/master/manila_tempest_tests/tests/api/test_replication.py#L244 | 12:21 |
| dviroel | hey felipe_rodrigues, i remember that issue :) | 12:26 |
| dviroel | the test don't check if there is 2 pools in the destination az, indeed | 12:27 |
| dviroel | it only checks the total number of pools | 12:27 |
| dviroel | the way that was working in netapp ci, was to have 2 pools in each AZ | 12:27 |
| dviroel | or | 12:27 |
| dviroel | you can fiz the test to check if there is more than one pool in the destination az, and skip the test in case if that check fails | 12:28 |
| dviroel | would be better to fix the test ;) | 12:28 |
| dviroel | only netapp driver currently test that, i don't think that there is another driver running this test | 12:29 |
| dviroel | the netapp setup was: 2 pools in backend A, 1 pool in backend B and another pool in backend C. Backends B and C were in the same cluster. | 12:30 |
| dviroel | somthing like that | 12:30 |
| dviroel | to increase the possibilities of create share from snapshot in another pool/backend, iirc | 12:31 |
| felipe_rodrigues | thank you Douglas! NetApp tests are passing, the only problem was that one of the pools in the replica AZ could not get it up (some issues in the netapp CI). Then, this test failed.. I spent a time thinking that my patch is doing something.. debugging it.. while whether the skip is coded correctly, I did not need to spend that time... So, I'm just confirming that this is a wrong test implementation, so I can | 12:47 |
| felipe_rodrigues | provide a fix for avoiding that misunderstanding while reading the test logs. | 12:47 |
| dviroel | felipe_rodrigues++ thanks | 12:51 |
| dviroel | the code to filter pools by backend or azs is already at some other tests | 12:52 |
| felipe_rodrigues | good.. i'll have a look at it! thks | 13:00 |
| dviroel | ++ | 13:02 |
| opendevreview | Merged openstack/manila master: Handle service client authorization errors https://review.opendev.org/c/openstack/manila/+/783904 | 14:08 |
| *** dviroel is now known as dviroel|lunch | 14:51 | |
| batrick | gouthamr: vkmc: tbarron: did the volume_prefix passed to CephFSVolumeClient change in manila? | 16:12 |
| batrick | did you ever set it? | 16:12 |
| batrick | frankly i'm confused how this code wasn't always stat'ing an empty string which should result in a fault | 16:14 |
| gouthamr | batrick: o/ nope, it was always /volumes | 16:15 |
| batrick | it's a mystery then | 16:15 |
| batrick | anyway, i have a fix in the works | 16:15 |
| *** dviroel|lunch is now known as dviroel | 16:16 | |
| gouthamr | batrick: awesome - i've an environment where this doesn't break, that we can trace the empty stat on it | 16:16 |
| batrick | actually you can apply my patch in the container to see if it resolves the issue for you? | 16:17 |
| batrick | it just changes ceph_volume_client.py | 16:17 |
| batrick | https://github.com/ceph/ceph/pull/42159 | 16:18 |
| batrick | i'll have a nautilus backport up shortly | 16:18 |
| batrick | rraja: let's chat here | 16:20 |
| rraja | batrick, sure | 16:24 |
| gouthamr | batrick: ack; lemme give it a shot | 16:24 |
| batrick | yes i suspected that nautilus qa would have instances of the empty path stat | 16:25 |
| rraja | thanks for the fix, batrick | 16:25 |
| batrick | i don't know why this is a problem only for manila | 16:25 |
| batrick | brb | 16:25 |
| rraja | yeah, recent testing of v14.2.22 's ceph_volume_client on centos also passed | 16:33 |
| batrick | gouthamr: https://github.com/ceph/ceph/pull/42162 | 16:43 |
| rraja | batrick, https://github.com/ceph/ceph/pull/42159#issuecomment-873130064 what do you think? | 17:06 |
| gouthamr | batrick++ no more aborts on volume creation; doing some more tests | 17:07 |
| rraja | great!! | 17:07 |
| batrick | rraja: thanks, i forgot that | 17:07 |
| rraja | batrick, thanks for root causing and sending the fix | 17:08 |
| gouthamr | crud works like a charm | 17:09 |
| gouthamr | will get some functional tests running to verify | 17:09 |
| gouthamr | o/ vhari: need some help | 17:17 |
| gouthamr | vhari: i've manually created/deleted shares on sealusa19 after applying batrick's fix; i was wondering if you could take a look for sanity | 17:18 |
| * gouthamr ¯\_(ツ)_/¯ wrong network | 17:18 | |
| vhari | gouthamr, aye good news .. sure will do | 17:20 |
| dviroel | (⊙‿⊙) | 17:27 |
| opendevreview | Goutham Pacha Ravi proposed openstack/manila stable/victoria: Fix ipaddress issues in the infinidat driver https://review.opendev.org/c/openstack/manila/+/799249 | 17:38 |
| gouthamr | ^ last branch where we support py38 | 17:46 |
| opendevreview | Merged openstack/manila stable/wallaby: TrivialFix: Fix the filter name in config helper https://review.opendev.org/c/openstack/manila/+/799137 | 19:07 |
| opendevreview | Goutham Pacha Ravi proposed openstack/manila stable/victoria: TrivialFix: Fix the filter name in config helper https://review.opendev.org/c/openstack/manila/+/799250 | 19:08 |
| simondodsley | gouthamr: I've set up a seperate tempest environment and it appears that the test instances created have no routes set so there is no way for the nfs mounts to ever work | 19:37 |
| simondodsley | no routes to the correct subnet where the share is, that is | 19:38 |
| simondodsley | gouthamr: I suspect the tempest config file is incorrect somehow. Who can help me debug it as I'm not sure I'm up to that from a networking perspective | 19:41 |
| gouthamr | simondodsley: i see, can you share your tempest.conf; i'll look at the default config in the meantime | 19:48 |
| *** dviroel is now known as dviroel|brb | 19:50 | |
| simondodsley | gouthamr: this is it https://www.irccloud.com/pastebin/7x7icmn8/ | 19:53 |
| gouthamr | thanks | 19:54 |
| simondodsley | gouthamr: the instances can ping the tempest host, but nothing else beyond that | 19:58 |
| gouthamr | the networking section in that config looks correct | 19:59 |
| gouthamr | https://github.com/openstack/tempest/blob/5dcdd41ddafa2e455ab0dd0ce0b1d3f3e32caa04/tempest/lib/common/dynamic_creds.py#L365-L367 | 19:59 |
| gouthamr | the router created will be connected to whatever the "public_network_id" points to | 19:59 |
| * gouthamr looks at the last run of the pure storage ci for tempest log | 20:00 | |
| gouthamr | Body: {"router": {"name": "tempest-TestShareExtendNFS-2101733040-router", "tenant_id": "61e09201775747d38f43ae76073cca15", "external_gateway_info": {"network_id": "f1eef55e-f950-4cf6-ba80-17560460714b"}}} | 20:01 |
| gouthamr | in tempest.conf's [network] section, you have: public_network_id = f1eef55e-f950-4cf6-ba80-17560460714b ; so that's doing the right thing | 20:02 |
| simondodsley | so that will explain why the instance can ping the tempest host I guess, but why can't it get out any further | 20:03 |
| opendevreview | Merged openstack/manila stable/wallaby: Add Ceph version check https://review.opendev.org/c/openstack/manila/+/797955 | 20:09 |
| *** dviroel|brb is now known as dviroel | 20:34 | |
| * gouthamr is on a call - will be slow to respond for a bit | 20:34 | |
| *** dviroel is now known as dviroel|out | 21:03 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!