Board label change of DUTs on shard should trigger the DUTs reassignment
Reported by
jrbarnette@chromium.org,
Jun 25 2016
|
||||||
Issue descriptionI was just trying to manually reassign selected DUTs from 'board:samus' to 'board:samus-cheets'. The change seemed to take effect, but then, magically, one of the DUTs lost the change. Here's the sequence according to my shell history: $ M=chromeos4-row12-rack5-host5,chromeos4-row12-rack5-host3,chromeos4-row12-rack5-host19 $ atest label remove -m $M board:samus-cheets $ atest label add -m $M board:samus-cheets [ ... oops! I'd meant to change the label ... ] $ atest label remove -m $M board:samus-cheets $ atest label add -m $M board:samus After this, I checked the database with more 'atest' commands to make sure the label changes were showing up. A little while later, I discovered that one of the hosts (chromeos4-row12-rack5-host5) had lost the changes; the board label was back to 'samus-cheets'. I note that label 'board:samus' is handled by a shard, whereas label 'board:samus-cheets' is handled by the master.
,
Jun 28 2016
,
Jun 28 2016
,
Jun 28 2016
+ Charlene as owner
,
Jun 28 2016
I have moved all the board:samus-cheets DUTs from shard chromeos-server33.cbf to master. Below is the list of DUTs moved: chromeos4-row12-rack6-host1 chromeos4-row12-rack5-host11 chromeos4-row12-rack5-host9 chromeos4-row12-rack5-host13 chromeos4-row12-rack6-host3 chromeos4-row12-rack6-host9 chromeos2-row24-rack1-host13 chromeos2-row24-rack1-host11 chromeos2-row24-rack1-host7 chromeos2-row24-rack1-host19 chromeos2-row24-rack2-host15 chromeos2-row24-rack3-host17 The pool label has been removed in the move process for the first five DUTs. They were either in the pool:bvt or in the pool:suites, but I don't know the exact pool they were in before. Richard, do you know what pool label I should add for these five DUTs: chromeos4-row12-rack6-host1 chromeos4-row12-rack5-host11 chromeos4-row12-rack5-host9 chromeos4-row12-rack5-host13 chromeos4-row12-rack6-host3
,
Jun 29 2016
,
Jun 29 2016
Charlene: We'd had similar issues with elm-cheets being relabeled as elm. Can you have a quick look if it's the same root cause, or if we need a separate bug for that?
,
Jun 29 2016
Re comment #5: Put all new DUTs into pool:suites. Then we can balance_pools to whatever counts we need. rohitbm@: can you comment on the expected pool counts for samus-cheets?
,
Jun 29 2016
Neither elm-cheets nor elm is assigned to shard. Is it accidentally relabeled by developer? Can you specify the hostname of the DUT(s) you are talking about?
,
Jun 29 2016
All the five DUTs have been added into pool:suites: chromeos4-row12-rack6-host1 chromeos4-row12-rack5-host11 chromeos4-row12-rack5-host9 chromeos4-row12-rack5-host13 chromeos4-row12-rack6-host3
,
Jun 29 2016
#9: Thanks, that looks like a completely different issue then, filed Issue 624207.
,
Jun 29 2016
For shard chromeos-server44.cbf, DUT chromeos4-row4-rack7-host7 (board:veyron_speedy) has been moved back to master. And the shard database has been cleaned up. For shard chromeos-server51.cbf, DUT chromeos2-row6-rack2-host1, chromeos4-row11-rack11-host17, chromeos4-row12-rack11-host19, chromeos4-row12-rack11-host9 (board:cyan-cheets) have been moved back to master. And shard database has been cleaned up.
,
Jun 29 2016
pool:cts needs 6 devices for -cheets boards. Same goes for Samus.
,
Jun 20 2017
I think this is not applied any more |
||||||
►
Sign in to add a comment |
||||||
Comment 1 by jrbarnette@chromium.org
, Jun 25 2016Here's the task/job history of the three DUTs around the time of the event: chromeos4-row12-rack5-host5 2016-06-24 17:10:49 -- http://cautotest/tko/retrieve_logs.cgi?job=/results/hosts/chromeos4-row12-rack5-host5/56835162-verify/ 2016-06-24 16:44:19 NO http://cautotest/tko/retrieve_logs.cgi?job=/results/hosts/chromeos4-row12-rack5-host5/56834850-repair/ 2016-06-24 16:37:36 -- http://cautotest/tko/retrieve_logs.cgi?job=/results/hosts/chromeos4-row12-rack5-host5/56834710-verify/ chromeos4-row12-rack5-host3 2016-06-24 17:45:47 -- http://cautotest/tko/retrieve_logs.cgi?job=/results/hosts/chromeos4-row12-rack5-host3/56835469-verify/ 2016-06-24 16:44:19 NO http://cautotest/tko/retrieve_logs.cgi?job=/results/hosts/chromeos4-row12-rack5-host3/56834853-repair/ 2016-06-24 16:37:37 -- http://cautotest/tko/retrieve_logs.cgi?job=/results/hosts/chromeos4-row12-rack5-host3/56834716-verify/ chromeos4-row12-rack5-host19 The actual changes happened around 17:06, so it was shortly before the recorded verify task on the problem DUT. I'll bet good odds at small stakes that that's related.