Commit Graph

48 Commits

Author SHA1 Message Date
Alex Dadgar 4bdccab550 goimports 2019-01-22 15:44:31 -08:00
Alex Dadgar 6d8bb3a7bd Duplicate blocked evals cancelling improved
The old logic for cancelling duplicate blocked evaluations by job id had
the issue where the newer evaluation could have additional node classes
that it is (in)eligible for that we would not capture. This could make
it such that cluster state could change such that the job would make
progress but no evaluation was unblocked.
2018-11-07 10:08:23 -08:00
Alex Dadgar bd420692f3 fix logging 2018-09-25 10:49:55 -07:00
Preetha Appan 86e725e84c Added logging around nacked evals in the scheduler worker 2018-09-25 10:49:02 -07:00
Alex Dadgar 99498da6ed Denormalize jobs in plan and ignore resources of terminal allocs
Denormalize jobs in AppendAllocs:
AppendAlloc was originally only ever called for inplace upgrades and new
allocations. Both these code paths would remove the job from the
allocation. Now we use this to also add fields such as FollowupEvalID
which did not normalize the job. This is only a performance enhancement.

Ignore terminal allocs:
Failed allocations are annotated with the followup Eval ID when one is
created to replace the failed allocation. However, in the plan applier,
when we check if allocations fit, these terminal allocations were not
filtered. This could result in the plan being rejected if the node would
be overcommited if the terminal allocations resources were considered.
2018-09-24 13:53:43 -07:00
Alex Dadgar 3c19d01d7a server 2018-09-15 16:23:13 -07:00
Alex Dadgar 405dab2253 integration test and basic fixes 2018-03-21 16:51:44 -07:00
Josh Soref b56ebdf350 spelling: invoke 2018-03-11 18:00:32 +00:00
Josh Soref 3232f6f27e spelling: abandoned 2018-03-11 17:34:16 +00:00
Alex Dadgar 6911bd7676 Worker waits til max ModifyIndex across EvalsByJob
This PR fixes a scheduling race condition in which the plan results from
one invocation of the scheduler were not being considered by the next
since the Worker was not waiting for the correct index.

Fixes https://github.com/hashicorp/nomad/issues/3198
2017-09-14 14:28:43 -07:00
Alex Dadgar 84d06f6abe Sync namespace changes 2017-09-07 17:04:21 -07:00
Luke Farnell f0ced87b95 fixed all spelling mistakes for goreport 2017-08-07 17:13:05 -04:00
Alex Dadgar b69b357c7f Nomad builds 2017-02-07 20:31:23 -08:00
Alex Dadgar bbe6e3d0b6 Larger delay on mismatch 2016-10-27 11:41:43 -07:00
Alex Dadgar a1d08c2aba Add scheduler version enforcement 2016-10-26 14:52:48 -07:00
Ben Barnard 83f647ed84 Replace "the the" with "the" in documentation and comments 2016-10-11 15:31:40 -04:00
Diptanu Choudhury dabb83063b Review comments 2016-07-25 17:26:38 -07:00
Diptanu Choudhury 7bafb7c675 Updating the job summary while mutating jobs and allocation objects 2016-07-25 17:26:38 -07:00
Alex Dadgar 3a8a27bcff refresh index eval id in log 2016-06-22 13:48:41 -07:00
Alex Dadgar 8ceb7ead20 Do not use snapshot 2016-06-22 09:33:15 -07:00
Alex Dadgar 25decca3ca Worker waitForIndex uses StateStore index, not Raft Applied Index 2016-06-22 09:04:22 -07:00
Alex Dadgar bfdd5846e1 Track unblock indexes and check evals on block to see if they missed an update while in the scheduler 2016-05-24 20:10:56 -07:00
Alex Dadgar 15936822a4 Worker annotates evals with their snapshot index 2016-05-24 20:10:56 -07:00
Alex Dadgar 18d9e89065 Reuse the same evaluation and reblock it until there is no more work to do 2016-05-24 20:10:56 -07:00
Alex Dadgar 88ddfbed31 Revert "Debug messages around the plan and plan response"
This reverts commit 7646657e6b8a892210779eaf5708341b94b29b24.
2016-02-22 22:24:52 -08:00
Alex Dadgar fa8e2d31ee Revert "err logs in worker and scheduler"
This reverts commit 7befc586521b70eb84013bff367310e4cfa45c27.
2016-02-22 22:23:57 -08:00
Alex Dadgar c2242552a1 Debug messages around the plan and plan response 2016-02-22 20:36:11 -08:00
Alex Dadgar f48eabe753 err logs in worker and scheduler 2016-02-22 14:47:59 -08:00
Alex Dadgar 18d2d9c091 Killing a driver handle is retried with an exponential backoff 2016-02-16 21:00:49 -08:00
Armon Dadgar df16cea2a4 nomad: worker supports create eval 2015-09-07 14:23:48 -07:00
Armon Dadgar c7773feced nomad: improve error messages at start for dev mode 2015-09-06 20:18:47 -07:00
Armon Dadgar 7e644b7cc9 nomad: use fast and slow exponential backoff in worker 2015-08-23 17:39:49 -07:00
Armon Dadgar 8c2bc337e6 nomad: adding ability to pause a worker 2015-08-23 10:52:31 -07:00
Armon Dadgar cae67b7f60 nomad: expose UpdateEval as a planner 2015-08-15 14:25:00 -07:00
Armon Dadgar 8dfcb99e7f nomad: rename SystemScheduler to CoreScheduler 2015-08-15 12:38:58 -07:00
Armon Dadgar f7007bfeb5 nomad: avoid split-brain in plan processing due to leader transition or eval retry 2015-08-12 15:44:36 -07:00
Armon Dadgar 183a238481 nomad: avoid split-brain eval handling after leader transition 2015-08-12 15:25:31 -07:00
Armon Dadgar c09da9fa51 nomad: measure invoke time per scheduler 2015-08-06 17:45:27 -07:00
Armon Dadgar 09a8c15d7e scheduler: adding service scheduler definition 2015-08-06 17:25:14 -07:00
Armon Dadgar 8c2f8cddd0 nomad: create system scheduler as needed 2015-08-06 17:08:40 -07:00
Armon Dadgar 5f1ebb9274 nomad: adding special 'system' scheduler 2015-08-06 17:04:35 -07:00
Armon Dadgar db78ce9a02 nomad: make worker more resilient to transient errors 2015-08-05 16:23:37 -07:00
Armon Dadgar 8e66fdfdab nomad: test waitForIndex 2015-07-28 17:11:00 -07:00
Armon Dadgar c65b2c472b nomad: testing worker dequeue 2015-07-28 17:02:54 -07:00
Armon Dadgar 9e6c0c9c3a nomad: guard against plan during shutdown 2015-07-28 16:52:13 -07:00
Armon Dadgar 2b4501bee8 nomad: improving worker, submit plans 2015-07-28 16:36:15 -07:00
Armon Dadgar d7e9927661 nomad: integrating worker and scheduler 2015-07-28 16:15:32 -07:00
Armon Dadgar 3c2a16038b nomad: first pass adding scheduling workers 2015-07-28 15:12:08 -07:00