Commit graph

2277 commits

Author SHA1 Message Date
Samuel Dionne-Riel 7889aef792
hydra-eval-jobs: Mirror eval errors in STDERR
Otherwise, errors will not be shown to end-users, which makes debugging
long evals pretty much impossible.

(cherry picked from commit 76299b9174)
2020-03-31 12:48:33 +02:00
Eelco Dolstra 15ae932488
Don't spam the journal with hydra-queue-runner status dumps 2020-03-26 15:30:37 +01:00
Eelco Dolstra 45ffe578b6
Keep track of the number of unsupported steps 2020-03-26 15:27:13 +01:00
Eelco Dolstra f5cdbfe21d
Abort unsupported build steps
If we don't see machine that supports a build step for
'max_unsupported_time' seconds, the step is aborted. The default is 0,
which is appropriate for Hydra installations that don't provision
missing machines dynamically.
2020-03-26 15:00:04 +01:00
Eelco Dolstra 5b731004da
Remove tabs 2020-03-26 14:59:52 +01:00
Eelco Dolstra d0688b93e1
Merge remote-tracking branch 'origin/master' into flake 2020-03-26 11:44:11 +01:00
Graham Christensen 12cc46cdb3
fixup: hydra-init: correct reference to hydra-backill-ids 2020-03-24 11:22:14 -04:00
Maximilian Bosch 73694087a0
Use the stopsign as icon for timeouts and exceeded log-limits
When I browse failed builds in a jobset-eval on Hydra, I regularly
mistake actual build-failures with temporary issues like timeouts (that
probably disappear at the next eval).

To prevent this kind of issue, I figured that using the stopsign-svg for
builds with timeouts or exceeded log-limits is a reasonable choice for
the following reasons:

* A user can now distinguish between actual build-errors (like
  compilation-failures or oversized outputs) and (usually) temporary issues
  (like a bloated log or a timeout).

* The stopsign is also used for aborted jobs that are shown in a
  different tab and can't be confused with timeouts for that reason.
2020-03-21 21:08:33 +01:00
Eelco Dolstra 4b5bb4e760
Merge remote-tracking branch 'origin/master' into flake 2020-03-04 15:28:23 +01:00
Eelco Dolstra 123bee1db5
Restore job type checking 2020-03-04 15:16:26 +01:00
Eelco Dolstra 69a6f3448a
Fix calling job functions
Fixes #718.
2020-03-04 15:16:04 +01:00
Eelco Dolstra 3cc1deb125
Merge pull request #721 from grahamc/one-by-one
hydra-evaluator: add a 'ONE_AT_A_TIME' evaluator style
2020-03-04 08:44:43 +01:00
Graham Christensen 994430b94b
treewide: allow nix command 2020-03-03 22:52:20 -05:00
Graham Christensen 117b9ecef1
Nix.pm: readNixFile: pass «--experimental-features nix-command»
Declarative jobsets were broken by the Nix update, causing
nix cat-file to break silently.

This commit restores declarative jobsets, based on top of a commit
making it easier to see what broke.
2020-03-03 22:36:21 -05:00
Graham Christensen 113a312f67
handleDeclarativeJobsetBuild: handle errors from readNixFile 2020-03-03 22:32:13 -05:00
Graham Christensen 5fae9d96a2
hydra-evaluator: add a 'ONE_AT_A_TIME' evaluator style
In the past, jobsets which are automatically evaluated are evaluated
regularly, on a schedule. This schedule means a new evaluation is
created every checkInterval seconds (assuming something changed.)

This model works well for architectures where our build farm can
easily keep up with demand.

This commit adds a new type of evaluation, called ONE_AT_A_TIME, which
only schedules a new evaluation if the previous evaluation of the
jobset has no unfinished builds.

This model of evaluation lets us have 'low-tier' architectures.

For example, we could now have a jobset for ARMv7l builds, where
the buildfarm only has a single, underpowered ARMv7l builder.
Configuring that jobset as ONE_AT_A_TIME will create an evaluation
and then won't schedule another evaluation until every job of
the existing evaluation is complete.

This way, the cache will have a complete collection of pre-built
software for some commits, but the underpowered architecture will
never become backlogged in ancient revisions.
2020-03-03 19:28:44 -05:00
Graham Christensen eaa65f51f4
hydra-evaluator: make the logic of the scheduler easier to read 2020-03-03 18:17:21 -05:00
Eelco Dolstra eb5873ae53
Fix build 2020-02-20 11:19:45 +01:00
Eelco Dolstra c642f787ee
hydra-eval-jobs: Parallelize
(cherry picked from commit be8eb9d00d)
2020-02-20 10:28:58 +01:00
Eelco Dolstra adf61e5cf8
Fix build
(cherry picked from commit 639c660abf)
2020-02-20 10:26:45 +01:00
Eelco Dolstra e4f5156c41
Build against nix-master
(cherry picked from commit e7f2139e25)
2020-02-20 10:24:04 +01:00
Eelco Dolstra b790a00729
Disable deprecation warnings
(cherry picked from commit 950e8bef6c)
2020-02-20 10:22:21 +01:00
Eelco Dolstra 2a50daa377
Update aggregate handling
(cherry picked from commit cf961ac893)
2020-02-20 10:13:39 +01:00
Eelco Dolstra 15187b059b
Remove hydra-eval-guile-jobs
This hasn't been used in a long time (Guix uses its own CI system),
and it probably doesn't work anymore.

(cherry picked from commit 23c9ca3e94)
2020-02-20 09:58:12 +01:00
Eelco Dolstra be8eb9d00d
hydra-eval-jobs: Parallelize 2020-02-19 21:10:22 +01:00
Eelco Dolstra 6f1d68bda4
Revert "hydra-eval-jobs -> nix eval-hydra-jobs"
This reverts commit 345512a6d0.
2020-02-19 20:36:52 +01:00
Eelco Dolstra cf961ac893
Update aggregate handling 2020-02-17 16:33:25 +01:00
Eelco Dolstra 68bdf834ee
Fix hydra-evaluator --unlock 2020-02-15 16:40:16 +01:00
Eelco Dolstra 345512a6d0
hydra-eval-jobs -> nix eval-hydra-jobs 2020-02-15 15:59:34 +01:00
Eelco Dolstra 23c9ca3e94
Remove hydra-eval-guile-jobs
This hasn't been used in a long time (Guix uses its own CI system),
and it probably doesn't work anymore.
2020-02-15 15:59:34 +01:00
Eelco Dolstra 639c660abf
Fix build 2020-02-15 14:54:35 +01:00
Graham Christensen ae1b9face3
Merge remote-tracking branch 'origin/master' into flake 2020-02-11 12:52:46 -05:00
Graham Christensen 027668f0db
hydra.sql: add an index for slow queries in production
These queries used to use (jobset, project) based indexes,
and the addition of jobset_id makes most of those indexes
unusable now.
2020-02-11 12:52:28 -05:00
Graham Christensen 53896ff19b
Merge remote-tracking branch 'origin/master' into flake 2020-02-11 08:36:40 -05:00
Graham Christensen 834793468f
fixup: d'oh, make the migrations from #710 part-2 sequential 2020-02-11 08:36:14 -05:00
Eelco Dolstra 881b7449fd Merge remote-tracking branch 'origin/master' into flake 2020-02-11 14:23:16 +01:00
Graham Christensen 2637a7ad76
Builds: index literally what latest-finished queries 2020-02-11 07:06:21 -05:00
Graham Christensen f0f41eaaff
LatestSucceededForJob{,set}: Filter with jobset_id 2020-02-11 07:06:20 -05:00
Graham Christensen 66fbbd9692
Jobs.builds: Fetch via Jobs.jobset_id 2020-02-11 07:06:20 -05:00
Graham Christensen 7c71f9df28
Jobsets.builds: Fetch via Jobsets.id 2020-02-11 07:06:20 -05:00
Graham Christensen 3c392b8cd8
Jobsets.jobs: Fetch via Jobsets.id 2020-02-11 07:06:20 -05:00
Graham Christensen 8ef08f1385
Builds.jobset_id: make not-null 2020-02-11 07:06:20 -05:00
Graham Christensen 2cdcc7f188
Jobs.jobset_id: make not-null 2020-02-11 07:06:17 -05:00
Eelco Dolstra 100e09a5b3
Merge remote-tracking branch 'origin/master' into flake
Also update flake.lock
2020-02-10 17:58:10 +01:00
Graham Christensen c4cc72f944
hydra-init: Warn about the schema version migration 2020-02-10 11:43:03 -05:00
Graham Christensen f69260118b
hydra-backfill-ids: create to add jobset_id values to Builds and Jobs
Vacuum every 10 iterations, update 10k at a time.
2020-02-10 11:43:03 -05:00
Graham Christensen f3a561aecd
Builds: populate Builds.jobset_id in hydra-eval-jobset 2020-02-10 11:43:02 -05:00
Graham Christensen ddf00fa627
Builds: add a nullable jobset_id foreign key to Jobsets.
Also, adds an explicitly named "builds" accessor to the Jobsets
Schema object, which uses the project/jobset name.
2020-02-10 11:43:02 -05:00
Graham Christensen 624f1d8d2d
Jobs: populate Jobs.jobset_id field when writing from hydra-eval-jobset 2020-02-10 11:43:02 -05:00
Graham Christensen efa1f1d4fb
Jobs: add a nullable jobset_id foreign key to Jobsets.
Also, adds an explicitly named "jobs" accessor to the Jobsets
Schema object, which uses the project/jobset name.
2020-02-10 11:43:02 -05:00