lix-project/hydra
3
0
Fork 6
Code Issues 2 Pull requests Packages Projects Releases Wiki Activity
3848 commits 2 branches 0 tags 20 MiB
Commit graph

3848 commits

This branch
This branch
All branches
Author SHA1 Message Date
Graham Christensen a1579007f4
Update t/test.pl: fixup the list of env vars yath looks at for job count 2022-02-10 11:53:10 -05:00
Graham Christensen b1b1b4b74b
t/lib/LDAPContext.pm: error message nit 
Co-authored-by: Cole Helbling <cole.e.helbling@outlook.com>
 2022-02-10 11:31:49 -05:00
Graham Christensen 33f4c4c13d build-locally-with-substitutable-path.t: give nix-store --delete a bit more time to run 
Under high load, like 64-128 tests at once, this can take more than a second.
 2022-02-10 11:13:31 -05:00
Graham Christensen e709a17508 gitea.t: try opening the file a few times 2022-02-10 11:01:09 -05:00
Graham Christensen 4acaf9c8b0 hydra-queue-runner: don't dispatch until the machines parser has completed one run 
Periodically, I have seen tests fail because of out of order queue runner behavior:

    checking the queue for builds > 0...
    loading build 1 (tests:basic:empty_dir)
    aborting unsupported build step '...-empty-dir.drv' (type 'x86_64-linux')
    marking build 1 as failed
    adding new machine ‘localhost’

This patch should prevent the dispatcher from running before any machines are
made available.
 2022-02-10 10:54:30 -05:00
Graham Christensen 09652475bd Set the YATH_JOB_COUNt to NIX_BUILD_CORES if no parallelism variables are set 2022-02-10 10:00:32 -05:00
Graham Christensen 5506962537 Test::Postgresql: update to a tagged fork to improve paralellism 2022-02-10 09:37:48 -05:00
Graham Christensen 4837a68d8d LDAPContext: don't background slapd 2022-02-10 09:28:53 -05:00
Graham Christensen 8bce8109e3 flake.nix: drop the LDAP server test 2022-02-09 20:56:10 -05:00
Graham Christensen 80c6525029 LDAP: Create a test which does not use a VM 2022-02-09 20:56:10 -05:00
Graham Christensen e4adec36f3 Set an env var in the build environment to point to the root of openldap 2022-02-09 20:56:10 -05:00
Graham Christensen 60adee7f94 hydra check inputs: add openldap 2022-02-09 20:56:10 -05:00
Graham Christensen 9ae7c8bddc Hydra::Helper::Exec add an expectOkay which dies with stdout / stderr on exit 2022-02-09 20:56:10 -05:00
Graham Christensen 5184e895df
Merge pull request #1148 from DeterminateSystems/hydra-helper-exec 
captureStdoutStderr*: move to Hydra::Helper::Exec which helps avoid some environment variable fixation problems
 2022-02-09 14:42:07 -05:00
Graham Christensen a374ef7d81
Merge pull request #1149 from DeterminateSystems/flake-format 
flake.nix: format with nixpkgs-fmt
 2022-02-09 14:40:48 -05:00
Graham Christensen 845e6d4760 captureStdoutStderr*: move to Hydra::Helper::Exec which helps avoid some environment variable fixation problems 2022-02-09 14:28:50 -05:00
Graham Christensen 68c6c3d556 flake.nix: format with nixpkgs-fmt 2022-02-09 10:43:40 -05:00
Graham Christensen 1abe7f4d80
Merge pull request #1145 from DeterminateSystems/eval-status 
`eval_*` events: change the API to include IDs and use \t separators
 2022-02-09 10:23:15 -05:00
Graham Christensen 517dce285a eval_added event: change interface to traceID\tjobsetID\tevaluationID 
I was not going to break the interface until I noticed
the current implementation uses the string literal \t.
 2022-02-08 09:51:35 -05:00
Graham Christensen d512e6220f eval_failed event: change interface to traceID\tjobsetID 
I was not going to break the interface until I noticed the other eval_* events used literal \ts
 2022-02-08 09:51:35 -05:00
Graham Christensen 2597fa8c11 eval_cached event: change interface to traceID\tjobsetID\tevaluationID 
I was not going to break the interface until I noticed
the current implementation uses the string literal \t.
 2022-02-08 09:51:35 -05:00
Graham Christensen be531c6c57 notifications.t: create a helper to parse the actual messages from the evaluator using the Event code 2022-02-08 09:51:35 -05:00
Graham Christensen c30f084f32 eval_started event: change interface to traceID\tjobsetID 
I was not going to break the interface until I noticed
the current implementation uses the string literal \t.
 2022-02-08 09:51:35 -05:00
Graham Christensen 3864ca820a
Merge pull request #1144 from DeterminateSystems/eval-event-cleanup 
Eval event cleanup
 2022-02-08 09:50:33 -05:00
Graham Christensen c0eb873379 notifications: document eval_* 2022-02-07 18:08:01 -05:00
Graham Christensen f648e91487 StepFinished.t: fixup use 2022-02-07 16:08:40 -05:00
Graham Christensen 7107ce2bc7 t/Event/* -> t/Hydra/Event/ 2022-02-07 16:08:40 -05:00
Graham Christensen 8a18326f2b Sort notification classes / events 2022-02-07 16:08:27 -05:00
Graham Christensen 7b7c03a521
Merge pull request #1141 from amerocu/amerocu/fix-dev-env 
missing develop dependencies
 2022-02-06 14:35:27 -05:00
Maximilian Bosch 5ae26aa760
Update Nix to 2.6 2022-02-06 15:05:15 +01:00
Marco Turchetto 73f2cc184f missing develop dependencies 2022-02-05 20:28:44 +01:00
Graham Christensen 85b842e0ac
Merge pull request #1137 from DeterminateSystems/runcommand-logs 
Store and display the output of RunCommands
 2022-01-31 16:26:31 -05:00
Cole Helbling b57345ba1f hydra.sql: add IndexRunCommandLogsOnBuildID index 2022-01-31 12:56:34 -08:00
Cole Helbling d0b6329aa8 sql/upgrade-81: remove unnecessary comment 2022-01-31 12:55:36 -08:00
Cole Helbling 9c4e6f78e7 hydra-module: don't bzip2 runcommand-logs 2022-01-31 12:55:36 -08:00
Cole Helbling 8c67e32480 RunCommand: ensure we reset the umask 2022-01-31 12:55:36 -08:00
Cole Helbling 34e4c119f4 build.tt: don't duplicate RunCommandLog buttons 2022-01-31 11:40:16 -08:00
Cole Helbling 61189ecca9 Helper/Nix: constructRunCommandLogPath: verify uuid is valid 
This shouldn't be possible normally, but it is possible to:

    $db->resultset('RunCommandLogs')->new({ uuid => "../etc/passwd" });

if you have access to the `$db`.
 2022-01-31 08:58:33 -08:00
Cole Helbling e381751564 Helper/Nix: constructRunCommandLogPath: return undef in case of an error 
This allows us to give a web request to an invalid UUID a 404.
 2022-01-31 08:58:33 -08:00
Cole Helbling 2c6487b8d7 t/Helper: test constructRunCommandLogPath 2022-01-31 08:58:33 -08:00
Cole Helbling 8bf3cdbc67 t/Helper: switch to using test_context() 2022-01-31 08:58:33 -08:00
Cole Helbling 8eab7b8543 Helper/Nix: constructRunCommandLogPath: take RunCommandLog as input 
This way we ensure that it actually exists in the database, rather than
blindly trusting user-generated input.
 2022-01-31 08:58:33 -08:00
Cole Helbling 61914d56c6 runcommand-log.tt: escape the command 2022-01-31 08:58:33 -08:00
Cole Helbling 71bbb042db build.tt: link to the pretty, raw, and tail versions of the log 
Also split it out to a new div -- there are now 3 lines per
RunCommandLog -- the first saying when it started, the second saying how
long it ran for (or has been running), and the third with the buttons
for the pretty, raw, and tail versions of the log.
 2022-01-31 08:58:33 -08:00
Cole Helbling 3594ba942a Controller/Build: use showLog in view_runcommandlog 
This also adds the `runcommandlog` object to the stash so that we can
access its uuid as well as command run in order to display more useful
and specific information on the webpage.
 2022-01-31 08:58:33 -08:00
Cole Helbling 1d0076408b Controller/Build: pass log_uri to showLog in place of drvPath 
This way, we can reuse the `showLog` sub for other things, such as
`view_runcommandlog` (which doesn't have a drvPath attached).
 2022-01-31 08:58:33 -08:00
Cole Helbling ff390e89a6 Controller/Build: remove unused parameter from showLog 2022-01-31 08:58:33 -08:00
Cole Helbling 38896db6b6 t/RunCommand: init http.t 
Test that we can indeed visit the pages of a valid runcommand log and
not of an invalid one.
 2022-01-31 08:58:33 -08:00
Cole Helbling 47c1f89d5a t/RunCommand: fixup basic.t to use uuid 2022-01-31 08:58:33 -08:00
Cole Helbling fc3cf4ecb2 RunCommandLogs: identify and access via uuid 
Using a sha1 of the command combined with the build ID is not a
particularly good or unique identifier:

* A build could fail, be restarted, and then succeed -- assuming no
configuration changes, the sha1 hash of the command as well as the build
ID will be the same. This would lead to an overwritten log file.

* Allowing user input to influence filenames is not the best of ideas.
 2022-01-31 08:58:33 -08:00