Commit graph

1872 commits

Author SHA1 Message Date
Eelco Dolstra b00bdefa98 Fix hydra-server signing 2016-02-26 17:28:16 +01:00
Eelco Dolstra 9de336de7c Proxy local binary caches via hydra-server 2016-02-26 17:27:30 +01:00
Eelco Dolstra b9afaadfb3 Keep better bytesReceived/bytesSent stats 2016-02-26 16:17:05 +01:00
Eelco Dolstra 6d741d2ffa Prevent download of NARs we just uploaded 2016-02-26 15:21:44 +01:00
Eelco Dolstra 02190b0fef Support hydra-build-products on binary cache stores 2016-02-26 14:45:03 +01:00
Eelco Dolstra 8e24ad6f0d Sync with Nix 2016-02-25 10:58:31 +01:00
Eelco Dolstra 8321a3eb27 Sync with Nix 2016-02-24 14:04:31 +01:00
Eelco Dolstra 7b509237cd Bleh Automake 2016-02-22 18:05:15 +01:00
Eelco Dolstra 6c3ae36648 hydra-queue-runner: Get store mode configuration from hydra.conf
To use the local Nix store (default):

  store_mode = direct

To use a local binary cache:

  store_mode = local-binary-cache
  binary_cache_dir = /var/lib/hydra/binary-cache

To use an S3 bucket:

  store_mode = s3-binary-cache
  binary_cache_s3_bucket = my-nix-bucket

Also, respect binary_cache_{secret,public}_key_file for signing the
binary cache.
2016-02-22 17:23:06 +01:00
Eelco Dolstra 94817d77d9 BinaryCacheStore: Respect build-use-substitutes 2016-02-22 17:21:39 +01:00
Eelco Dolstra 5668aa5f71 After uploading a .narinfo, add it to the LRU cache 2016-02-20 10:35:16 +01:00
Eelco Dolstra 88a05763cc Pool local store connections 2016-02-20 00:04:08 +01:00
Eelco Dolstra 1cefd6cac8 Fix log message 2016-02-20 00:02:37 +01:00
Eelco Dolstra a593ebc58e Add missing file 2016-02-19 20:58:40 +01:00
Eelco Dolstra 2b76094a23 S3BinaryCacheStore::isValidPath(): Do a GET instead of HEAD 2016-02-19 17:41:11 +01:00
Eelco Dolstra bd76f9120a Cache .narinfo lookups 2016-02-19 16:19:40 +01:00
Eelco Dolstra a0f74047da Keep some statistics for the binary cache stores 2016-02-19 14:24:23 +01:00
Eelco Dolstra dc4a00347d Use a single BinaryCacheStore for all threads
This will make it easier to do caching / keep stats. Also, we won't
have S3Client's connection pooling if we create multiple S3Client
instances.
2016-02-18 17:31:19 +01:00
Eelco Dolstra 00a7be13a2 Make queue runner internal status available under /queue-runner-status 2016-02-18 17:11:46 +01:00
Eelco Dolstra 8c9fc677c1 Typo 2016-02-18 16:43:24 +01:00
Eelco Dolstra db3fcc0f5e Enable substitution on the build machines
If properly configured, this allows them to get store paths directly
from S3, rather than having to receive them from the queue runner.
2016-02-18 16:42:05 +01:00
Eelco Dolstra 2d40888e2e Add an S3-backed binary cache store 2016-02-18 16:18:50 +01:00
Eelco Dolstra 0e254ca66d Refactor local binary cache code into a subclass 2016-02-18 14:06:17 +01:00
Eelco Dolstra a992f688d1 Rename class 2016-02-18 13:02:20 +01:00
Eelco Dolstra de77cc2910 Rename file 2016-02-18 13:02:20 +01:00
Michael Mortensen d9a4f7a211 Construct uri using uri_for, which respects X-Request-Base. 2016-02-17 21:27:44 +01:00
Eelco Dolstra ce5790285a Merge remote-tracking branch 'origin/master' into binary-cache 2016-02-17 11:54:59 +01:00
Eelco Dolstra d7a123fcd4 Keep track of the time we spend copying to/from build machines 2016-02-17 10:30:23 +01:00
Eelco Dolstra 25022bf5fd hydra-queue-runner: Support generating a signed binary cache 2016-02-16 16:41:42 +01:00
Eelco Dolstra 744cee134e hydra-queue-runner: Compress binary cache NARs using xz 2016-02-15 21:56:53 +01:00
Eelco Dolstra 2d0dd7fb49 hydra-queue-runner: Write directly to a binary cache 2016-02-15 21:10:29 +01:00
Eelco Dolstra e46acbf05b Remove the errorMsg column from the Builds table
The queue runner no longer uses this field, and it doesn't provide
very interesting historical data (mostly SSH failures), but it takes
up a lot of space. Also, it contained some bad UTF-8 which was
preventing an upgrade to Postgres 9.5, so a good occasion to get rid
of it.
2016-02-12 17:21:55 +01:00
Eelco Dolstra 5c5bb3762c Fix display of builds that don't have an eval
Like http://hydra.nixos.org/build/22872.
2016-02-12 16:27:25 +01:00
Eelco Dolstra 92d8b59361 Process Nix API changes 2016-02-11 15:59:47 +01:00
Eelco Dolstra 993647d1e3 Use Google's verifier 2016-01-14 12:54:47 +01:00
Eelco Dolstra 5a580b1bb2 Add support for logging in via a Google account
The required configuration in hydra.conf:

  enable_google_login = 1
  google_client_id = 238429sdjkds....apps.googleusercontent.com

and optionally persona_allowed_domains to restrict to one or more
domains.
2016-01-13 17:32:52 +01:00
Eelco Dolstra f11ce7e219 Bump evaluation timeout to 6 hours
This is necessary given the current size of the Nixpkgs/NixOS
jobsets. Once we have a Nix store + Postgres on SSD, we can reduce
this again.

Should really make this configurable...
2016-01-07 16:19:54 +01:00
Eelco Dolstra 0ca6bb79af Shut up warning 2016-01-07 16:19:54 +01:00
Eelco Dolstra 97f8c61928 Fix hydra-queue-runner --build-one 2015-12-29 17:53:33 +01:00
Eelco Dolstra 29db16bc69 Fix 'Can't locate object method "buildstepoutputs"'
Template::Toolkit braindamage strikes again. See
8f85fe67d4.
2015-12-15 11:55:57 +01:00
Eelco Dolstra b081133dcc Move GC roots to /nix/var/nix/gcroots/hydra
The uid split a while back caused the web interface to create GC roots
in /nix/var/nix/gcroots/per-user/hydra-www, where they wouldn't be
purged by hydra-update-gc-roots. Thus restarted builds would
accumulate forever. The fix is to keep the roots in a shared directory
with gid=hydra.
2015-12-14 13:31:24 +01:00
Eelco Dolstra e04fb5c8e7 Sort build steps 2015-12-14 13:01:35 +01:00
Eelco Dolstra 8f7614030e Better fix for dots in jobset names 2015-11-17 11:31:11 +01:00
Eelco Dolstra ea80a75204 Move getBaseUrl 2015-11-17 11:21:20 +01:00
Rob Vermaas dddb9a281d Allow dots in job specifier of input type 'Previous build' 2015-11-17 08:36:46 +00:00
Eelco Dolstra b6d2c6f03a Further fix for the channel regression 2015-11-02 14:58:23 +01:00
aszlig 20b412224e ListBuilds: Fix legacy channel URLs.
Regression introduced by 1fdc258de0.

The commit introduced a channel/custom PathPart which uses the new
custom channel expressions, but I forgot to remove CaptureArgs, so the
URL really is channel/latest/ignored-value.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Reported-by: Peter Simons <simons@cryp.to>
2015-11-02 14:51:42 +01:00
Eelco Dolstra c087472c71 Remove superfluous "has" function 2015-11-02 14:29:12 +01:00
Eelco Dolstra a5879cbeb8 Consistent terminology: "Clear" -> "Cancel" 2015-11-02 14:22:25 +01:00
Eelco Dolstra e1f9a209c5 Fix cancelling non-current builds 2015-11-02 14:20:27 +01:00
Eelco Dolstra 2d128d2a6f Don't show redundant "removing machine..." messages 2015-10-30 18:22:43 +01:00
Eelco Dolstra d8d188301d Fix division-by-zero crash
Not clear why step_->jobsets was empty...
2015-10-30 18:01:48 +01:00
Eelco Dolstra 1ff48da3d3 int2String -> std::to_string 2015-10-30 18:01:38 +01:00
Eelco Dolstra 19b52e907f Add action to restart all failed builds in an eval 2015-10-28 15:04:12 +01:00
Eelco Dolstra b9ebd38edf Fix restarting builds
Failed paths were not cleared because the $builds query becomes empty
as soon as the "finished" flag has been cleared. So this needs to be
done last.
2015-10-28 14:54:54 +01:00
Eelco Dolstra bbe1fb2f9f Ignore non-existence of IndexBuildsOnFinishedBusy
Apparently for some reason not all Hydra installations had this
index. Another -1 for non-declarative schema definition languages.
2015-10-28 14:09:45 +01:00
Eelco Dolstra 4d1816b152 Remove obsolete Builds columns and provide accurate "Running builds"
This removes the "busy", "locker" and "logfile" columns, which are no
longer used by the queue runner. The "Running builds" page now only
shows builds that have an active build step.
2015-10-27 15:37:17 +01:00
Eelco Dolstra 9fb91460e3 Restart jobset evals efficiently
Fixes

  DBIx::Class::Storage::DBI::_dbh_execute(): DBI Exception: DBD::Pg::st execute failed: ERROR: stack depth limit exceeded
  HINT: Increase the configuration parameter "max_stack_depth" (currently 2048kB), after ensuring the platform's stack depth limit is adequate. [for Statement "UPDATE Builds SET busy = ?, finished = ?, iscachedbuild = ?, locker = ? WHERE ( ( id = ? OR id = ? OR ...
2015-10-27 13:43:19 +01:00
Eelco Dolstra 53c80d9526 getQueuedBuilds(): Periodically stop to handle priority bumps
Previously, priority bumps could take a long time to get noticed if
getQueuedBuilds() was busy processing zillions of queue
additions. (This was made worse by the reintroduction of substitute
checking.)
2015-10-22 17:00:46 +02:00
Eelco Dolstra 71bf7e02d5 Use nix::willBuildLocally() 2015-10-21 15:44:29 +02:00
Eelco Dolstra 30823078c4 Merge branch 'custom-channels' of https://github.com/aszlig/hydra 2015-10-16 17:00:29 +02:00
Eelco Dolstra 4759323514 Tweak 2015-10-16 16:59:07 +02:00
dantheho 8ecd84c160 Fix schema update for schedulingShares 2015-10-16 16:58:37 +02:00
aszlig 92e85f05e2
hydra.sql: Make isChannel a not null column.
We have this set in upgrade-42.sql, so it's better to stay consistent
with the basic SQL file to avoid problems with new Hydra installations.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Reported-by: Eelco Dolstra <eelco.dolstra@logicblox.com>
2015-10-14 21:37:30 +02:00
Eelco Dolstra 1a92f971d4 Restart: Clear failed dependencies 2015-10-09 15:39:41 +02:00
Eelco Dolstra 3fa3e6414c Fix "Use of uninitialized value $mode" 2015-10-09 15:06:57 +02:00
Eelco Dolstra ca58e97691 Prevent inputs from being garbage collected during evaluation
There is still a tiny window between the calls to nix-prefetch-* and
addTempRoot. This could be eliminated by adding a "-o" option to
nix-prefetch-*, or by not using those scripts at all (and use
addToStore directly).
2015-10-09 12:50:23 +02:00
Eelco Dolstra d959afebe1 Store unset descriptions etc. as nulls 2015-10-08 12:37:56 +02:00
Eelco Dolstra 4b31b23c04 Handle attrsets in meta.license (i.e. lib.licenses) 2015-10-08 11:56:30 +02:00
Eelco Dolstra 8e8e31ce86 Re-implement log size limits
The old queue runner already had this. However, we now store "log
limit exceeded" as a separate status code in the database.
2015-10-06 17:35:08 +02:00
Eelco Dolstra 82504fe010 hydra-queue-runner: Use substitutes
This allows Hydra to use binaries from available binary caches. It
makes the queue monitor thread quite a bit slower, so if you don't
want to use binary caches, it's better to add "--option
build-use-substitutes false" to the hydra-queue-runner invocation.

Fixed #243.
2015-10-05 14:57:44 +02:00
Eelco Dolstra ae2cc61be6 Fix hydra.queue.machines.total stat
This was broken by the fact that "machines" now also contains
removed/disabled machines.
2015-09-22 12:59:19 +02:00
Eelco Dolstra f8141fdc98 Set propagatedFrom for cached failed build steps 2015-09-11 15:55:26 +02:00
Eelco Dolstra 2dc9b9e32a Don't show "localhost" as machine for cached failed build steps 2015-09-11 14:12:07 +02:00
Eelco Dolstra aea1ed7445 hydra-evaluator: Fix error handling in fetching inputs
This caused bogus "cannot find the input containing the job
expression" errors.
2015-09-11 13:49:46 +02:00
aszlig 99cbe06b0d
channel-contents: Fix display of last paragraph.
The last paragraph states about package installation of the "following"
jobs, but it only applies to generic channels, so let's only display it
there.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:39:55 +02:00
aszlig 8b4f51f657
Channel.pm: Actually deliver channel expressions.
So this is the final part which is needed in order to be able to deliver
custom channels, everything else is now just polishing.

We do this by simply redirecting to the build product download URL and
we use binary_cache_url the same way as in NixChannel.pm.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:39:55 +02:00
aszlig db0ef9e49b
Tie custom channels into channel-contents.tt.
We should now get an overview and help text on how to add a particular
channel and also a bit of information about the builds that are required
for a channel to get upgraded.

Right now we only select the latest successful build in the latest
successful evaluation, so if someone wants to have more information about
which channel has failed, (s)he still has to look at the "Channels" tab
of the jobset.

We can make this more fancy at some later point if this is really
needed, because right now we're only interested in the latest build,
because it's the only thing necessary to deliver the channel contents.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:39:55 +02:00
aszlig f1dd08afc8
build.tt: Fix reference to build.ischannel.
It's actually lower-case _despite_ the spelling in the SQL file(s),
because the schema auto-generator from DBIx::Class doesn't take it into
account because it's working on SQLite and the latter seems to ignore
case.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:39:55 +02:00
aszlig d1de12cac1
product-list: Handle channel along with "file".
We want to have contents and detauls of channel expressions as well and
we already have that in product.type == file, so why not reuse the same
for the channel expression?

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:38:45 +02:00
aszlig c9aadd8b7d
Jobset.pm: Factor out build->eval mapping.
We now have a searchBuildsAndEvalsForJobset, which creates such a
mapping for us, so we don't need to duplicate code in jobs_tab and
channels_tab.

Also, we're going to use this for the overview of a particular channel
as well, so it makes sense to put it in CatalystUtils instead of
directly in Jobset.pm.

Instead of eval->jobs, it's now eval->builds, because it's really an
aggregate over the builds schema, rather than the job schema.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:38:45 +02:00
aszlig 1fdc258de0
ListBuilds: Simplify nix channel dispatcher.
We only allow channel/latest anyway, so it really doesn't make sense to
explicitly specify this in the PathPart and provide other dispatcher
once we have more than just "latest", which greatly simplifies the
dispatch tree.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:38:44 +02:00
aszlig ed361a0145
Build.pm: Don't get isChannel using buildproducts.
We now have a column for that, so no need for counting rows which was a
bit inefficient anyway, because we only would have needed the first row
in the result.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:38:44 +02:00
aszlig ba7402ed6c
Jobset.pm: Don't show channels in "Jobs" tab.
Now that we have our dedicated "Channels" tab, there is no need anymore
to show redundant information.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:36:38 +02:00
aszlig 8f7486cd9c
Jobset.pm: Detect channels by isChannel attribute.
We now no longer need that additional join of the build outputs and can
solely use the isChannel column of the Builds table to determine whether
it's a channel build.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:36:38 +02:00
aszlig 06b76ab275
Add isChannel column and meta attribute.
This is to properly separate channels from regular jobs and also make
sure that we can always iterate on them, no matter whether the build has
failed. The reason why we were not able to do this until now was because
we were iterating on the build products, and whenever some constituent
of a channel job has failed, we didn't get a build output.

So whenever there is a meta.isHydraChannel, we can now properly
distinguish it from the other jobs.

I still don't have any clue, why "make -C src/sql update-dbix" without
*any* modifications tries to create additional schema definitions. But
I've checked the md5sums of the existing schema definitions and they
don't seem to match, so it seems that they already have been tampered
with.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:36:38 +02:00
aszlig a235de3ad7
Make channels a product type instead of subtype.
Now we can provide different channel expressions for one particular
channel build. Not sure yet how this would be useful, but I found it
more appropriate to use a type instead of a subtype of "file".

This should get us consistent with the provious commit.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:12:13 +02:00
aszlig f6bb8afc30
Show build products on aggregates with channels.
This is to get a bit more consistency among channel builds but doesn't
do a radical change on the display. Ideally we may want to have a
channel overview with all the constituents and a small help showing how
the user can add the channel.

Unfortunately, this also introduces an inconsistency: We previously used
the *subtype* "channel", but now we're expecting "channel" as the type
of the product, so we need to change this for the channels overview as
well.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:12:12 +02:00
aszlig 7b60aed5ed
jobset: Add a new "channels" tab.
It's very similar to "jobs" and the code is pretty much the same, except
that we don't do filtering on it. At least it doesn't waste space for a
filter option when there are usually WAY less channel jobs than ordinary
jobs.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:08:24 +02:00
aszlig 19e7114f0e
Add channel as a product type.
Currently I'm using a (not very well) downscaled version of the NixOS
logo, so we want to replace it by a proper image ASAP.

Other than that, the idea is to have something like this in
hydra-build-products:

file channel $out/channel.tar.bz2

Right now of course, it's only displayed at the corresponding builds, so
we might want to have aggregates on all channels for a project, jobset
or maybe even single jobs?

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
2015-09-10 17:08:24 +02:00
Eelco Dolstra c1dd3fe4be Machine status page: Show disabled machines and some machine stats 2015-09-09 16:51:43 +02:00
Eelco Dolstra 6075ac6fed Remove localhost hack 2015-09-09 16:50:59 +02:00
Eelco Dolstra ee9bf7ace7 Account steps with preferLocalBuild as a separate system type
They will show up in machineTypes as (e.g.) x86_64-linux:local instead
of x86_64-linux. This is to prevent the Hydra provisioner from
creating machines for steps that are supposed to be executed locally.
2015-09-02 13:42:25 +02:00
Eelco Dolstra 7e954aff03 Keep machine stats even when a machine is removed from the machines file
This is important for the Hydra provisioner, since it needs to be able
to see whether a disabled machine still has jobs running on it.
2015-09-02 13:31:47 +02:00
Eelco Dolstra 2d116d758a Return JSON for /eval 2015-09-02 13:13:55 +02:00
Eelco Dolstra 4f018b0612 Include the eval IDs in the JSON /build result 2015-09-02 12:57:58 +02:00
Eelco Dolstra 2b9bf7526d hydra-send-stats: Keep track of the number of machines 2015-08-27 12:05:06 +02:00
Eelco Dolstra 2a7fbd57cc Allow the machines file to specify host public keys
It's easier for the Hydra provisioner to put host public keys in the
machines file than to separately manage the known_hosts file
(especially when the provisioner runs on a different machine).
2015-08-26 13:43:02 +02:00
Eelco Dolstra 88d7eb5247 Remove double entry for localhost on machines status page 2015-08-26 13:01:45 +02:00
Eelco Dolstra 7aa52517e9 Support multiple machines files
This is primarily useful for the Hydra provisioner, which can write
its machines to another file than /etc/nix/machines.
2015-08-25 15:34:53 +02:00