Age | Commit message (Collapse) | Author |
|
cbuildbot by default bootstraps in a directory under /tmp. This CL adds
a flag that lets the user specify the top level directory under which
all bootstrapping directories should be created.
This is especially useful on mobbuild, where /tmp is mounted noexec.
This will also be used on our builders where we don't want to execute
code from /tmp.
The default behaviour remains unchanged, so that's what developers get.
BUG=chromium:490935
TEST=`cbuildbot --test-bootstrap --bootstrap-dir path/to/custom/dir \
--local -p chromiumos/chromite binhost-pre-cq`
and verify that all bootstrap re-executions use the custom
directory.
Change-Id: Iba4bcf29bc092815b30c9ccbb5a5cfa29ef1d583
Reviewed-on: https://chromium-review.googlesource.com/274913
Reviewed-by: Paul Hobbs <phobbs@google.com>
Commit-Queue: Prathmesh Prabhu <pprabhu@chromium.org>
Tested-by: Prathmesh Prabhu <pprabhu@chromium.org>
|
|
Since we are moving away from "board-specific" overlays, this renames
the relevant constants and locations so they are more generally
applicable.
BUG=brillo:918
TEST=Unit tests
Change-Id: I960b5dde91a70921f73e1ecfda03c86f6b68cd3a
Reviewed-on: https://chromium-review.googlesource.com/275004
Reviewed-by: David James <davidjames@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
|
|
Introduced by copy-paste.
BUG=brillo:918
TEST=Unit tests still work
Change-Id: Id7f43c5aae84b6c096aa065c50ff573ca5f1be89
Reviewed-on: https://chromium-review.googlesource.com/275003
Reviewed-by: David James <davidjames@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
|
|
BUG=chromium:416022
TEST=None
This reverts commit 6397d6409001da437be2a61e1edb16466b000b27.
Change-Id: I2ed4399dfcb02f75dbafb7266fe41337be2955bd
Reviewed-on: https://chromium-review.googlesource.com/275341
Reviewed-by: David James <davidjames@chromium.org>
Tested-by: Aviv Keshet <akeshet@chromium.org>
|
|
We have ignored image tests so far because some tests are difficult to
really fix, e.g. the missing symbols imported by libthread_db. But
recently we started whitelisting binaries such as this, and fixing other
tests, image tests should be stable enough to flip the bit now. This CL
de-marks ImageTestStage as a ForgivingBuilderStage.
BUG=chromium:416022
TEST=trybots on {x86,arm,amd64,mipsel-o32}-generic-full
Change-Id: I5a1460714413b3fa9c5a3893d41b56ba333d2e90
Reviewed-on: https://chromium-review.googlesource.com/275064
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Tested-by: Nam Nguyen <namnguyen@chromium.org>
Commit-Queue: Nam Nguyen <namnguyen@chromium.org>
|
|
Adds an optional stage to cbuildbot for generating and uploading GCE-ready
image tarballs. These uploads are easy to convert GCE images and then run as
GCE VM instanes.
In most cases this will be of limited utility, since ChromeOS images do
not come with SSH running by default and GCE does not provide serial
acccess. The test images, however, do run SSH on startup and can be
accessed while running on GCE.
The option is disabled by default.
For example:
$ gcloud compute images create foo-image \
--source-uri gs://${GCS_PATH}/chromiumos_test_image_gce.tar.gz
$ gcloud compute instances create foo-instance --image foo-image
$ gcloud compute ssh root@foo-instance [ --ssh_key_file=... ]
BUG=brillo:782
TEST=None
Change-Id: I4b769d3ad757dd7a8610ba6e5ce342df6e137a22
Reviewed-on: https://chromium-review.googlesource.com/264772
Commit-Queue: Amey Deshpande <ameyd@google.com>
Tested-by: Amey Deshpande <ameyd@google.com>
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
|
|
BUG=chromium:496752
TEST=Unit tests enhanced to use fake_cidb, and use a master_build_id.
Verified that unit tests are exercising new code path.
Change-Id: I89497e449e29cc2c52474e37688891a086e29c48
Reviewed-on: https://chromium-review.googlesource.com/275314
Trybot-Ready: Aviv Keshet <akeshet@chromium.org>
Tested-by: Aviv Keshet <akeshet@chromium.org>
Reviewed-by: David James <davidjames@chromium.org>
Commit-Queue: Aviv Keshet <akeshet@chromium.org>
|
|
This patch adds constants for for dev, recovery and test
image file names. It also creates dicts for mapping image types to image
file names and vice versa. Finally, it replaces references to these
strings under chromite/cbuildbot/ and chromite/lib.
There are more such references under chromite/cli/, chromite/lib/ and
chromite/scripts/, which will be fixed in a later CL.
BUG=brillo:782
TEST=Unittests and 'lumpy-paladin' trybot passed.
Change-Id: I37fb2098924a4f1196b375bea6f44b10794bbc07
Reviewed-on: https://chromium-review.googlesource.com/274920
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Commit-Queue: Amey Deshpande <ameyd@google.com>
Tested-by: Amey Deshpande <ameyd@google.com>
|
|
I originally named the module generate_chromeos_config because I
intended to only use it for generate the chromeos specific
configuration Json file.
However, we are using it for some Chrome OS specific scripts, and
won't be able to generalize them for a long time. So... rename the
module to reflect how it's being used.
BUG=None
TEST=cbuildbot/run_tests
Change-Id: I8b111c06c48181a43e01c6d494a8a493a0371718
Reviewed-on: https://chromium-review.googlesource.com/275026
Reviewed-by: David James <davidjames@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
Trybot-Ready: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
|
|
Nobody uses this new value yet, but this CL makes the site config
available to all build stages. This is prep for getting rid of the
cbuildbot_config global access.
Follow on changes will remove references in various stages by using
the new shared site_config value instead.
BUG=chromium:483566
TEST=run_tests + lint
Change-Id: I54532bdc5456cdbcc7eefa99e248a646bc5a23a9
Reviewed-on: https://chromium-review.googlesource.com/274871
Trybot-Ready: Don Garrett <dgarrett@chromium.org>
Reviewed-by: David James <davidjames@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
|
|
image_test module imports some third party modules. In order for us to
not depend on them at bootstrap time, this CL reorganizes those imports
into an image_test_lib module, and keep only the tests in image_test
module.
BUG=brillo:1158
TEST=unittests
TEST=trybots on release groups
Change-Id: Ia60e74414abf49e8f9f81fbabd28b19d59a0cadf
Reviewed-on: https://chromium-review.googlesource.com/274447
Commit-Queue: Nam Nguyen <namnguyen@chromium.org>
Tested-by: Nam Nguyen <namnguyen@chromium.org>
Trybot-Ready: Nam Nguyen <namnguyen@chromium.org>
Reviewed-by: David James <davidjames@chromium.org>
|
|
This CL adds the ability to use templates for configs, but doesn't turn
it on yet. It also adds tests for configs to ensure the names of the
configs consistently match the template names, and adjusts configs to
match. A list of changes:
- Remove some incorrectly configured "*-chrome-pfq-informational" bots.
These incorrectly configured bots were unused and building Chromium,
not Chrome, as the name implies. These were deprecated long ago in
favor of correctly configured bots.
- Renamed *-telem-chrome-pfq-informational bots to
telem-chromium-pfq-informational, since they build Chromium, not Chrome.
- Removed pre-cq-group. As a side-effect, this converts storm-pre-cq back to a
regular pre-cq rather than being a compile-only-pre-cq (the previous
code had accidentally converted storm-pre-cq to be compile-only.)
- Update amd64-generic-asan-paladin to actually be a paladin builder.
- Removed the unused 'binary' config type.
- Clarified documentation link for 'asan' builder.
This doesn't clean up every inheritance bug / messy situation in the file but
it cleans it up enough that it now passes the new tests.
Since we haven't turned templates on yet, you can see the changes made
by this CL by just looking at the config dump.
BUG=chromium:494824
TEST=New unit tests. Inspection of config dump.
Change-Id: I7d94923e27ec219d8f05bacfdd976e6531e3c921
Reviewed-on: https://chromium-review.googlesource.com/274236
Reviewed-by: David James <davidjames@chromium.org>
Tested-by: David James <davidjames@chromium.org>
|
|
This test has been broken for a while, but we never noticed because
the builders don't run with '--network'.
BUG=chromium:493981
TEST=Ran test with --network.
Change-Id: Id247e69b65804202871ad426527786a922459f1d
Reviewed-on: https://chromium-review.googlesource.com/274119
Trybot-Ready: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
Reviewed-by: Richard Barnette <jrbarnette@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
|
|
Config is an overloaded term, rename the Config class to something
more specific.
BUG=None
TEST=lint + run_tests
Change-Id: I5dbb40596fa47ad4ed50deb8eb86c22232968536
Reviewed-on: https://chromium-review.googlesource.com/274122
Trybot-Ready: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
Reviewed-by: Shuqian Zhao <shuqianz@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
|
|
This CL gets rid of all cbuildbot_config references, other than those
related to GetConfig, and OverrideConfigForTrybot.
It also removes most symbols from cbuildbot_config to prove they
aren't being used.
BUG=chromium:483566
TEST=cbuildbot/run_tests
Change-Id: Id289d1aad190d865cd4c609c5550a64f9dfdd463
Reviewed-on: https://chromium-review.googlesource.com/273913
Trybot-Ready: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
|
|
This places board toolchain overlays in /tmp/sdk-board-overlays inside
the chroot, instead of dumping them in the build root. This way they're
easier to harvest during prebuilt upload.
BUG=brillo:918
TEST=Unit tests
Change-Id: I44a1b2f7fe3a9cf6f2c46153c3d1c03fd6dac8bb
Reviewed-on: https://chromium-review.googlesource.com/273601
Reviewed-by: Gilad Arnold <garnold@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
|
|
Turns out that unmounting is not easy to get right, because the mount
point might appear busy even if unmount was successful, and so trying to
delete the directory shortly thereafter fails. One example is the use of
MountOverlayContext context inside a TempDir context in
sdk_stages.SDKPackageBoardToolchainsStage: the latter fails to delete
the directory even thought the former "successfully" unmounted it. This
fails sdk_stages_unittest.py intermittently.
To mitigate this, we let MountOverlayContext do the cleanup itself
via osutils.UmountDir, which has built-in retry logic exactly for this
purpose. Ugly, but works.
BUG=chromium:493269
TEST=cbuildbot/run_tests works
Change-Id: I394d2e3d65062a05381584ccb6606ac42ff63881
Reviewed-on: https://chromium-review.googlesource.com/273922
Reviewed-by: Gilad Arnold <garnold@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
|
|
Move this function from generate_chromeos_config to config_lib.Config,
and update all call sites.
The unittests related to this method seem to really be testing the
configuration information, not the method itself, so they remain in
generate_chromeos_config_unittest.
BUG=chromium:483566
TEST=cros lint + run_tests
Change-Id: Ia9b38b4c868c6800df77c819dcb7ba63448f7893
Reviewed-on: https://chromium-review.googlesource.com/272831
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
|
|
This may help diagnosing issues with the pre-cq-launcher, as well as
help tracing down why some counter stats seem to be getting lost.
BUG=None
TEST=None
Change-Id: Id8c199fc824e7e10f3dc6eee2e220a864cb2471b
Reviewed-on: https://chromium-review.googlesource.com/273546
Tested-by: Aviv Keshet <akeshet@chromium.org>
Trybot-Ready: Aviv Keshet <akeshet@chromium.org>
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Aviv Keshet <akeshet@chromium.org>
|
|
This new SDK stage produces per-board toolchain overlay tarballs. It
does so by mounting an overlayfs layer on top of the vanilla amd64-host
target created in a previous stage, then invoking cros_setup_toolchains
to install the necessary toolchains for each board and packaging them in
a tarball.
Currently this isn't yet wired and so has no effect on the builders.
CQ-DEPEND=CL:272933
BUG=brillo:918
TEST=Unit tests
Change-Id: I8424b3dbbb2e1db4fb30d6a97c83640709caf5fe
Reviewed-on: https://chromium-review.googlesource.com/272922
Reviewed-by: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
|
|
This CL makes a cosmetic change to the build_started stat, so that the
chromite stats namespace is not filled with entries for every config
ever launched.
It renames the pre-cq launch_count stat slightly, and adds an additional
len(plan) * len(configs) counter which should exactly correspond with
the number of trybot_launching actions in cidb. This is to help track
down an anomaly in the launch counts. (the delta= argument is also
removed from counters; statsd documentation indicates this is not
necessary, and I wonder if it is responsible for the anomaly).
BUG=None
TEST=pre-cq-launcher tryjob
Change-Id: Ie6cff9708861f25269678fba1be8d4539a54d5ae
Reviewed-on: https://chromium-review.googlesource.com/273274
Trybot-Ready: Aviv Keshet <akeshet@chromium.org>
Tested-by: Aviv Keshet <akeshet@chromium.org>
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Aviv Keshet <akeshet@chromium.org>
|
|
BUG=chromium:469860
TEST=unit tests
Change-Id: I50b7eaedeecababbb60763cfd4c0348e706585db
Reviewed-on: https://chromium-review.googlesource.com/272743
Trybot-Ready: Aviv Keshet <akeshet@chromium.org>
Tested-by: Aviv Keshet <akeshet@chromium.org>
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
|
|
We need to be able to reuse some of the tarball packaging logic in a new
(upcoming) SDK builder stage, so refactoring for this purpose.
1) Moved some constants to the global context, replacing a string
literal with a constant (SDK_TARBALL_NAME)
2) Moved tarball creation to a global function.
3) Small touch-ups.
BUG=brillo:918
TEST=sdk_stages_unittest
Change-Id: I95db70330e6d90ca78465f132c710e686f4fd852
Reviewed-on: https://chromium-review.googlesource.com/272749
Trybot-Ready: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
|
|
BUG=None
TEST=None
Change-Id: If6e8ea9ecebb49b26d34afed992e535696dfb5eb
Reviewed-on: https://chromium-review.googlesource.com/272685
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
Tested-by: Aviv Keshet <akeshet@chromium.org>
|
|
Some of the CL statuses are showing up, but not the full expected set.
Adding some logging to help diagnose.
BUG=chromium:469860
TEST=None
Change-Id: Ifc5bc370751e25128765eeb1b5909de7c84e074a
Reviewed-on: https://chromium-review.googlesource.com/272697
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
Tested-by: Aviv Keshet <akeshet@chromium.org>
|
|
This reverts commit 18c8e8fd7eb7cbebd480e053b358bb23be58a9a8.
Breaks on mipsel-o32-generic due to missing symbols in its
mesa drivers. This also isn't enabled in the CQ which means
we're going to be letting through CLs that break on canaries
and full bots. This needs to be turned on in the CQ first.
BUG=chromium:490546
Change-Id: I73f331ddc98fb5fc599cbee1bde702ee8a2383b8
Reviewed-on: https://chromium-review.googlesource.com/272365
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Tested-by: Mike Frysinger <vapier@chromium.org>
|
|
We have ignored image tests so far because some tests are difficult to
really fix, e.g. the missing symbols imported by libthread_db. But
recently we started whitelisting binaries such as this, and fixing other
tests, image tests should be stable enough to flip the bit now. This CL
de-marks ImageTestStage as a ForgivingBuilderStage.
BUG=chromium:416022
TEST=trybots on release groups, image tests should be green
Change-Id: I1548ed0e9451c0fe7bdd8151ac0f7840dc9799b1
Reviewed-on: https://chromium-review.googlesource.com/272270
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Commit-Queue: Nam Nguyen <namnguyen@chromium.org>
Trybot-Ready: Nam Nguyen <namnguyen@chromium.org>
Tested-by: Nam Nguyen <namnguyen@chromium.org>
|
|
This will disable the incremental compatibility check on the PFQ only.
We don't need to preserve incremental compatibility there as we will
provide new, compatible prebuilts to use.
BUG=chromium:489741
TEST=trybot run on binhost-pre-cq: all 4 tests are run.
TEST=trybot run on master-chromium-pfq: all 3 non-incremental tests run.
Change-Id: I566f55acdd695d91bc71324b1e23614682a19e57
Reviewed-on: https://chromium-review.googlesource.com/272199
Reviewed-by: David James <davidjames@chromium.org>
Tested-by: Bertrand Simonnet <bsimonnet@chromium.org>
Commit-Queue: Bertrand Simonnet <bsimonnet@chromium.org>
|
|
logging.warn is deprecated and we should always use logging.warning
(http://bugs.python.org/issue13235).
BUG=None
TEST=Unittests.
Change-Id: I8ca32a6f0d2d96baecb4a1e3f70d2273d1894a06
Reviewed-on: https://chromium-review.googlesource.com/272445
Reviewed-by: Christopher Wiley <wiley@chromium.org>
Trybot-Ready: Bertrand Simonnet <bsimonnet@chromium.org>
Tested-by: Bertrand Simonnet <bsimonnet@chromium.org>
Reviewed-by: Steve Fung <stevefung@chromium.org>
Commit-Queue: Bertrand Simonnet <bsimonnet@chromium.org>
|
|
This reverts commit ee6b9a60a10d6c578f72e0eb8f5027d364373f85.
BUG=chromium:489870
TEST=None
Change-Id: I15ec483865280a9c1ef7cda93a0170aa96afbbcb
Reviewed-on: https://chromium-review.googlesource.com/272284
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
Reviewed-by: David James <davidjames@chromium.org>
Tested-by: David James <davidjames@chromium.org>
|
|
Port of non-release-specific parts of:
https://chromium-review.googlesource.com/#/c/264929
https://chromium-review.googlesource.com/#/c/266093
- The 'internal_pfq_branch'-derived configs mark themselves as present
on the release waterfall.
- Allow explicitly-specified waterfall presense (i.e., don't override
with default).
- Fix bug with non-canary masters in 'sync_stages' (from release
branch).
- Fixed some formatting errors.
BUG=chromium:456203
TEST=`./cbuildbot/run_tests`
Change-Id: I1bbd849b7b3875601f86e08a714057112c8d002f
Reviewed-on: https://chromium-review.googlesource.com/265727
Trybot-Ready: Daniel Jacques <dnj@chromium.org>
Reviewed-by: Daniel Jacques <dnj@chromium.org>
Commit-Queue: Daniel Jacques <dnj@chromium.org>
Tested-by: Daniel Jacques <dnj@chromium.org>
|
|
These seem to have been squished together even though they had separate
subnames. Instead of relying on subnames, lets give them explicitly distinct
gauge stats.
BUG=chromium:469860
TEST=Ran run_tests.
Change-Id: I6df85f9785dfa6008f2d0c6a95345c2ce968dd80
Reviewed-on: https://chromium-review.googlesource.com/271551
Reviewed-by: Dan Shi <dshi@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
|
|
Pre-CQ launcher now submits each tryjob one by one (4 Pre-CQ configs ->
4 tryjobs). This is wasteful. This CL fixes that.
BUG=chromium:446554
TEST=Unit tests.
TEST=Pre-cq-launcher trybot run.
Change-Id: I2ef3f3d31ed308b73b80374dda1c3351bbf63c3a
Reviewed-on: https://chromium-review.googlesource.com/268851
Trybot-Ready: David James <davidjames@chromium.org>
Reviewed-by: Aviv Keshet <akeshet@chromium.org>
Tested-by: David James <davidjames@chromium.org>
Commit-Queue: David James <davidjames@chromium.org>
|
|
These stats measure how many CLs are in various states, and how many tryjobs
have been launched. Since a CL being in a particular state is a persistent
condition, I'm measuring those with gauges. Launching a trybot is an event
whose rate we want to track over time, and that sounds like what a counter
is for.
BUG=chromium:469860
TEST=Ran run_tests.
Change-Id: Idfac735b1b41b403972ed0d890e312ff25ebbca2
Reviewed-on: https://chromium-review.googlesource.com/271324
Reviewed-by: Dan Shi <dshi@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
|
|
This reverts CL:270559, which broke the Pre-CQ launcher.
See chromegw/i/chromeos/builders/Pre-CQ%20Launcher/builds/4722
20:48:43: ERROR: <type 'exceptions.AssertionError'>:
Traceback (most recent call last):
File "/mnt/data/b/cbuild/internal_master/chromite/cbuildbot/failures_lib.py", line 160, in wrapped_functor
return functor(*args, **kwargs)
File "/mnt/data/b/cbuild/internal_master/chromite/cbuildbot/stages/sync_stages.py", line 1601, in PerformStage
builder_run=self._run)
File "/mnt/data/b/cbuild/internal_master/chromite/cbuildbot/validation_pool.py", line 1330, in AcquirePool
if pool.AcquireChanges(gerrit_query, ready_fn, change_filter):
File "/mnt/data/b/cbuild/internal_master/chromite/cbuildbot/validation_pool.py", line 1234, in AcquireChanges
self, self.changes, self.non_manifest_changes)
File "/mnt/data/b/cbuild/internal_master/chromite/cbuildbot/stages/sync_stages.py", line 1504, in ProcessChanges
status_guage.send(status, count)
File "/mnt/data/b/cbuild/internal_master/chromite/lib/graphite_lib/stats.py", line 139, in send
statsd.Gauge.send(self, subname, value)
File "/usr/local/lib/python2.7/dist-packages/statsd/gauge.py", line 31, in send
assert isinstance(value, compat.NUM_TYPES)
AssertionError
BUG=chromium:469860
TEST=none
Change-Id: I11977afa2feb3ba2c239d561bfd23f97c6dc7154
Reviewed-on: https://chromium-review.googlesource.com/271232
Reviewed-by: David James <davidjames@chromium.org>
Tested-by: David James <davidjames@chromium.org>
|
|
These stats measure how many CLs are in various states, and how many tryjobs
have been launched. Since a CL being in a particular state is a persistent
condition, I'm measuring those with guages. Launching a trybot is an event
whose rate we want to track over time, and that sounds like what a counter
is for.
BUG=chromium:469860
TEST=Ran run_tests.
Change-Id: Ib8b4f5ed712710051dbd4193a86769d61f2aa646
Reviewed-on: https://chromium-review.googlesource.com/270559
Reviewed-by: Dan Shi <dshi@chromium.org>
Reviewed-by: Aviv Keshet <akeshet@chromium.org>
Commit-Queue: Gabe Black <gabeblack@chromium.org>
Tested-by: Gabe Black <gabeblack@chromium.org>
|
|
CQ-DEPEND=CL:270493
BUG=brillo:1032
TEST=cbuildbot/run_tests
TEST=Trybot
Change-Id: I0e48e254af76cedbc9ccb069b4b9800073678516
Reviewed-on: https://chromium-review.googlesource.com/269267
Reviewed-by: Gilad Arnold <garnold@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
|
|
BUG=brillo:1032
TEST=cbuildbot/run_tests
TEST=Trybot
Change-Id: I7403d5af1ddea24ee8caaec648c7c52f83878c73
Reviewed-on: https://chromium-review.googlesource.com/269266
Reviewed-by: Bertrand Simonnet <bsimonnet@chromium.org>
Tested-by: Gilad Arnold <garnold@chromium.org>
Commit-Queue: Gilad Arnold <garnold@chromium.org>
|
|
This method is specific to how Chrome Os build configs are generated
and used, so restrict it to the Chrome Os specific config library.
This doesn't solve the problem, but does call it out more
specifically. Eventually, anything which references
generate_chromeos_config will clearly be Chrome OS specific and easy
to identify.
BUG=chromium:483566
TEST=lint + run_tests.
Change-Id: Ia4b43ab798fc3f3d625bbb4f1387e885967bec8b
Reviewed-on: https://chromium-review.googlesource.com/270112
Tested-by: Don Garrett <dgarrett@chromium.org>
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
|
|
This CL adds an argument to the branch-util subgroup of cbuildbot to
skip pushing to remote repositories completely. This will be used by an
integration test that uses branch_util end-to-end, skipping just the
push-to-remote part.
BUG=chromium:470690
TEST=`cbuildbot --local branch-util --skip-remote-push \
--branch-name blah --version 7041.0.0`
Change-Id: I73cb1afdcc1daa9ac7ced916f9b5876e7ba7e06a
Reviewed-on: https://chromium-review.googlesource.com/270043
Tested-by: Prathmesh Prabhu <pprabhu@chromium.org>
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Prathmesh Prabhu <pprabhu@chromium.org>
|
|
By default, if a test image is built, the builder creates the
update.gz file from that test image. This is the wrong thing for
beaglebone_servo: we want the base image, not the test image.
This fixes builds to include a new config option for specifying
which build should be the source for non-paygen update payload
generation. By default, it chooses according to the old rules,
but if the option is provided it overrides the default.
BUG=None
TEST=Unit tests
Change-Id: I5f7460a614cf48f7a5cfcf8b1caf764fa9ae31fc
Reviewed-on: https://chromium-review.googlesource.com/269836
Commit-Queue: Richard Barnette <jrbarnette@chromium.org>
Tested-by: Richard Barnette <jrbarnette@chromium.org>
Reviewed-by: Richard Barnette <jrbarnette@chromium.org>
|
|
Create minimal cbuildbot_config that imports symbols from
gen_cbuildbot_config. This is done to clearly show what external
symbols are needed from cbuildbot_config.
Eventually cbuildbot_config will read a json file to populate itself,
but we need to fully understand the interface to know all of the
constraints.
BUG=chromium:483566
TEST=run_tests + linting
Change-Id: I71cb9d541df2a63efb7342f465418678bd515dc9
Reviewed-on: https://chromium-review.googlesource.com/269953
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
|
|
Sometimes the indentation of things like "Args:" is incorrect.
Add a linter check for it and fix up all the existing errors.
BUG=chromium:485005
TEST=`cros lint` on all files pass, and catches some bad cases
Change-Id: I79ea60d1e07c40d9283284bdaebfcda5af4e6246
Reviewed-on: https://chromium-review.googlesource.com/269642
Trybot-Ready: Mike Frysinger <vapier@chromium.org>
Reviewed-by: Prathmesh Prabhu <pprabhu@chromium.org>
Commit-Queue: Mike Frysinger <vapier@chromium.org>
Tested-by: Mike Frysinger <vapier@chromium.org>
|
|
Stop exporting the variable 'config' from cbuildbot_config and
instead export 'GetConfig()'.
Instead of giving every script access to the global variable 'config',
export via a function. This is to simplify upcoming changes that may
need to load the config dynamically, or handle not having one.
TEST=cbuildbot/run_tests && cros lint *.py
BUG=chromium:483566
CQ-DEPEND=CL:269775
Change-Id: Ifaa33a597907cfc18a7e5e99c218609e26eed683
Reviewed-on: https://chromium-review.googlesource.com/269770
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
|
|
The method GetManifestVersionsRepoUrl never really made sense inside
cbuildbot_config, since it's not really a config value.
So... get rid of the method altogether. It was either being used in
places where it had a static result (now replaced with constant
values), or via a wrapper in sync_stages.
Move the real implementation into sync_stages.
TEST=cbuildbot/run_tests && cros lint *.py
BUG=chromium:483566
Change-Id: Iddfee41340bfdbd63b6376469e3fbbad18361673
Reviewed-on: https://chromium-review.googlesource.com/269165
Reviewed-by: Aviv Keshet <akeshet@chromium.org>
Reviewed-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
|
|
The _config class is used in a number of unittests so officially
export it. Also makes it generally easier to search for.
TEST=cbuildbot/run_tests && cros lint *.py
BUG=chromium:483566
Change-Id: Ia5a921ff53ab3397927e19cd33de69970f88f1c9
Reviewed-on: https://chromium-review.googlesource.com/268915
Reviewed-by: Mike Frysinger <vapier@chromium.org>
Tested-by: Don Garrett <dgarrett@chromium.org>
Commit-Queue: Don Garrett <dgarrett@chromium.org>
|
|
Looks like the fwgdb script was missed in the big migration, as was the
sync_stages module (and there looks like a linting bug there).
The others are more insidious -- they weren't logging their output, but
instead using the builtin exceptions.Warning to construct an object.
That's why the linter didn't complain -- the code was valid.
BUG=brillo:600
TEST=`cros lint` is happy
TEST=`./cbuildbot/run_tests` passes
Change-Id: I45207422fbea059bdee70a2eedd120b5688b6350
Reviewed-on: https://chromium-review.googlesource.com/268383
Trybot-Ready: Mike Frysinger <vapier@chromium.org>
Reviewed-by: Ralph Nathan <ralphnathan@chromium.org>
Commit-Queue: Mike Frysinger <vapier@chromium.org>
Tested-by: Mike Frysinger <vapier@chromium.org>
|
|
The factory configs are for producing factory install shims only.
They also produce recovery images, but we really don't want them:
- factory branches have lots of hacks to support factory
- they never actually get used anywhere
- they're pure overhead to upload/sign
Change the pushimage logic to only push factory images when it's
a factory config.
BUG=chromium:482342
TEST=`./cbuildbot/run_tests` passes
TEST=`cbuildbot daisy-factory` runs pushimage just for factory images
Change-Id: Id6ad8ec8b25d944991c3b08eb8c16f36424f5097
Reviewed-on: https://chromium-review.googlesource.com/267850
Trybot-Ready: Mike Frysinger <vapier@chromium.org>
Tested-by: Mike Frysinger <vapier@chromium.org>
Reviewed-by: Hung-Te Lin <hungte@chromium.org>
Commit-Queue: Mike Frysinger <vapier@chromium.org>
|
|
Syncing to a pinned manifest ensures that we have the specified
revisions, but, unfortunately, repo won't bother to update branches.
Sync with an unpinned manifest first to ensure that branches are updated
(e.g. in case somebody adds a new branch to a repo.)
BUG=chromium:482077
TEST=Confirm that the new repo sync -n is run in a trybot run, and that
all repos are updated with new branches. Confirm that repos are not
updated with new branches without the fix.
Change-Id: I9ed27975eb58c20015d8f8e5d0a6f43baafdd2da
Reviewed-on: https://chromium-review.googlesource.com/268365
Reviewed-by: Gabe Black <gabeblack@chromium.org>
Commit-Queue: David James <davidjames@chromium.org>
Trybot-Ready: David James <davidjames@chromium.org>
Tested-by: David James <davidjames@chromium.org>
|
|
BUG=chromium:478176
TEST=Test that CQ slaves are now printing chumped changes.
Change-Id: I9b6bddbeac940d0fe16f1acc985210be5a8637fd
Reviewed-on: https://chromium-review.googlesource.com/268354
Reviewed-by: David James <davidjames@chromium.org>
Commit-Queue: David James <davidjames@chromium.org>
Trybot-Ready: David James <davidjames@chromium.org>
Tested-by: David James <davidjames@chromium.org>
|