New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 840588 link

Starred by 2 users

Issue metadata

Status: Fixed
Owner:
Last visit > 30 days ago
Closed: May 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug



Sign in to add a comment

cyan HWTest failure: mysql is not started on cros-full-0031.mtv.corp.google.com

Project Member Reported by philipchen@chromium.org, May 7 2018

Issue description

Owner: xixuan@chromium.org
Status: Assigned (was: Untriaged)
PTAL
Cc: -nxia@chromium.org ayatane@chromium.org
Owner: nxia@chromium.org
Summary: cyan HWTest failure: mysql is not started on cros-full-0031.mtv.corp.google.com (was: HWTest: swarming internal error)
1. Failure for veyron_mickey is just test failure: "Reason: Tests were retried."

http://cautotest-prod/afe/#tab_id=view_job&object_id=198118981

2. Failure for snappy is suite got aborted: "Reason: Suite job failed."

http://cautotest-prod/afe/#tab_id=view_job&object_id=198122975

3. Failure for cyan is "Reason: No test views found.", which means run_suite.py fails to call afe or tko db to get the test results. 

Swarming is working well mostly, but I do see a bot died during one of these tasks, that's probably the "internal error". But it will get retried automatically by another bot. So nothing to worry about it.


TODO:

For fixing 3, after checking, I think the problem is afe db is not connectable. The problematic server is "cros-full-0031.mtv.corp.google.com":

chromeos-test@cros-full-0031:~$ mysql -uroot -pautotest chromeos_autotest_db
ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)

Assign it to @deputy to fix it.


Comment 3 by nxia@chromium.org, May 9 2018

There's no mysql under /etc/init.d/ in drones, so 'service mysql start' doesn't work for drone like "cros-full-0031". Is puppet responsible for installing and starting mysql in drones ? @ayatane
Drones don't have mysql.  Drone should be connecting to master db.

Re #2 sounds like CloudSQL TKO flake from recent migration

Comment 5 Deleted

Comment 6 Deleted

Comment 7 by nxia@chromium.org, May 9 2018

if it's TKO connection flake, we've just switched the IP to the primary external IP today. 

Comment 8 by nxia@chromium.org, May 9 2018

The TKO migration has been committed into CQ. All the servers should be talking to the TKO primary external IP address now. Closing this bug for now. If there are other infra issues. Please feel free to open separate bug.

Comment 9 by nxia@chromium.org, May 9 2018

Status: Fixed (was: Assigned)
The CL which switched the IP.

https://chrome-internal-review.googlesource.com/c/chromeos/chromeos-admin/+/620845

Sign in to add a comment