You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
roachtest: extract Fatal-level log messages for easier triage
Currently, when triaging an issue that originates from a node that has failed
with a fatal error (regardless if it was being watched by a monitor or not), you
get a generic error message that will most likely require you to go to TeamCity
and download and inspect the logs. As mentioned in the linked issue, a simple
grep on the node's logs can help to identify the issue quickly and there are
cases where the roachtest failure can be categorized as an infra related flake
(e.g. clock sync).
Now after the teardown test step after downloading the artifacts, we have a new
inspectArtifacts step where we will search node logs for any fatal logs and if
found, we will log to test.log and append those logs to the error message that
is sent to the github poster.
Then in pkg/cmd/bazci/githubpost/issues/formatter_unit.go, a new formatter will
check for this message and if found, will format the markdown such that the logs
are in a separate code block for better readability.
Resolves#147360
Epic: None
Release note: None
[This test on roachdash](https://roachdash.crdb.dev/?filter=status:open%20t:.*github_test.*&sort=title+created&display=lastcommented+project) | [Improve this report!](https://github.com/cockroachdb/cockroach/tree/master/pkg/cmd/bazci/githubpost/issues)
0 commit comments