Open knz opened 4 years ago
Seems like another instance of #50466 occurred with RocksDB (#55146).
We have marked this issue as stale because it has been inactive for 18 months. If this issue is still relevant, removing the stale label or adding a comment will keep it active. Otherwise, we'll close it in 10 days to keep the issue queue tidy. Thank you for your contribution to CockroachDB!
Context: #51162, #50466
Today storage issues a
panic
orlog.Fatal
when OS "no space left" or "read-only filesystem" occur (normal operational events) or "permission error" (sysadmin mistakenly updates perms 'under' crdb).Because of this CRL receives Sentry crash reports. This is undesirable, because there's nothing CRL can do about these situations.
Suggestion: trying to ensure that the OS errors don't get treated as a fatal error / panic, so as to not trigger Sentry reporting. If that's architecturally hard to do, I'd like to hear that - may be an argument in favor of an intermediate severity which stops the server but is not considered a reportable error.
cc @petermattis for triage
Jira issue: CRDB-3838