Currently on large data volumes / range counts, one feed's initial scan can take days and block other feeds from (re)starting. https://github.com/cockroachdb/cockroach/pull/133789 is aimed at mitigating the latter, but we should still improve this, by eg. printing a NOTICE if the range count is such that initial scans will take forever.
Currently on large data volumes / range counts, one feed's initial scan can take days and block other feeds from (re)starting. https://github.com/cockroachdb/cockroach/pull/133789 is aimed at mitigating the latter, but we should still improve this, by eg. printing a NOTICE if the range count is such that initial scans will take forever.
Jira issue: CRDB-44198