Maksim Milyutin

This issue have to be resolved after changes discussed in https://github.com/postgrespro/pg_wait_sampling/issues/43

Hi @legrandlegrand ! Currently `pg_wait_sampling` doesn't account queryId for utility statements at all (except CTAS). And this is definitely downside that have to be repaired. Thanks for your proposal to...

Good catch! Thanks for the issue. Generally, such asynchronous behavior of collector and requesting process might incur different failures.

PR https://github.com/postgrespro/pg_wait_sampling/pull/52 resolves this issue. Placing acquiring of collector lock before reading of `collector_hdr->request` would allow to complete history request, that from collector's side will fail on message queue, before...

> List of functions could be added, but i'm not sure about function editor. You can always use custom sql query tab for that Yes, sql query tab is intended...

Rebased PR on actual state of 6X_STABLE due to some critical change in https://github.com/greenplum-db/gpdb/commit/cf25bc0f0609afeedddcee2c5c1731cef2d648a8 was added recently. I'm a little bit confused about relevance of `AbortStrongLockAcquire()` call and temporary interrupt...

Hello! Any interest to this PR?

> But when the master restart, the standby should then start correctly or am I missing something? Your statement is correct when master node restarts earlier than standby one after...

> Users can already do this. IMO manual (or automatic from external script) postmaster killing looks like duct tape. Users have to has the more convenient way to restart PostgreSQL...

Not exactly, the problem here that standby relies on config data from control file on startup stage to determine that critical to restart parameters satisfy invariant from doc: > For...