Re: RFC: Add 'taint' field to pg_control.

Hi,

On 2018-02-28 16:16:53 -0600, Justin Pryzby wrote:

Unfortunately your list seems to raise the bar to a place I don't see us
going soon :(

> - pg_control versions used on this cluster (hopefully a full list..obviously
> not going back before PG11);

That needs arbitrary much space, that's not feasible for
pg_control. Needs to be <= 512 bytes

> - did recovery (you could use "needed recovery" instead, but then there's the
> question of how reliable that field would be);
> + or: timestamp of most recent recovery (attempt?)

What'd that be useful for?

> - index corruption detected (on system table?); Note that "CREATE IF NOT
> EXIST" doesn't avoid unique key errors on system tables, so it's not useful
> to log every ERROR on system tables.

I don't think we have a easy way to diagnose this specifically enough

> - page %u is uninitialized --- fixing

Doesn't really indicate a bug / problem.

> - here's one I just dug up: ERROR: missing chunk number 0 for toast value 10270298 in pg_toast_2619

Hm.

> - XID wraparound?

why?

> - autovacuum disabled?

why?

> - checksum failue (in system relation?)

Hm.

> - local_preload_libraries?

Hm?

> - started in single user mode or with system indices disabled?

why?

> - hit assertion or PANIC ??

We certainly don't write to persistent data in those case.

> - UPDATE/DELETE/INSERT on system table ? (or maintenance command like
> vacuum/analyze/cluster/reindex?)

Hm, the former I could see some use for, but it might end up being
pretty ugly locking and layering wise.

Greetings,

Andres Freund