Horton

Re: error context for vacuum to include block number

Justin Pryzby <pryzby@telsasoft.com>

View thread View on postgr.es
From: Justin Pryzby <pryzby@telsasoft.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: pgsql-hackers@postgresql.org, Andres Freund <andres@anarazel.de>
Date: 2019-12-11T14:36:48Z
Lists: pgsql-hackers

Attachments

On Wed, Dec 11, 2019 at 09:15:07PM +0900, Michael Paquier wrote:
> On Fri, Dec 06, 2019 at 10:23:25AM -0600, Justin Pryzby wrote:
> > Find attached updated patch:
> >  . Use structure to include relation name.
> >  . Split into a separate patch rename of "StringInfoData buf".
> > 
> > 2019-11-27 20:04:53.640 CST [14244] ERROR:  canceling statement due to statement timeout
> > 2019-11-27 20:04:53.640 CST [14244] CONTEXT:  block 2314 of relation t
> > 2019-11-27 20:04:53.640 CST [14244] STATEMENT:  vacuum t;
> > 
> > I tried to use BufferGetTag() to avoid using a 2ndary structure, but fails if
> > the buffer is not pinned.
> 
> No problem from me to add more context directly in lazy_scan_heap().

Do you mean without a callback ?  I think that's necessary, since the IO errors
would happen within ReadBufferExtended, but we don't want to polute that with
errcontext.  And cannot call errcontext on its own:
FATAL:  errstart was not called

> So I would suggest the following instead:
> "while scanning block %u of relation \"%s.%s\"" 

Done in the attached.

Commits

  1. Avoid calls to RelationGetRelationName() and RelationGetNamespace() in

  2. Introduce vacuum errcontext to display additional information.

  3. Fix mesurement of elapsed time during truncating heap in VACUUM.

  4. Allow vacuum command to process indexes in parallel.

  5. Refactor code dedicated to index vacuuming in vacuumlazy.c

  6. Remove duplicated progress reporting during heap scan of VACUUM