| From: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> |
|---|---|
| To: | Brian Ghidinelli <brian(at)vfive(dot)com> |
| Cc: | pgsql-bugs(at)postgresql(dot)org |
| Subject: | Re: BUG #13970: Vacuum hangs on particular table; cannot be terminated - requires `kill -QUIT pid` |
| Date: | 2016-02-19 21:45:55 |
| Message-ID: | 20160219214555.GA150237@alvherre.pgsql |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | Postg범퍼카 토토SQL : Postg범퍼카 토토SQL 메일 링리스트 : 2016-02-19 이후 PGSQL-BUGS 21:45 |
Here's another idea: just try SELECT FOR UPDATE on the offending tuple.
As I recall, that code path has enough protections that we wouldn't try
to read the members of the multixact even if we don't see it as below
the horizon, simply by checking the infomask bits as I mentioned before.
You could try SELECT FOR UPDATE the complete table. If you had debug
symbols you could grab the specific TID that's giving you trouble from
the backtrace ... but then the table is not *that* large.
--
Álvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Brian Ghidinelli | 2016-02-20 03:16:45 | Re: BUG #13970: Vacuum hangs on particular table; cannot be terminated - requires `kill -QUIT pid` |
| Previous Message | Tom Lane | 2016-02-19 20:04:25 | Re: BUG #13974: temp_file_limit effects vacuum |