| From: | Greg Stark <gsstark(at)mit(dot)edu> |
|---|---|
| To: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
| Cc: | Gregory Maxwell <gmaxwell(at)gmail(dot)com>, Greg Stark <gsstark(at)mit(dot)edu>, Simon Riggs <simon(at)2ndquadrant(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Christopher Kings-Lynne <chriskl(at)familyhealth(dot)com(dot)au>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Reducing relation locking overhead |
| Date: | 2005-12-03 15:23:06 |
| Message-ID: | 87mzjidx2d.fsf@stark.xeocode.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> The problem is that you need to upgrade the lock at the end of the
> operation. This is very deadlock prone, and likely to abort the whole
> operation just when it's going to finish. Is this a showstopper? Tom
> seems to think it is. I'm not sure anyone is going to be happy if they
> find that their two-day reindex was aborted just when it was going to
> finish.
How likely is this really to be a problem in this particular case? Obviously
if two people try to reindex the same index they'll have a problem, but that's
not really a problem. (Postgres should probably do something to block that up
front rather than wait until the end to fail.)
Other than that case is there any other case the reindex could deadlock with?
I'm a bit hampered thinking about this because I'm not really sure exactly
what locks a reindex needs and what else takes those locks.
--
greg
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bruce Momjian | 2005-12-03 15:27:37 | Re: [HACKERS] Should libedit be preferred to libreadline? |
| Previous Message | Greg Stark | 2005-12-03 15:15:25 | Re: Reducing relation locking overhead |