Re: Meaning of transaction pg_locks?
Philip Warner <pjw@rhyme.com.au>
From: Philip Warner <pjw@rhyme.com.au>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-hackers@postgresql.org
Date: 2008-11-12T03:08:45Z
Lists: pgsql-hackers
Tom Lane wrote:
> Neither are we, because you left out all the columns that might tell
> that ...
>
The columns are actually blank....it's the other rows I left out with
the row-level locks:
925282231 | 925280527 | | 62814 | RowExclusiveLock | t
925282208 | 925280527 | | 62814 | RowExclusiveLock | t
| | 1192675195 | 62814 | ExclusiveLock | t
925282207 | 925280527 | | 62814 | RowExclusiveLock | t
925282025 | 925280527 | | 62814 | AccessShareLock | t
925282025 | 925280527 | | 62814 | RowExclusiveLock | t
925282206 | 925280527 | | 62814 | RowExclusiveLock | t
925282212 | 925280527 | | 62814 | RowExclusiveLock | t
925282210 | 925280527 | | 62814 | RowExclusiveLock | t
925282209 | 925280527 | | 62814 | RowExclusiveLock | t
925281338 | 925280527 | | 62814 | AccessShareLock | t
925281338 | 925280527 | | 62814 | RowExclusiveLock | t
925282211 | 925280527 | | 62814 | RowExclusiveLock | t
925282213 | 925280527 | | 62814 | RowExclusiveLock | t
So I assume the processes waiting on the TX were waiting for one or more
of those rows.
Now I just need to figure out why the rows were locked for such a long
time (the row level locks are mostly on one table and various indexes of
that table).