Re: Small patch to replace 'idle' by 'trans' if transaction is still open

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Philip Warner <pjw@rhyme.com.au>
Cc: Christof Petig <christof.petig@wtal.de>, Bruce Momjian <pgman@candle.pha.pa.us>, pgsql-patches@postgresql.org
Date: 2000-10-11T15:15:54Z
Lists: pgsql-hackers
Philip Warner <pjw@rhyme.com.au> writes:
> The idea was to indicate if the process is likely to be causing a locking
> problem - but now I think about it, a more useful pice of information would
> be if any tables/rows/whatever are locked in such a way that another user
> might be prevented from doing an update. Is this possible?

Don't see how you could convey a useful amount of info in the space
available.  There are way too many levels of way too many locks that
might or might not conflict with someone else.  (Even a read-only
transaction grabs read locks, so it's not like the presence or absence
of any lock is sufficient info to be useful.)

There was another thread recently about adding a SHOW command or some
such to dump out the state of the lock manager's table in a readable
form.  I think that'd be a more useful thing to work on than trying
to cram one or two bits' worth of info into the ps display.

			regards, tom lane