Re: [HACKERS] select * from ..;vacuum crashes

From: Bruce Momjian <maillist(at)candle(dot)pha(dot)pa(dot)us>
To: t-ishii(at)sra(dot)co(dot)jp
Cc: t-ishii(at)sra(dot)co(dot)jp, pgsql-hackers(at)postgreSQL(dot)org
Subject: Re: [HACKERS] select * from ..;vacuum crashes
Date: 1998-10-12 01:16:10
Message-ID: 199810120116.VAA28615@candle.pha.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

> >You must enable Assert to see the crash.
>
> I saw the crash without assertion enabled? This is FreeBSD 2.2.6.
>
> >The cause may be because you are doing a vacuum INSIDE a transaction. I
> >think that also explains the psql -e thing, because that does both
> >commands in the same transaction.
> >
> >Perhaps we need to disable vacuum inside transactions. Vadim?
>
> FYI, it is reported that 6.3.2 does not have the crash.

OK, this is fixed now. The problem is that my new cache-use code finds
tuples by looking in the cache, getting the t_ctid value, and using
heap_fetch to get the tuple, rather than the older sequential
scan/ScanKey method.

However, when you vacuum a table, as the rows are moved, the t_ctid
changes, but there was no call to invalidate the system cache for the
row, so when you vacuum pg_class, and later use the cache to look up
something, the cache points to an old tuple.

This is fixed now, along with a little cache code cleanup that removes
some unused code that was confusing things.

--
Bruce Momjian | http://www.op.net/~candle
maillist(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bruce Momjian 1998-10-12 01:19:44 Re: [HACKERS] Problem dumping PL/pgsql functions.
Previous Message Bruce Momjian 1998-10-12 01:12:40 Re: [HACKERS] Linux and -export-dynamic