variant: CLUSTER indexname ON tablename

Maybe there should be something about this in the docs, so that users don't get surprised when this happens and start sending stupid emails to the pgsql-bugs list

/Mikael



-----Original Message-----
From: Tom Lane [mailto:tgl@sss.pgh.pa.us]
Sent: den 17 november 2005 15:51
To: Mikael Carneholm
Cc: pgsql-bugs@postgresql.org
Subject: Re: [BUGS] Autovacuum deadlock - bug or not?


"Mikael Carneholm" <Mikael.Carneholm@WirelessCar.com> writes:
> Don't know if this is a bug or just undocumented, but it seems as you should turn off autovacuum before you run CLUSTER, otherwise you might run into a deadlock:

Which variant of CLUSTER were you using? ISTR that some of them lock
the index before the table, which is prone to deadlock against nearly
all other operations on the table (not just vacuum).

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings

"Mikael Carneholm" <Mikael.Carneholm@WirelessCar.com> writes:
> variant: CLUSTER indexname ON tablename

Hmph. Looking at the code, that should always lock the table first,
so I don't see where the problem is. Would you look up the numbers
for us --- exactly which relations were involved in the deadlock,
and (if you can tell) which process was which?

Also, what PG version is this exactly?

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly