VACUUM error on CVS build 07-JAN-98

Keith Parks <emkxp01@mtcc.demon.co.uk>

From: Keith Parks <emkxp01@mtcc.demon.co.uk>
To: hackers@postgresql.org
Date: 1998-01-07T00:35:20Z
Lists: pgsql-hackers
Hi All,

I'm just investigating some regression failures on a SPARC-Linux
build of PostgreSQL from the 7th Jan CVS tree.

Although there are many failures I can't explain, the failure
of the sanity_check VACUUM has attracted my interest.

Doing a VACUUM on the regression database I'm getting:-

regression=> vacuum;
ABORT:  nodeRead: Bad type 0
regression=>

The log shows:-

DEBUG:  Rel equipment_r: Pages 1: Changed 0, Reapped 0, Empty 0, New 0; Tup 4: 
Vac 0, Crash 0, UnUsed 0, MinLen 62, MaxLen 75; Re-using: Free/Avail. Space 0/0; 
EndEmpty/Avail. Pages 0/0. Elapsed 0/0 sec.
DEBUG:  Rel iportaltest: Pages 1: Changed 0, Reapped 0, Empty 0, New 0; Tup 2: 
Vac 0, Crash 0, UnUsed 0, MinLen 120, MaxLen 120; Re-using: Free/Avail. Space 
0/0; EndEmpty/Avail. Pages 0/0. Elapsed 0/0 sec.
ABORT:  nodeRead: Bad type 0



So what comes next?

It looks like the Rel iportaltest vac'd OK.

If I VACUUM each relation individually everything seems to be OK.

If I try to vacuum a VIEW I get the same error.

regression=> vacuum toyemp;
ABORT:  nodeRead: Bad type 0
regression=> 


Anyone have any insight into this?

Keith.