Okay, this is weird.
I've got a server-side program that opens a backend connection to a postgres
6.5.2 database that came with Red Hat 6.1. It issues the following
statements:
BEGIN;
DECLARE qbdbportal CURSOR FOR SELECT next_number from counter WHERE name =
'local';
FETCH ALL IN qbdbportal;
CLOSE qbdbportal;
UPDATE counter SET next_number = 32531 WHERE name = 'local';
COMMIT;
and I get the following error:
ERROR: cannot write block 60 of pg_log
/var/lib/pgsql/pg_log is 491520 bytes long (which is 10x a nice, round
number in hex - C000h - by the way), and I'm at the end of my rope, so I
truncate it this way:
cat /dev/null > /var/lib/pgsql/pg_log
and try the server-side program again. It works, and pg_log is now 499712
bytes long.
What's going on here? Everything seems to work fine now - it's just that
postgres couldn't write to the file that one time or something. Am I going
to have to deal with this all the time? Did I break something by truncating
pg_log?
Thanks,
Neil