"Ed L." <pgsql@bluepolka.net> writes:
> Someone flipped a breaker switch, and evidently triggered
> corruption in one of our major clusters:
> 2005-06-06 14:31:11.984 [20124] PANIC: Invalid page header in block 22376 of 79189398
It's possible that you are one minor release short of having dodged this
problem, as I see in the 7.3.5 CVS log
2003-12-01 11:53 tgl
* src/backend/storage/buffer/bufmgr.c (REL7_3_STABLE): Force
zero_damaged_pages to be effectively ON during recovery from WAL,
since there is no need to worry about damaged pages when we are
going to overwrite them anyway from the WAL. Per recent
discussion.
This doesn't really address the question of how the page header got
clobbered in the first place, though. Did you by any chance make a dump
to see what data was in there?
regards, tom lane