A few weeks back one of my PostgreSQL servers logged the following errors during the nightly dump:
pg_dump: ERROR: invalid memory alloc request size 4294967293
pg_dump: SQL command to dump the contents of table "xxxx" failed: PQendcopy() failed.
pg_dump: Error message from server: ERROR: invalid memory alloc request size 4294967293
pg_dump: The command was: COPY public.xxxx (gx, rx, cx, rxx, px, lx, vx, rxxx, sx, ex, cxx, ux, dx, rarrx) TO stdout;
pg_dumpall: pg_dump failed on database "yyy", exiting
The error repeated each night for a few nights and then stopped. Unfortunately I was not in a position to pursue the
causeand effects of this error at the time they occurred.
I have not found evidence of any problems with the regular operation of the server nor any evidence of data damage and
thisserver continues to handle hundreds of thousands of transactions per day without fail.
This server is in-queue for updates but is currently running PostgreSQL 8.1.2 on i686-pc-linux-gnu, compiled by GCC gcc
(GCC)4.0.2 20050901 (prerelease) (SUSE Linux).
The table in question typically has 10,000-100,000+ inserts/day, a similar number of deletes due to a nightly
move/archiveprocess, and a handful of updates/day.
I'd appreciate any advice on how I should pursue this (even if the advice is that it is OK to ignore at this point).
Cheers,
Steve
