Greetings,
* Don Seiler (don@seiler.us) wrote:
> Started seeing "invalid memory alloc" errors in a non-production DB (9.6.6)
> all of a sudden this afternoon. It's running on CentOS 7.4 on VMWare.
>
> I ran a pg_dump on that DB and got this error almost immediately:
>
> pg_dump: [archiver (db)] query failed: ERROR: invalid memory alloc request
> size 8830452760576
That's interesting. An invalid memory alloc request size complaint such
as that tends to be an indicator of corruption or a bug in the backend
possibly.
Certainly curious that you were able to run the query and not hit that..
Are you sure you were running it exactly the way pg_dump does? There's
a bunch of setup that pg_dump does early on, maybe log all queries when
you run the pg_dump and then in a psql session try running them..?
If you can reliably reproduce this with either pg_dump or psql, then the
next step would perhaps be jumping in with gdb to see where that error
is being hit..
Thanks!
Stephen