"Tomeh, Husam" <htomeh@firstam.com> writes:
> ThanksTom. I've asked the SA to bounce the server thinking it may be a
> memory fragmentation issue on the server. After bouncing the server, I
> was able to run vacuum successfully. (Before bouncing the server, I
> tried a lower value of maintenance_work_mem of 512MB and vacuum worked
> fine which made me feel better the status of data not being corrupted).
> Do you agree with me at this point that it's not a data corruption
> issue? Thanks again.
No, I don't, unless you have fields in the database that *should* be
1-billion-and-change bytes long.
regards, tom lane