Thank you Tom for the suggestion and the workaround.
I think we will be able to hold off on the temporary workaround.
Would the 2nd processes in the deadlock show up in pg_locks? (It didn't)
An insert from another table that i created to test the datafile did show up
as blocked.
I got confirmation from developers that no other process was running at the
time.
Also, even after a reboot, the COPY to the unlogged table hung, until it was
dropped and recreated.
After the latest occurrence, we took some downtime and I have rebuilt all
unlogged tables, no recurrences yet.
--
View this message in context:
http://postgresql.1045698.n5.nabble.com/What-else-could-I-ve-done-COPY-to-unlogged-tbl-hung-locked-the-table-tp5721983p5722754.html
Sent from the PostgreSQL - admin mailing list archive at Nabble.com.
