Hi Jason, oddly enough the setting on or off does not affect this particular
issue. As a rule I generally enable this option on my instances that support
it. I recently tried upping the nodes to the latest generation (m3) to try
and rectify/improve this issue. Unfortunately right now m3 won't work
because we rely on a lot of space in mnt for temporary data work and the new
instances don't have much space there (though it is much faster). So I went
back to m1.large and left EBS optimized off. I'm not seeing any note-worthy
change in performance.
So I went and fired up an RDS postgres instance here. Eventually I want to
move to RDS anyways, but it's not a good short term solution to the right
now issue. Restore is running now, so I'll know within the next day or so if
this is much faster.
I'm puzzled by the change from 9.0 to 9.2 coinciding with this though.
Before the upgrade this job never had any issues. But I am 100% aware that
could be a red herring.
--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Incredibly-slow-restore-times-after-9-0-9-2-upgrade-tp5824701p5824871.html
Sent from the PostgreSQL - performance mailing list archive at Nabble.com.