Yep I know its out of date but thank you for replying anyways :)
I retried my test with the autovacuum logs turned on and confirmed that a postgresql restart fix the issue.
thanks your your help!
-HH
On Thu, Feb 26, 2015 at 11:02 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Hanns Hartman <hwhartman@gmail.com> writes: > I am running postgres 8.3.17 on a RedHat linux derivative using a mips64 > architecture.
You realize of course that 8.3 is long out of support ... > I've recently noticed some odd autovacuum behavior.
> Recently we've had some systems deployed with the system clock set to the > year 2016. Postgres was installed with that date and things were fine > until a user noticed the incorrect date. They reset the system time back > to 2015 and then we started seeing weird behavior where the autovacuum > proccess does not seem to be maintaining the disk space.
I would not be surprised if the autovacuum launcher thinks that it doesn't need to touch that database again until 2016. A quick DB restart should fix it.