hankiewicz@gmail.com writes:
> The list of steps to reproduce bug:
> 1) Create two database on single server instance (production and sandbox)
> 2) On production perform normal operations (we use Trac with over 50 active
> users)
> 3) on the sandbox database start explain plan of query
> 4) while explain plan is executing drop sandbox database
> 5) after recovery was completed we have discovered that sequences on
> production database had wrong values
It's normal for sequences to be advanced by a few counts (up to 32)
after a database crash. You seem to have done a "kill -9" on the
backend that was running the EXPLAIN, which would count as a crash.
So I don't see any bug here.
regards, tom lane