From: | Csaba Nagy <nagy(at)ecircle-ag(dot)com> |
---|---|
To: | Iv Ray <pobox(at)verysmall(dot)org> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: a strange error |
Date: | 2009-05-27 12:57:37 |
Message-ID: | 1243429057.4056.23.camel@PCD12478 |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Hi Ray,
On Wed, 2009-05-27 at 14:24 +0200, Iv Ray wrote:
> When the db is refreshed, the first click inside the web application
> that uses it (php 5.x, Apache 2.x), and some of the next clicks (i. e.
> the 3rd, 5th, 8th, 12th) result in the following error -
> PGSQL ERROR: server closed the connection unexpectedly
> This probably means the server terminated abnormally
> before or while processing the request.
> After two dozen of clicks and several such messages, the application
> operates normally until the next refresh.
Although I don't know the exact mechanics of how this happens, I guess
it has to do with "permanent connections" used by PHP (a kind of
connection caching), and the reused connections discovering that the
data base backend they were connected to is gone. Given that an idle
connection will not notice that the backend is gone until the next
query, it can happen well after the "data base refresh" if the cached
connection stays idle for that time. So probably you also want to clear
the connection cache of php, but for that I have no idea how can be
done.
BTW, I think it would be less disruptive if you truncate your tables
(supposing you don't need the content)... that can also be done using a
plsql script which iterates through all tables in the public schema and
truncates them, if you're concerned with changing schema... of course
you would need to lock all tables exclusively for a short time, but that
would be for sure less disruptive than deleting the DB.
Cheers,
Csaba.
From | Date | Subject | |
---|---|---|---|
Next Message | Alexander Schöcke | 2009-05-27 13:51:14 | Bloated Table |
Previous Message | Marc Schablewski | 2009-05-27 12:48:09 | Re: a strange error |