From: | Steve Atkins <steve(at)blighty(dot)com> |
---|---|
To: | sfpug(at)postgresql(dot)org |
Subject: | Re: Server crash on cancelling expensive query. |
Date: | 2006-06-14 22:33:19 |
Message-ID: | F71B4E0D-FF59-4BC4-9C38-C140F7021F5F@blighty.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | sfpug |
On Jun 14, 2006, at 3:12 PM, Larry Wissink wrote:
> A developer at my company was running some queries that shouldn't
> have been run because they selected from a very large table and
> ordered by an unindexed field. The odd thing is that when the
> developer cancelled the query the whole server rebooted. (Not just
> the postgres process, the whole server!) It seems that the
> cancellation rather than the query itself caused the reboot. It
> seems like running a big query with a where clause or order by on
> fields that are not indexed and
>
> then cancelling the query after say minutes of running causes the
> server to crash and reboot. It happened several times this week
> already on several occasions. This seems not quite right.
That's not an application problem. That's a kernel bug or a hardware
failure. I'm putting money on bad hardware.
Anything interesting in the system logs? Do you have backups?
Cheers,
Steve
From | Date | Subject | |
---|---|---|---|
Next Message | David Fetter | 2006-06-14 22:45:24 | Re: Server crash on cancelling expensive query. |
Previous Message | Josh Berkus | 2006-06-14 22:29:02 | Re: Server crash on cancelling expensive query. |