[Gllug] Sending lots of emails "crashes a server"

[Tom Schutzer-Weissmann]

On Wed, 21 Jul 2004 09:51:21 +0100
John Hearns <john.hearns at clustervision.com> wrote:

> On Wed, 2004-07-21 at 09:42, John Hearns wrote:
> 
> > > 
> > > What do you have in /proc/sys/fs/file-max?  

I have 65000

> > 
> > Good thought - I've been tripped up over that one on a mail server.
> > IIRC though, there should be an entry in the logs saying that this
> > has been reached.

Nope, not that I can find, I'll have another look tho'

> Keep an eye on  /proc/sys/fs/file-nr and the load average,
> then submit a batch of emails. Don't send them to us, thanks!

2006    780     65000, currently. 

Last column is the max, what about the other two? Only the middle one changes when I start new processes.

It seems you might be right that the load climbed until the server stopped responding, and then the user asked for it to be restarted. But there's a whole hour unaccounted for, which I find odd.

Thank you all for the help,
Tom
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug