Re: [SQU] diskd stats

From: Adrian Chadd <adrian@dont-contact.us>
Date: Tue, 27 Feb 2001 15:53:43 -0700

On Tue, Feb 27, 2001, Florin Andrei wrote:
>
> I'm looking at the "DISKD Stats" menu in cachemgr.cgi. It's a cache that
> gets only a few requests per second and is up since maybe 20 hours.
> >From a few thousands operations (open, create, close...), i only have
> two fails: one open and one unlink.
>
> How significant are these fails? Should i have to worry about them?

They only happen when the diskd queue grows bigger than the diskd
queue sizes.

The whole point behind it is to "throw away" events when the disks
get too busy. This stops squid from quickly spiralling into death
when your load slightly exceeds your disks.

I bet that the fails are from when squid starts up - it tries to swapout
all the icons and it can sometimes fail. Its ok though - internal objects
are kept in memory anyway. :-)

Adrian

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
Received on Tue Feb 27 2001 - 15:57:02 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:58:14 MST