Re: [squid-users] Re: storage rebuild slow with reiserfs

From: Matus UHLAR - fantomas <[email protected]>
Date: Fri, 4 Feb 2005 00:46:42 +0100

> Martin Marji Cermak <mc1 <at> trimedia.cz> writes:
> > I have two 36 GB, 10 rpm scsi disk dedicated to Squid.2.5.STABLE7.
> > No raid.Adaptec 29320LP Ultra320 SCSI adapter, SEAGATE ST336607LW

On 02.02 16:27, Maik Ihde wrote:
> Our setup is quite different since we have 2 15krpm Disks configured as
> a stripe set and are using diskd...

You haven't read the FAQ before you cofiguref the proxy, have you?
especially http://www.squid-cache.org/Doc/FAQ/FAQ.html#toc3.11

> > 2005/02/01 16:12:38| 0 Swapfile clashes avoided.
> > 2005/02/01 16:12:38| Took 864.1 seconds (3441.4 objects/sec).
>
> However we had a power failure today and so squid was restarted, it took
> 168 seconds (18.022,9 obj/sec) for about 3 Million entries. The Squid
> partition is about 36 GB of Size.
>
> So that's quite different. From what I have learned in the past ReiserFS
> should be faster than ext3 when it comes to lots of small files in lots
> of dirs...

That is what everybody says and it is probably true. However, I haven't
performed real checks and there is a possibility that squid catches one on
situations where ext3 is more efficient than reisersfs.

it is also true that loading and validating the cache does not say
anything on runtime performance...

-- 
Matus UHLAR - fantomas, [email protected] ; http://www.fantomas.sk/
Warning: I wish NOT to receive e-mail advertising to this address.
Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
I intend to live forever - so far so good. 
Received on Thu Feb 03 2005 - 16:46:46 MST

This archive was generated by hypermail pre-2.1.9 : Tue Mar 01 2005 - 12:00:01 MST