Re: [SQU] Tunning Diskd (squid 2.4) + Solaris 7 (Aaaarghh! :-)

From: Duane Wessels <wessels@dont-contact.us>
Date: Thu, 4 Jan 2001 15:05:23 -0700 (MST)

On Thu, 4 Jan 2001, BERTOLI Javier E. wrote:

> Hi all;
>
> First thing (and important :-): Happy new year for you all!
>
> Now to business. I have the following scenario:
>
> Server: Sun SPARC Ultra10
> 512MB RAM
> 4 x 18 GB SCSI disks for cache (each one assigned an
> individual cache_dir entry)
> SO: Solaris 7
>
> I'm currently running squid 2.4 using AsyncIO, but I think that
> using diskd should be better, as I've read previously in the list that
> squid would run a diskd process for each cache_dir entry, distributing the
> disk IO operations better. Am I right up to this point? :-)
>
> Well. The point is that when I change the squid.conf entries to
> use diskd instead of aufs, I get these errors in my logs:
>
> 2001/01/04 15:22:48| storeDiskdSend: msgsnd: (22) Invalid argument
> 2001/01/04 15:22:48| storeDiskdSend OPEN: (22) Invalid argument
> 2001/01/04 15:22:48| storeDiskdSend: msgsnd: (22) Invalid argument
> 2001/01/04 15:22:48| storeDiskdSend OPEN: (22) Invalid argument
> 2001/01/04 15:22:51| storeDiskdSend: msgsnd: (22) Invalid argument
>
> I don't see any diskd process running, and squid dies repeatedly.
> Searching in the mailing-list archive, I guess that it should be a shared
> memory configuration problem, but I can't figure out which parameter is
> wrong.
>
> Is there any kind soul that would like to help me? :-)
>
> Here is my /etc/system:
>
> set semsys:seminfo_semmap=64
> set semsys:seminfo_semmni=4096
> set semsys:seminfo_semmns=4096
> set semsys:seminfo_semmnu=4096
> set semsys:seminfo_semume=64
> set semsys:seminfo_semmsl=100
>
> set shmsys:shminfo_shmmax=536870912
> set shmsys:shminfo_shmmin=1
> set shmsys:shminfo_shmmni=4096
> set shmsys:shminfo_shmseg=32767
>
> set msgsys:msginfo_msgmap=8192
> set msgsys:msginfo_msgmax=8192
> set msgsys:msginfo_msgmnb=1048576
> set msgsys:msginfo_msgmni=256
> set msgsys:msginfo_msgssz=64
> set msgsys:msginfo_msgtql=1024

What does the output of 'ipcs -o' look like?

I got this working on Solaris once, but it was a long time ago.

I think I had to force the kernel to load the msgsys and shmsys
modules. There was some line in /etc/system where you could
force a module to be loaded.

Sorry I cant be more help. I'm not a Solaris person...

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
Received on Thu Jan 04 2001 - 15:08:08 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:57:20 MST