On Thursday 22 May 2003 20.55, Adam wrote:
> only for the /cache3 dir. My main concern is does having multiple
> cache's sharing controllers lose the gain I sought in adding extra
> separate disks in the first place?
No. As far as a Squid workload is concerned the load is on the disk
spindles, not the controller. There is no problem to run very many
Squid cache disks on the same controller.
> The FAQ says (10MB * cache_dirGB ) + cache_mem + 20MB ) * 2
> (twice the () for physical RAM) = total RAM
> That FAQ formula becomes:
> Insert known values: ((10 * xGB ) +
> 256MB + 20MB ) * 2 = 1024MB
> Sum cache_mem + extra 20MB :
You can go down a little in cache_mem if you want more disk cache.
256MB for a proxy cache is a bit overkill.. (but makes sense if you
are running an accelerator or have the memory to spare).
> (10x + 276 ) * 2 = 1024
> divide both sides by 2 to remove "*2":
> 10x + 276 = 512
> subtract 276 from both sides :
> 10x = 236
> divide both sides by 10 :
> x = 23.6 GB
>
> With the two extra cache_dirs I now have in total: 7.26GB + 3.6GB
> + 6.8GB = 18GB
> So it *seems* to me that 18GB is well under the 23.6GB limit
Yes. You should be quite well within the recommended operational
parameters for a Squid server.
> calculated above so we should be fine though the FAQ does warn that
> 64bit systems do take more RAM (our Squid 2.5STABLE2 is 32bit but
> the Solaris 8 OS is 64bit). So if I have made an egregious error,
> please let me know!
In the 64/32 bit question the important aspect is how Squid is
compiled, not the hardware buswidth.
-- Donations welcome if you consider my Free Squid support helpful. https://www.paypal.com/xclick/business=hno%40squid-cache.org If you need commercial Squid support or cost effective Squid or firewall appliances please refer to MARA Systems AB, Sweden http://www.marasystems.com/, [email protected]Received on Thu May 22 2003 - 16:58:53 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:16:53 MST