Hi
Thanks for your response. What other variables are there in the
squid.conf that are configurable besides the cache_mem value
that determines the amount of memory squid uses on your system?
Do you know how to find the amount of memory used via mallinfo()?
as I compiled using --enable-dlmalloc.
Thanks, any help appreciated
On Wed, 12 Jul 2000 23:59:44 +0200, you wrote:
>Hillel Bilman wrote:
>
>> I'm running Squid 2.3 Stable 3 on Sun420R, with Solaris 7 in http
>> Memory: 2048M real, 172M free, 75M swap in use, 10G swap free
>>
>> Any ideas why top says 172M is free but I'm using swap memory?
>>
>> I'm using 400M as my cache_mem value. What's the best way to
>> find out if it's not being used, and forcing memory to be unavailable
>> to the system?
>
>On one hand, it is totally legal with Solaris to have the system
>report that some swap is used - as long as it is not swapping.
>Check with the manpage for vmstat, and run vmstat accordingly.
>Also check with iostat your swap devices.
>
>On the other hand, 400 MB cache_mem sounds *way* too much -
>depending on the number of objects and other factors,
>cache_mem is between 1/3 to 1/10 of the memory squid actually
>needs (rule of thumb). I think it is repeatedly and again
>mentioned every few weeks on the list that cache_mem is not
>the upper memory limit for Squid. it is rather one (of many)
>limits for a tiny fraction of the memory Squid needs.
Received on Fri Jul 14 2000 - 07:16:11 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:54:31 MST