Re: [squid-users] Random dropping of requests

On Tue, Aug 03, 2004 at 02:33:54AM -0700, unixware wrote:
>
> try to find out where is bottleneck .
>
> 1) are you using 32 GB for cache_dir this too much big
> my. try to 8 to 10 GB for it.

There is 2GB of RAM and nothing else is running on the system so I doubt
the cache size is the problem.

> 2) set half_closed_clients off in squid.conf

Worth a try though I don't see what this has to do with the stability
issue.

> 3) verify that dns working properly. through cache
> manager

Yes, DNS is correctly working.

> 4) check that your dont have too much DISK I/O
> through iostat command

iostat isn't included with any Linux distro as far as I know. Will check
if there is a port available.

Still, I doubt the hardware is being pushed to any limit at this point.
Any other io/cpu/mem intensive tasks (such as compiling) run very well
without any slowdowns or problems.

> 5) verify that cache manager that your squid
> performing well go to Generenl Run time information
>
> Median Service Times (seconds) 5 min 60 min:
> HTTP Requests (All): 1.17732 1.17732
> Cache Misses: 1.38447 1.31166
> Cache Hits: 0.00865 0.00865
> Near Hits: 1.24267 1.24267

The proxy is down right now, but I'll check again once the service is
back.

> 6) what u get in access.log when your squid dont
> fullfil request.

Nothing. The logs just stop growing and the normal entries resume once
the service again becomes available.

-- 
A. Sajjad Zaidi
GnuPG Key ID: 0xD7AD0E13