As mentioned before, try the new version of calamaris (v2.99). There is
a new report included, which shows you in the 'Requested extensions'
report some nice information about object freshness:
- ratio fresh/stale
- ratio unmod/mod
This helps you to improve your squid refresh_pattern.
Michael
On Thu, 2004-12-02 at 22:39, Adam Aube wrote:
> Martin Marji Cermak wrote:
>
> > I have been playing with Squid under a heavy load and there are some
> > stats. I am trying to maximise the "Byte Hit Ratio" value. I got 13%
> > average, but I am not happy about this number - I want it higher
>
> To increase your byte hit ratio, you can:
>
> 1) Switch to one of the heap cache replacement policies
> 2) Tune your refresh_pattern settings to make Squid cache more aggressively
>
> See the FAQ and default squid.conf for details on these items.
>
> However, before going through the tuning, run an analysis tool (such as
> Calamaris) on your logs to see what your traffic pattern is like. This will
> show you what a reasonable byte hit ratio would be.
>
> If, for example, 70% of your traffic is dynamic content (which usually
> cannot be cached), then a 13% byte hit ratio is actually pretty good.
>
> > USED HARDWARE:
> > Processor: P4 1.8GHz
> > Memory: 1 GB
> > Hardisk: 40 GB IDE 7200rpm
>
> > Requests: 180 req/sec (peak), 60 req/sec (day average).
>
> According to posts from Squid developers, a single caching Squid box has an
> upper limit of about 300 - 400 requests/second. This isn't too bad,
> considering you are using a single IDE disk for the entire system.
>
> > maximum_object_size 51200 KB (SHOULD I MAKE IT HIGHER ???)
>
> Actually, you might want to make it lower. Most web requests will not be for
> 50 MB files, and your byte hit ratio will be hurt if a 50 MB file that is
> requested once forces out fifty 1 MB files that are accessed twice each.
>
> The default is generally acceptable, unless log analysis shows large numbers
> of requests for larger files.
>
> > cache_dir aufs /cache 25000 16 256
>
> You should size your cache to hold about a week's worth of traffic. Just
> watch your memory usage (1 GB of cache ~ 10 MB of memory for metadata).
>
> > cache_mem 8 MB
>
> This is generally fine - the OS will generally use free memory to cache
> files anyway, which will have the same effect as boosting this setting.
>
> > I am going to install a new box with SCSI disks so I will report to you
> > how the performance will change.
>
> Best disk performance will be achieved with multiple small, fast SCSI disks
> dedicated to Squid's cache, each with its own cache_dir (no RAID), and
> round-robin between the cache_dirs.
>
> Adam
Received on Fri Dec 03 2004 - 03:54:04 MST
This archive was generated by hypermail pre-2.1.9 : Sat Jan 01 2005 - 12:00:01 MST