Re: Actually...description of following...

Marc G. Fournier wrote:

> >> ICP messages? those are 'intra-proxy requests', if I recall correctly?

Inter Cache Protocol is an UDP based query/response protocol for finding
out which neughbor caches has an given object. The protocol is overrated
by many people and is often not very useful outside thigtly coupled
caches.

If you have a single cache server then there should not be any ICP
queries/responses.

> Request failure ratio: 0.00%
> HTTP requests per minute: 511.3
>
> >> Is this an average since the server was started, or how many requests
> >> over the last 60 seconds?

Don't know actually. Lets look into the code here..

   storeAppendPrintf(sentry, "\tRequest failure ratio:\t%5.2f%%\n",
        request_failure_ratio);

Looks like there is some kind of 100 points decaying average behind this
value. See the checkFailureRatio maths in client.c for details.

    storeAppendPrintf(sentry, "\tHTTP requests per minute:\t%.1f\n",
        Counter.client_http.requests / (runtime / 60.0));

So HTTP requests per minute is obviously average since startup.

> ICP messages per minute: 0.0
> Select loop called: 1651221 times, 10.985 ms avg
>
> >> Am I supposed to read anything out of this one? *raised eyebrow*

The select loop measurements are mostly aimed at us developers, and the
reading is useless unless the Squid has been under constant load.

> Cache information for squid:
> Request Hit Ratios: 5min: 52.2%, 60min: 65.2%
> Byte Hit Ratios: 5min: 25.8%, 60min: 28.9%
> Storage Swap size: 14866393 KB
>
> >> Huh?

Exacly what are you unsure about here? The last line is the amount of
data in your on-disk cache.

> Storage Mem size: 131064 KB
>
> >> This is the 128MB that I set for 'cache_mem', right?

Right.

> Memory accounted for:
> Total accounted: 234372 KB
>
> >> Again, not sure what I'm supposed to read out of this one...

It is how much of it's allocated memory Squid has accounted what it uses
it for. There is a separate cachemgr entry breaking this down in more
details.

> File descriptor usage for squid:
> Maximum number of file descriptors: 1064
> Largest file desc currently in use: 186
> Number of file desc currently in use: 143
> Files queued for open: 0
>
> >> If this goes >0, then I assume I've hit the limit on the number of
> >> file descriptors and requests to open files are pending?

No. It will go above 0 only if you use async-io where file opens are
asyncronous.

> >> would it not also be useful to have a 'Largest number of files queued
> >> for open in here? basically I'm thinking that if I look at this once a day,
> >> and, at that time, its zero...but mid-day, we got this massive spike
> >> of activity and that value went >0, it would be nice to know that...no?

There will be spikes in this value during normal activity. The thing you
need to worry about is if there is prolonged times where it is >0, but
it is not very useful as a performance measurements and is included more
to indicate what is going on in the code.

--
Henrik Nordstrom
Squid hacker