Re: [squid-users] Re: How much disk space is reclaimed by the cache replacement algorithms?

RW wrote:
> On Wed, 23 Dec 2009 13:44:54 -0900
> Chris Robertson <crobertson_at_gci.net> wrote:
>
>> RW wrote:
>>> On Wed, 23 Dec 2009 17:58:41 +0100
>>> Matus UHLAR - fantomas <uhlar_at_fantomas.sk> wrote:
>>>
>>>
>>>
>>>>> 1. How often is the replacement done?
>>>>>
>>>> when the percentage of configured cache_dir size crosses the
>>>> cache_swap_high value.
>>>>
>>> I haven't looked at the code for a long-time and I only looked at
>>> 2.x, but cache_swap_high didn't seem to play any part that I could
>>> see. It's certainly not a simple high/low watermark algorithm.
>>>
>
>>>
>> From http://www.squid-cache.org/Doc/config/cache_swap_high/...
>>
>>
>> The low- and high-water marks for cache object replacement.
>> Replacement begins when the swap (disk) usage is above the
>> low-water mark and attempts to maintain utilization near the
>> low-water mark. As swap utilization gets close to high-water
>> mark object eviction becomes more aggressive. If utilization
>> is close to the low-water mark less replacement is done each time.
>
>
> But as I said cache_swap_high doesn't actually do anything in the 2.x
> code, and given that that description has been around a long time I
> doubt anything has changed in 3.x.
>

Garbage collection occurs at specific intervals. Default is hourly.
AFAIUI (from the paper, not the code)... between high and low items are
checked for freshness according to the replacement algorithm. Above high
watermark everything marked for potential removal by the replacement
algorithm is removed regardless of staleness.

So ...

  * the low-volume cache stays low because live requests are performing
the replacement of stale objects instead of garbage collection.

  * between low and high water marks an old but fresh object will be
retained over a newer but definitely stale object.

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE7 or 3.0.STABLE20
   Current Beta Squid 3.1.0.15