Re: [PATCH] SMP SNMP

On 01/02/11 04:17, Tsantilas Christos wrote:
> I am sending a new version of the patch.
> Please my comments bellow for the fixes and my comments.
>
> Also I am including a patch for Ranges. The original patch which
> included in squid-smp-snmp patch did not support 64bit range sizes which
> required by the HttpRanges.
>
>
> On 01/29/2011 03:06 AM, Amos Jeffries wrote:
>> On 29/01/11 07:04, Tsantilas Christos wrote:
>>> I am attaching the new version.
>>>
>>>
>>> On 01/27/2011 05:10 PM, Amos Jeffries wrote:
<snip>
>>>
>>>>
>>>> * PERF_SYS_CURLRUEXP again is tricky to combine. best match would be a
>>>> min() for oldest timestamp.
>>>
>>> Currently it is just return "0".
>>>
>>
>> Oh I see.
>>
>> The meaning of LRU is "oldest timestamped object in cache, if LRU
>> algorithm is used". The fact that the workers are not presenting the
>> timestamp can be left for future fix. There is a bug open for that.
>>
>> What this SMP support needs to do is aggregate via a special filter
>> equivalent to min() to retain the semantic oldest-object meaning. A
>> special one is needed that works as unsigned and ignores '0' values.
>
> I just add your comment near the related source code.
>
>>
>> As I understand it, if there are no values available from any worker we
>> should return the SNMP no-data message instead of '0'.
>
> I did not found an quick/easy way to return no-data message (or an
> example how to do it). I let it as is for now. Is it important?
>

The default switch case should handle it nicely if the hard-coded block
is "#if 0"'d out of existence.

In the range patch I think src/HttpHeaderRange.h typedef should be
uint64_t for the S parameter. Once that is fixed and any side effects it
can be commited.

+1. Please commit when those two are resolved to your liking.

Amos

-- 
Please be using
   Current Stable Squid 2.7.STABLE9 or 3.1.10
   Beta testers wanted for 3.2.0.4