Hi,
I am using squid cache as a forward caching proxy.
CONTEXT:
For my use case since:
1. the average object size is ~80KB (moreover > 32KB),
2. the proxy server has multiple cores available
3. the throughput requirement is high (upto 1Gbps)
I have configured squid to use SMP + LargeRock. I am
(resending to squid-users)
On Thu, Feb 13, 2014 at 12:19 PM, Alex Rousskov
wrote:
> On 02/12/2014 08:04 PM, Rajiv Desai wrote:
>> Hi,
>>
>> I am using squid cache as a forward caching proxy.
>>
>> CONTEXT:
>>
>> For my use case since:
>> 1
On Thu, Feb 13, 2014 at 3:02 PM, Alex Rousskov
wrote:
> On 02/13/2014 03:01 PM, Rajiv Desai wrote:
>
>> When using LargeRock, what doe the io pattern correspond to? 16KB
>> random reads if slot size is 16KB?
>
> I am not 100% sure the reads are always 16KB. Squid may reque
On Sat, Feb 15, 2014 at 10:12 AM, Dr.x wrote:
> @Rajiv Desai
>
> have u found increasing in bandwidth saving when u used large rock ??
Yes. Large rock works pretty well with multiple SMP workers (in my
limited experience for past 5 days).
I get 85% hit rate for previously read (an
I am using Squid Cache:
Version 3.HEAD-20140127-r13248
My cache dir is configured to use rock (Large rock with SMP):
cache_dir rock /mnt/squid-cache 256000 max-size=4194304
My refresh pattern is permissive to cache all objects:
refresh_pattern . 129600 100% 129600 ignore-auth
I uploaded 30 GB of
On Sun, Feb 16, 2014 at 3:39 PM, Amos Jeffries wrote:
> On 17/02/2014 11:41 a.m., Rajiv Desai wrote:
>> I am using Squid Cache:
>> Version 3.HEAD-20140127-r13248
>>
>> My cache dir is configured to use rock (Large rock with SMP):
>> cache_dir rock /mnt/squid-cache
that subsequent GETs for the same objects
after PUTs do get a cache MISS.
On Sun, Feb 16, 2014 at 3:45 PM, Rajiv Desai wrote:
> On Sun, Feb 16, 2014 at 3:39 PM, Amos Jeffries wrote:
>> On 17/02/2014 11:41 a.m., Rajiv Desai wrote:
>>> I am using Squid Cache:
>>> Ver
hInsert: Inserting Entry e:=p2DV/0x168f990*3 key
'AC671962CFC5644F4B22DA51C242DA50'
2014/02/17 00:06:55.128 kid1| ctx: exit level 0
2014/02/17 00:06:55.128 kid1| store.cc(858) write: storeWrite: writing
17 bytes for 'AC671962CFC5644F4B22DA51C242DA50'
On Sun, Feb 16, 2014 at 11:07 PM, Rajiv
(!method.respMaybeCacheable())
return false;
-Rajiv
On Mon, Feb 17, 2014 at 12:54 AM, Rajiv Desai wrote:
> FWIW, from debug logs in cache.log, it seems like PUT responses are
> being cached.
> I am fairly new to using squid so I am be completely misreading these.
> Just trying t
Hello,
Need some guidance for optimal sharing of cache amongst SMP workers
using Large rock.
Context:
I am using squid cache with 8 SMP workers and a 200 GB rock cache stored on SMP.
(Using squid-3.HEAD-20140127-r13248 which has LargeRock support to
cache objects > 32 KB).
I have set :
maximum_
Internal Data Structures:
477 StoreEntries
477 StoreEntries with MemObjects
468 Hot Object Cache Items
39070 on-disk objects
On Tue, Feb 18, 2014 at 1:52 AM, Rajiv Desai wrote:
> Hello,
>
> Need some guidance for optimal sharing of cache amongst SMP workers
> usin
On Tue, Feb 18, 2014 at 2:42 AM, Rajiv Desai wrote:
> Some more info:
>
> Following are mgr:storedir stats after back to back downloads for 4 GB
> data (ie same 2 GB twice).
> Perhaps the 477 StoreEntries with MemObjects AND 468 Hot Object
> Cache Items are not shared?
Nah
I dug deeper into the original issue I reported where some objects
were not gettings cache HITs on subsequent reads.
It seems like there is a bug where an object overwrites previously
written object:
For a similar "multiple downloads on same data via squid" test, I
directed both store and cache lo
On Tue, Feb 18, 2014 at 9:28 PM, Alex Rousskov
wrote:
> On 02/18/2014 04:11 PM, Rajiv Desai wrote:
>
>> It seems like there is a bug where an object overwrites previously
>> written object:
>
> This is not really a bug -- hash collisions do happen, as you have
> discove
When the cache is nearing capacity, will the collision rate increase
dramatically?
Is there any special handling to reduce collisions and follow lru eviction?
On Tue, Feb 18, 2014 at 10:30 PM, Rajiv Desai wrote:
> On Tue, Feb 18, 2014 at 9:28 PM, Alex Rousskov
> wrote:
>> On 02/18/2
Hi,
I use ignore-auth for caching urls with signature and that works great.
For some endpoints, signature is added as a query (which is otherwise
an "Authorization" header). Is there a way to perhaps ignore-query?
Currently, I see that replies are cached even when they have queries,
however query
16 matches
Mail list logo