[squid-users] I was wondering if someone has ever tried to use a SAN\NAS as the cache backend?

[Kinkie]

Hi Eliezer,
  it depends.
The problem is not the NAS/SAN per se, but the disk access patterns.
Squid's disk access pattern, regardless the technology, is always
randomly-timed 4kb writes (in case of Rock, they are sequential, in
*ufs scattered).
If the NAS/SAN uses a write-back policy, it is possible that by the
time it decides to flush to disk, squid has written to a full stripe
and everyone will be happy (except for RAID5 or 10); this is
relatively likely in case of Rock, unlikely in case of *ufs.
But every time a write is not stripe-aligned, the NAS/SAN will have to
read and write N stripes (N >= 2 depending on the type of RAID). This
is a bit  suboptimal for the NAS/SAN in case of Rock, but it will
likely hurt the SAN/NAS performance in case of *ufs.

In case the SAN/NAS policy is not write-back but write-through, any
option (including rock) will adversely affect the SAN/NAS performance.

On Thu, Jun 25, 2015 at 2:09 PM, Eliezer Croitoru <eliezer at ngtech.co.il> wrote:
> Hello list,
>
> I was wondering if someone has ever tried to use a SAN\NAS as the cache
> backend?
> Since rock cache type\dir changed the file handling way from "lots of files
> db" into a single(and one more) cache db There is surly a way to benefit
> from nas and SAN.
>
> If someone have used san(ISCSI) or nas(NFS) for any of the cahed dirs type I
> would like to run some tests and you can help me not repeat old tests
> resolts.
>
> Thanks,
> Eliezer
>
> _______________________________________________
> squid-users mailing list
> squid-users at lists.squid-cache.org
> http://lists.squid-cache.org/listinfo/squid-users



-- 
    Francesco