[Gluster-users] Some benchmarks for anyone that's interested..

[lejeczek]

glusterfs is a distributed file system, fair enough, easy to 
maintain and very friendly to the user
still, comparing it against a raw (local) file system, like 
I do via local mount point back ended with a single brick 
volume would be a valid route to see what glusterfs does 
when most of the variables are out of the equation.
I mean a basic logic one would follow is, unless a volume is 
a smartly distributed it would slow down even more (with 
some formula) as soon as other media get involved
thus I believe for simpler scenarios glusterfs won't do, for 
instance one would like to run a live replica of a storage,
a glusterfs two bricks replicated vol VS even only 
bidirectional lsyncd
lsyncd wins by miles, even for very deep data trees with 
lots of files

all may appreciate great bonus of clear and easy maintenance 
gluster offers (yet still no AFR-like setups with command 
utils possible) which is important for more complex 
configurations, for simpler ones this bonus does not 
outweigh poor performance gluster suffers from, well, in my 
opinion.

thanks



On 02/05/12 13:09, Amar Tumballi wrote:
> On 05/02/2012 02:22 PM, lejeczek wrote:
>> thanks for posting
>> I'd be curious to see what kind of disproportion you get 
>> between:  raw
>> fs / single brick volume with local fuse mountpoint which 
>> effectively
>> points back to the same raw fs
>> from my quick tests I saw massive gap between the two
>> thanks
>>
>>>
>>> Tests are:
>>> Single Disk (direct, no gluster)
>>> Gluster Replicated
>>> Gluster Striped Replicated
>>> Gluster Distributed Replicated
>>> Gluster Stripe
>>>
>
> Hi All,
>
> I would like to clarify few things before some one does 
> performance runs on GlusterFS.
>
> First of all, GlusterFS is not designed/intended to be 
> used as a local filesystem, ie, without n/w in picture it 
> should not be used for any kind of benchmark. Please do 
> let us know the exact use cases to use GlusterFS without 
> n/w in picture, and we can consider that in our designs.
>
> If you are comparing GlusterFS's performance to your local 
> file system (like XFS/ext4/btrfs etc), performance numbers 
> would look bad, for sure (at least for short future).
>
> This is the main reason, we recommend understanding the 
> use-case before deploying GlusterFS. Try to run with 
> similar workload on the setup to run benchmarks, because 
> the pattern of fops, type of volume, type of hardware/ 
> type of network, all of these has a effect on benchmark 
> numbers you would get.
>
>
> Regards,
> Amar
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20120510/dde01f50/attachment.html>