[Gluster-users] Gluster on Thumper
a s p a s i a
aspasia.sf at gmail.com
Mon Dec 1 22:41:39 UTC 2008
Apologize for my replication question - I have read through the User's
Guide and Volume Spec documentation and was happy to read about AFR
and Unify ...
I have one question - how is the Automated File Replication feature
implemented in Gluster? is this implemented sort of like an rsync?
if not, what would the underlying mechanism to perform this remote
mirror between sites, via AFR?
thanks in advance,
On Mon, Dec 1, 2008 at 12:12 PM, a s p a s i a <aspasia.sf at gmail.com> wrote:
> Hello Harshavardhana,
> Thank you for the reply and the reference on how to compile, etc....
> Have a few more questions though:
> 1. Is there a documentation that addresses replication to an offsite
> DR site so that I could deploy a GlusterFS in Site 1 and
> mirror/replicate same data on a GlusterFS in Site 2?
> 2. Any feedback or opinions on how GlusterFS's performance with very
> small millions of files (256k or less), random IO patterns?
> thanks in advance,
> On Sat, Nov 29, 2008 at 10:52 AM, Harshavardhana Ranganath
> <harsha at zresearch.com> wrote:
>> Hi Aspasia,
>> Sorry for late reply, Yes we have successfully deployed GlusterFS on
>> Sun thumper platforms and are currently working in production from months.
>> Please follow the below link to understand how to compile and prerequisites
>> for compiling GlusterFS
>> on Solaris/OpenSolaris.
>> Please let us know if you face any problems.
>> On Thu, Nov 27, 2008 at 5:08 AM, a s p a s i a <aspasia.sf at gmail.com> wrote:
>>> Hello all,
>>> Am newbie to opensolaris and SUN HW; am curious and considering the
>>> Thumper 4540/ZFS/Gluster/Unify solution to provide cluster file system
>>> services to a bunch of clients ...
>>> question is -
>>> 1. Has anyone deployed this in production? if so, what type of
>>> application environment is this appropriate for - ie., large seq IO -
>>> like rendering, HPC; how appropriate is it for random and small files?
>>> 2. Can anyone point me to any links that could give me some idea of
>>> the ideal architecture where this technology is deployed?
>>> Is there an entry level configuration that one can grow into as
>>> environment grows, YET maintain a simplified level of administration
>>> (ie., still have single namespace, no mount map problems, etc.)?
>>> Any suggestions or help will be greatly appreciated.
>>> Gluster-users mailing list
>>> Gluster-users at gluster.org
>> [y4m4 on #gluster at irc.freenode.net]
>> "Samudaya TantraShilpi"
>> Z Research Inc - http://www.zresearch.com
> A S P A S I A
> . . . . . . . . . . ..
A S P A S I A
. . . . . . . . . . ..
More information about the Gluster-users