Re: Let's get the ball rolling: Hardware requirements for Zimbra

[Will Froning <wfroning@aus.edu>]

Hello Tom,

----- "Tom Golson" <tgolson@neo.tamu.edu> wrote:

> Hi Jeff,
>      Well ... that's kind of one of those things.  Right now, we're
> living and dying on the fact that we're running "highly available"
> disk.  The mailstore vendor we were going to did very little database
> magic, and simple file system backups once a week were "ok".
>      Zimbra, though, does a lot more database stuff, and simple,
> filesystem backups aren't enough.  The Zimbra backup engine though, is
> _great_.  Right now we're keeping it on the same array, but we're
> looking at putting a separate controller out there, with big, slow
> disk that we can park those backups on and get a little more
> availability.  Then, these backups we can spool off to tape
> periodically.
>      Generally speaking, though, our backup "goal" for mail is
> disaster recovery.  Mailbox-level restore is something we've only been
> able to recently consider, thanks to the Zimbra backup engine.

If you are considering dumping the backups to a SATA shelf on the NetApps, would you still be using iSCSI?  I saw some mentions on the Zimbra forums of people putting the backups on a NetApp via NFS.  The obvious benefit is that you don't even have to touch the Zimbra Servers for a backup (NDMP).

Also I've noticed a lot of NetApps being mentioned with iSCSI.  I guess you all are confident with iSCSI stability on Linux?  Not being a Linux guy (FreeBSD & Solaris lover), I'm inexperienced with iSCSI on Linux.  My current plan is to run 2 2-way Quad-Core blades with 32GB on RHEL-64 with cluster to a NetApp 3020c over FCP for the mailstore.  Hearing iSCSI mentioned makes me wonder if the reduced complexity of iSCSI shouldn't be considered for our deployment...

Thanks,
Will

> 
> --Tom
> 
> ----- Original Message -----
> From: "Jeff Wasilko" <Jeff.Wasilko@tufts.edu>
> To: "Steve Hillman" <hillman@sfu.ca>, tgolson@neo.tamu.edu
> Cc: "zimbra-hied-admins" <zimbra-hied-admins@sfu.ca>
> Sent: Wednesday, March 5, 2008 11:32:14 AM GMT -08:00 Tijuana / Baja
> California
> Subject: Re: Let's get the ball rolling: Hardware requirements for
> Zimbra
> 
> On Wed, Mar 05, 2008 at 11:25:29AM -0800, Steve Hillman wrote:
> > 
> > Zimbra doesn't seem to be computationally intensive (8 cores is kind
> of overkill), but it loves memory.  Even with more, smaller servers,
> I'm not sure how small the connecting population would have to be, per
> server, to get away with less than 8 GB of RAM on a mailstore.  We'll
> probably be adding two more LS41's to the mix, in order to give folks
> a warmer fuzzy about capacity, but we're really not doing too badly,
> right now.  We just don't have a lot of extra capacity for unexpected
> demand peaks.  But, we are delivering around 300,000 messages a day to
> the complex, and supporting peak usage accross the complex of about
> 6,000 concurrent connections.
> 
> How are backups going for you?


-- 
Will Froning
Sr. Systems Architect
American University of Sharjah
PO Box 26666
Sharjah, UAE

Tel:  +97165152124