Oracle and NetApps - they can be supported via iSCSI, or via NFS - but 
Oracle has their own NFS drivers, they do not use the operating system's 
NFS drivers (at least if you want throughput). I have this issue with 
another product with an underlying database that thrashes 'getattrs' 
calls in NFS (it's Btrieve underneath the product), and both NetApps and 
EMC (with their NFS frontends) give abysmal performance. Oracle used to 
give abysmal performance too, until they wrote their own NFS drivers. 
NFS isn't inherently bad, it all depends on how you talk to it (as long 
as the server implementation is solid, as in the NetApps).

As to Exchange - $WORK has hundreds of thousands of users on a large 
Exchange installation. It does work. I access it primarily using IMAP, 
which also works, and means that I don't have to get my mail via 
Outlook, which means that I don't curse continuously.

I won't comment on the cost effectiveness of this $$$$ solution (oops, 
guess I just did!) - but with 'sufficient resources' ($ and cluefull 
staff) it can be made to work. YMMV. :-)

Exchange (along with Outlook mail/calendaring/contacts) sells extremely 
well to Dilbert's boss and his bosses' marketing and sales droid 
departments. Dilbert, on the other hand, tends to gag at the mere 
thought. Dilbert needs to learn how to educate his bosses, unless he 
wants to gag ad infinitum.

- Richard


Brad Knowles wrote:
> on 3/3/09 10:01 PM, John Clear said:
>
>   
>> At $WORK, we do Exchange over iSCSI to NetApp.  We run it over a
>> dedicated network, but the traffic is low enough (even with 500+
>> mailboxes) that it would work just fine over the shared network.
>> The Exchange admins love it.
>>     
>
> I'm not looking forward to the day when management kills the only 
> scalable mail system on campus (which hasn't seen any significant 
> upgrades in eight years and still handles ~65k users), thus forcing a 
> total of more than 20k users into Exchange, which is currently having 
> major heartburn handling about one-third that load.
>
>
> I can't hope that someone will get fired for this fiasco in progress, 
> because then the people who would be fired are the systems staff who are 
> simply doing what they are told by management, and are themselves good 
> people who are being given an impossible task.
>
> If I could be sure that the people to be fired would be the management 
> types forcing this decision for totally asinine reasons, then I'd be 
> happy to hope that multiple someones would get fired.
>
>
> Sigh....
>
>   

_______________________________________________
Discuss mailing list
[email protected]
http://lopsa.org/cgi-bin/mailman/listinfo/discuss
This list provided by the League of Professional System Administrators
 http://lopsa.org/

Reply via email to