Interesting.  Thanks for the input.  LTO 100 tapes are the media being
used to write to.  The T2000 is way under utilized now which is why I
wanted to consolidate.  It will be a big boost either way since the
current backup server is a ultra 5 (don't ask).  I also have an Ultra 20
AMD 64 box that the previous sys admin thought would be a good idea for
a backup server and it is what was purchased.  I finally convinced them
that Sol 10 x86 was not the way to go.  Still working on convincing them
to stop using desktops as server(whole other battle)  I was going to use
it as a storage node but it may do better than the T2000.  I also have a
240V that might be able to be used.  Seems like that might be the best
bet.  Anyone use the Solaris AMD 64 bit server package?  I will say
T2000's rock.  I have one that is finally getting utilized and that's
because it's running Puremessage spam system the postgres backend also I
have my r&d blackboard and oracle running there too temporarily due to
limited hba cards.  The box is reaching 50% cpu usage only after I added
oracle and blackboard.  Okay let me stop plugging sun.

Ben



Jonathan Loran wrote:
> Hi Ben,
> 
> I have a few comments on the T2000 as a backup machine.  We bough a 6
> CPU 8GB RAM T2000 for our backup server.  This system is indeed an
> impressive box, as you probably know, since it sounds like you have some
> experience with it.  However, you need to be aware that as a backup
> server, it can have bottlenecks.  The single task processing power of
> the Sun T1 processor architecture is not very high.
> We were using a storage array on an iSCSI SAN as an advanced file system
> device.  We used software iSCSI on the T2000 for the initiator.  When
> backups were running full speed, the maximum throughput we saw on the
> T2000 to the iSCSI SAN was about 10-15MB/sec, and if there were more
> than 2 sessions running at a time, this would slow down to 5-7MB/sec. 
> All the while, the T2000 would only be using 10-20% of it's total CPU. 
> IO wait was always 0, the SAN was not the bottleneck.  Basically, what
> was happening was a few of the 24 processing threads on the T2000 were
> pegged, and the rest of the machine was idle.
> 
> We recently changed the backup server to a 2 CPU Sun V215, a much less
> powerful system, by all accounts, but it has about 4 times the single
> thread integer performance of the T2000.  The write speeds to the SAN
> now top out at about 30-40MB/sec, while the CPU usage on the V215 hits
> 80-90% sometimes.
> 
> I'm not sure what backup media you're storing your data to, but I will
> bet you will find the T2000 under utilized at best, and at worst, you
> may experience poor throughput.
> 
> Good luck,
> 
> Jon
> 
> Benjamin Harner wrote:
>> Hey All,
>>
>> I am in the midst of taking the big leap here and planing an upgrade
>> from 7.2.1 to 7.4.  I plan on putting it on a T2000 that runs other apps
>> now(internal admin box you could say).  I am going to move everything on
>> there now to a local zone and run networker from the global zone.
>> Anyone have a setup like this?  If so any issues or gotchas?
>>
>> Thanks
>>
>> Ben
>>
>> To sign off this list, send email to listserv AT listserv.temple DOT edu and
>> type "signoff networker" in the body of the email. Please write to
>> networker-request AT listserv.temple DOT edu if you have any problems with
>> this list. You can access the archives at
>> http://listserv.temple.edu/archives/networker.html or
>> via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
>>   
> 

To sign off this list, send email to listserv AT listserv.temple DOT edu and 
type "signoff networker" in the body of the email. Please write to 
networker-request AT listserv.temple DOT edu if you have any problems with this 
list. You can access the archives at 
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER