If I understand you right you need to pull 500 GB/hour.
That should be around 140 MB/s and you would need a bandwith of at least 2
Gbit/s network for that.
I don't know LTO drives, but I know that the STK 9840 and 9940 can take
that kind of data without any problem day after day.
I have seen people use STK 9840 and 9940 and getting 35 MB/s to the tape
but that is a max. and depends mostly on your data.
If I had to do it in our shop (and we only have 750 GB disk) I would use
some 8 to 10 STK 9940 drives pulling the data directly from the server over
the SAN to the tape drives.
This should work fine with Sun, but I don't know anything about SGI
Is all the data on one server and is it something you need to do daily?
/johnny
At 12:20 24-01-2002 -0500, you wrote:
>Hmm,
>
>Any sugestions to backup 5 TB over the network in 10 hours ;)
>I guess now they are supporting (or trying to) 3250 and
>Seagate... Perhaps the
>way out is direct SCSI attachment.
>
>Regards,
>
>Yucel
>
>---------------------------------------------------------
>Yucel M. Erbilgic
>erbilgic.ym AT pg DOT com
>(513) 945-1859
>
>
>From: Andrew Shinkarev <shinkara AT pprd.abbott DOT com> on 01/24/2002 05:03
>PM GMT
>
>
> Andrew Shinkarev To: Yucel
> Erbilgic-YM/PGI
> <shinkara AT pprd.abbott DOT com> Cc:
> Veritas-bu AT mailman.eng.auburn DOT edu
> Subject: Re: [Veritas-bu] SGI and
> Netbackup/
> Drive problems with
> LTO
> 01/24/2002 12:03
> PM
> Please respond to
> Andrew
> Shinkarev
>
> <shinkara AT pprd.abbott DOT com>
>
>
>
>
>
>
>
>
>On Thu, Jan 24, 2002 at 11:36:19AM -0500, erbilgic.ym AT pg DOT com wrote:
>
>Just backup you SGI boxes over network and you will be fine.
>The only supported SAN configuration by SGI is
>FC-ready STK tape drives, at least it was 6 months ago.
>
> >
> > Hi,
> >
> > Does anyone use LTO (seagate) with SGI? What about over fibre with
> > STK/Crossroads briges, qlogic card?
> >
> > I am having some problems ;)
> > 1st only one drive per bridge shows up, if there is more than one, the
> > scsi/hwpath is created but the corresponding /dev/rmt/tpsXXYYZZ is not...
>Does
> > not show up on hinv -vmm either.. (sounds like a MAKEDEV bug).
> >
> > 2nd I can do a mt -f /dev/rmt/tps** status, but unload etc. times out: As
>well
> > as the Sense information... I note the ql (qlogic probibly) rather
> than the
>tps
> > in there too... Needless to say my robotics work mighty fine.
> >
> > Jan 24 10:00:36 6A:pie unix: ql15d1l0: command timeout: 0x12, notified
> demon t
> > o abort iocb, port id 0xe8
> > Jan 24 10:00:36 6A:pie unix: ql15d1l0: command timeout: abort of iocb
> complete
> > , port id 0xe8
> > Jan 24 10:02:21 6A:pie unix: ql15d1l0: command timeout: 0x12, notified
> demon t
> > o abort iocb, port id 0xe8
> > Jan 24 10:02:21 6A:pie unix: ql15d1l0: command timeout: abort of iocb
> complete
> > , port id 0xe8
> >
> > Any one running into the same problems? Also which special file do you all
>use
> > for LTO/netbackup? /rmt/tps*****nrvc,? nrcsc etc??
> >
> > Any help would be appriciated.
> >
> > Regards,
> >
> > Yucel
> >
> >
> > ---------------------------------------------------------
> > Yucel M. Erbilgic
> > erbilgic.ym AT pg DOT com
> > (513) 945-1859
> >
> > _______________________________________________
> > Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
> > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>
>--
>----
>Andrew Shinkarev
>Network Systems Specialist
>Abbott Laboratories
>(847) 9387559
>shinkara AT pprd.abbott DOT com
>
>
>
>_______________________________________________
>Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
>http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
|