Veritas-bu

[Veritas-bu] RE: perfomance problem using Sun E450 with SDLT drives

2003-08-08 15:29:50
Subject: [Veritas-bu] RE: perfomance problem using Sun E450 with SDLT drives
From: Karl.Rossing AT Federated DOT CA (Karl.Rossing AT Federated DOT CA)
Date: Fri, 8 Aug 2003 14:29:50 -0500
This is a multipart message in MIME format.
--=_alternative 006B199686256D7C_=
Content-Type: text/plain; charset="US-ASCII"

i had a freaky scsi problem. I managed to confirm it with scsiinfo from 
http://www.cs.toronto.edu/~jdd/

the channel was only sync'ing up at 5mb/s.

I rebooted the box and that seemed to fix the problem.

veritas-bu-admin AT mailman.eng.auburn DOT edu wrote on 08/08/2003 11:10:35 AM:

> Hello dennis
> 
> Iam using an E450 with 4 processors. also I have been running this for
> quite a while.This server is my master server and i began to notice
> performance issues only after significant increase in my catalog data.
> I ran a couple of tests
> i did a simple tar tests to all the 4 drives at the same time and
> iam seeing same slow performance with tar also.
> 
> Then i did a tar of a huge directory to /dev/null and this completed 
with
> 30MB/sec.That means that the O.S. seems to working fine.
> 
> Iam unable to find whether the problem is with the scsi subsystem or 
with
> the controller card
> 
> 
> Thanks
> vR
> 
> 
> On Fri, 8 Aug 2003, Dennis Dwyer wrote:
> 
> > A couple of questions first.
> >
> > How many processors in your E450?
> >
> > Did you just implement SDLT220 drives or have you been running them 
awhile?
> >
> > If you didn't just implement them, have they been running 
> satisfactorily and then all of a sudden started running badly?
> >
> > The reason I ask is I was considering adding SDLT drives to my 
> L700 library last year (there were 8 DLT8000's in there). The vendor
> (STK) was adamant about the fact that for SDLT you need to have one 
> processor per tape drive in order to take full advantage of its 
> performance. Perhaps someone can confirm or deny that fact for me 
> but that's was the party line from STK as recently as last October.
> >
> > So ... If you just implemented the drives and you're running 4 
> SDLT drives on a base E450 with two 400Mhz CPU's, that could be a 
> problem for you. Obviously this reasoning might not apply if they 
> had been running well and all of a sudden the performance problem 
> started (unless you had a significant increase in the amount of data
> you're transferring).
> >
> > Regards,
> > Dennis
> >
> > Dennis F. Dwyer
> > Manager, Systems Software
> > Tampa Electric Company
> >
> > (813) 225-5181  - Voice
> > (813) 275-3599  - FAX
> >
> > Visit our corporate website at www.tecoenergy.com
> >
> > The Colonel Says: "Time is not a test of the truth"
> > Translation: Just because you've always done it that way, doesn't 
> make it right
> >
> > >>> Vijaykumar Rajendra Rao <vrajendr AT mipos2.intel DOT com> 08/07/2003
> 12:52:04 PM >>>
> >
> > hello all
> >
> > i have a sun E 450 Netbackup 4.5 Fp3 master server connected to a 
Quantum
> > P4000 using SDLT220 drives. I have 4 tape devices hooked up to the 
system.
> > Iam seeing extremely slow performance while runnig 4 jobs. Iam getting
> > around 2Mb/sec per drive.. also I have tweaked a lot of NB parameters 
with
> > out much success.
> >
> >
> >
> > Please help
> >
> >
> >
> > Thanx
> >
> >
> >
> > _______________________________________________
> > Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
> >
> >
> 
> _______________________________________________
> Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
> 

--=_alternative 006B199686256D7C_=
Content-Type: text/html; charset="US-ASCII"


<br><font size=2 face="sans-serif">i had a freaky scsi problem. I managed
to confirm it with scsiinfo from http://www.cs.toronto.edu/~jdd/</font>
<br>
<br><font size=2 face="sans-serif">the channel was only sync'ing up at
5mb/s.</font>
<br>
<br><font size=2 face="sans-serif">I rebooted the box and that seemed to
fix the problem.</font>
<br>
<br><font size=2><tt>veritas-bu-admin AT mailman.eng.auburn DOT edu wrote on 
08/08/2003
11:10:35 AM:<br>
<br>
&gt; Hello dennis<br>
&gt; <br>
&gt; Iam using an E450 with 4 processors. also I have been running this
for<br>
&gt; quite a while.This server is my master server and i began to notice<br>
&gt; performance issues only after significant increase in my catalog data.<br>
&gt; I ran a couple of tests<br>
&gt; i did a simple tar tests to all the 4 drives at the same time and<br>
&gt; iam seeing same slow performance with tar also.<br>
&gt; <br>
&gt; Then i did a tar of a huge directory to /dev/null and this completed
with<br>
&gt; 30MB/sec.That means that the O.S. seems to working fine.<br>
&gt; <br>
&gt; Iam unable to find whether the problem is with the scsi subsystem
or with<br>
&gt; the controller card<br>
&gt; <br>
&gt; <br>
&gt; Thanks<br>
&gt; vR<br>
&gt; <br>
&gt; <br>
&gt; On Fri, 8 Aug 2003, Dennis Dwyer wrote:<br>
&gt; <br>
&gt; &gt; A couple of questions first.<br>
&gt; &gt;<br>
&gt; &gt; How many processors in your E450?<br>
&gt; &gt;<br>
&gt; &gt; Did you just implement SDLT220 drives or have you been running
them awhile?<br>
&gt; &gt;<br>
&gt; &gt; If you didn't just implement them, have they been running <br>
&gt; satisfactorily and then all of a sudden started running badly?<br>
&gt; &gt;<br>
&gt; &gt; The reason I ask is I was considering adding SDLT drives to my
<br>
&gt; L700 library last year (there were 8 DLT8000's in there). The vendor<br>
&gt; (STK) was adamant about the fact that for SDLT you need to have one
<br>
&gt; processor per tape drive in order to take full advantage of its <br>
&gt; performance. Perhaps someone can confirm or deny that fact for me
<br>
&gt; but that's was the party line from STK as recently as last October.<br>
&gt; &gt;<br>
&gt; &gt; So ... If you just implemented the drives and you're running
4 <br>
&gt; SDLT drives on a base E450 with two 400Mhz CPU's, that could be a
<br>
&gt; problem for you. Obviously this reasoning might not apply if they
<br>
&gt; had been running well and all of a sudden the performance problem
<br>
&gt; started (unless you had a significant increase in the amount of data<br>
&gt; you're transferring).<br>
&gt; &gt;<br>
&gt; &gt; Regards,<br>
&gt; &gt; Dennis<br>
&gt; &gt;<br>
&gt; &gt; Dennis F. Dwyer<br>
&gt; &gt; Manager, Systems Software<br>
&gt; &gt; Tampa Electric Company<br>
&gt; &gt;<br>
&gt; &gt; (813) 225-5181 &nbsp;- Voice<br>
&gt; &gt; (813) 275-3599 &nbsp;- FAX<br>
&gt; &gt;<br>
&gt; &gt; Visit our corporate website at www.tecoenergy.com<br>
&gt; &gt;<br>
&gt; &gt; The Colonel Says: &quot;Time is not a test of the truth&quot;<br>
&gt; &gt; Translation: Just because you've always done it that way, doesn't
<br>
&gt; make it right<br>
&gt; &gt;<br>
&gt; &gt; &gt;&gt;&gt; Vijaykumar Rajendra Rao &lt;vrajendr AT mipos2.intel DOT 
com&gt;
08/07/2003<br>
&gt; 12:52:04 PM &gt;&gt;&gt;<br>
&gt; &gt;<br>
&gt; &gt; hello all<br>
&gt; &gt;<br>
&gt; &gt; i have a sun E 450 Netbackup 4.5 Fp3 master server connected
to a Quantum<br>
&gt; &gt; P4000 using SDLT220 drives. I have 4 tape devices hooked up to
the system.<br>
&gt; &gt; Iam seeing extremely slow performance while runnig 4 jobs. Iam
getting<br>
&gt; &gt; around 2Mb/sec per drive.. also I have tweaked a lot of NB parameters
with<br>
&gt; &gt; out much success.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; Please help<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; Thanx<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; _______________________________________________<br>
&gt; &gt; Veritas-bu maillist &nbsp;- &nbsp;Veritas-bu AT mailman.eng.auburn 
DOT edu<br>
&gt; &gt; http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; <br>
&gt; _______________________________________________<br>
&gt; Veritas-bu maillist &nbsp;- &nbsp;Veritas-bu AT mailman.eng.auburn DOT 
edu<br>
&gt; http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu<br>
&gt; <br>
</tt></font>
--=_alternative 006B199686256D7C_=--


<Prev in Thread] Current Thread [Next in Thread>