ADSM-L

Re: slow running windows 2000 client backup

2004-07-12 16:15:41
Subject: Re: slow running windows 2000 client backup
From: "Martinez, Matt" <matt AT IDEXX DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Mon, 12 Jul 2004 16:14:43 -0400
Hi Ralph,

        Make sure your network settings are correct. I have had bandwidth
problems with our network switch (cisco 6509) with auto negotiation so I had
the network guys force the port the 100 full duplex on the switch and set
the client to the same setting. Then the bandwidth problems went away, and
the client backed up normally.

Thanks,
Matt


-----Original Message-----
From: Richard Sims [mailto:rbs AT BU DOT EDU]
Sent: Monday, July 12, 2004 3:28 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: slow running windows 2000 client backup

>Thanks for your quick response.  Unfortunately we reviewed everything
>that is running on the box and came up with nothing.  We also ran
>performance monitor all night and all we see (especially during the long
>waits) is that nothing is going on (memory, processing etc).  Once we
>start transmitting data, the stats start fluctuating and look normal (as
>it should when processes are running).  This has us totally stumped...

Hi, Ralph -

A good one you have there.  You are up to date on W2000 and sounds like
you've
covered the basics in looksee.  The problem smells like a file system or
disk
problem - perhaps the latter in that the manifestation seems to involve so
many files (severe fragmentation may contribute to it).  A DOS command which
runs through the file system reporting attributes may help reveal there
things
are mired.  A commercial disk analyzer may be of help.  Personally, I would
pursue a TSM client trace.  The latest trace issuance info is in the 4.2
Trace
Facility manual.  I have notes on Client Tracing in ADSM QuickFacts.  You
can
also take cues from info in IBM's web pages, such as in
 http://www-1.ibm.com/support/docview.wss?uid=swg21158015
Such a client trace should reveal where the time is being spent.
We'd be interested in what's found on this problem.

   thanks, Richard Sims