ADSM-L

Re: Wait percentages

1997-04-27 07:42:50
Subject: Re: Wait percentages
From: Dave Sanders <DSanders AT MASSMUTUAL DOT COM>
Date: Sun, 27 Apr 1997 07:42:50 -0400
Tim, thanks a lot for those suggestions.  I'll see about changing what
needs to be changed.

About those wait percentages, though.  Would you think that those
interpretations of the waits are right??  Thanks again.

----------
> From: Pittson, Timothy ,HiServ/NA <tpittson AT HIMAIL.HCC DOT COM>
> From: Pittson, Timothy ,HiServ/NA <tpittson AT HIMAIL.HCC DOT COM>
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: Wait percentages
> Date: Saturday, April 26, 1997 1:37 PM
>
> Dave,
>         A couple of things on the Netware side you can try....
>
> 1) Make sure you're running with the most current version of TCPIP for
> Netware available.  Also check to make sure the network drivers are
> current.
>
> 2) MTU - Netware breaks data down into 512 byte packets if you're more
> than 1 hop away from your mainframe. - there's a parameter you can add
> to tell Netware not to do that - I think it goes in /etc/tcpip.cfg
>
> TCPMSSInternetlimit off    (I think this is it - check the latest ADSM
> client readme)
>
> 3) Transaction related parameters - you can increase the value for
> TXNGROUPMAX on the server and TXNBYTELIMIT on the client - this will
> tell ADSM to batch more data/files in each transaction.
>
> 4) ADSM Compression - if you're running on an older server (386, 486, or
> older Pentium),  try disabling it.  Also, in Netware 4.1, it's usually a
> good idea to disable ADSM compression if you have Netware compression
> enabled.  Otherwise, ADSM will try to compress an already compressed
> file and this ends up causing a lot of retransmissions.
>
> 5) Processorutilization  - you can increase the value of this in the
> DSM.OPT client file - from experience, I've found this helps when you
> have ADSM compression enabled.
>
> Tim Pittson
> tpittson AT himail.hcc DOT com
>
> >----------
> >From:  Dave Sanders[SMTP:DSanders AT MASSMUTUAL DOT COM]
> >Sent:  Friday, April 25, 1997 2:35 PM
> >To:    ADSM-L AT VM.MARIST DOT EDU
> >Subject:       Wait percentages
> >
> >I've been looking at the performance of our backups.  I'm getting a
> >60megs/hour average transfer rate on NETWARE clients to the MVS
mainframe
> >server.  But, on my OS2 and HPUX servers, I'm getting around
240megs/hour
> >transfer rate.  Obviously, I'm concerned with the NETWARE thruput.
> >
> >I know that there have been suggestions in a general nature here,, but
> >could someone respond with NETWARE specific tweeking to be done??
> >
> >Also, when I look at the percentage of wait times; can I interpret the
> >"Idle Wait Time" as possibly being the time the session waits while the
> >client compression is working??
> >
> >Can I interpret the "Pct. Comm Wait" time as an indicator of the how
busy
> >the network is, and therefore need something to be beefed up??
> >
> >Thanks in advance for responses,,, and (please don't do attachments so
that
> >everyone can just read your response without having to bop over to some
> >other reader?????))
> >
<Prev in Thread] Current Thread [Next in Thread>