Leo,
Thanks for the info...
(Embedded
image moved Leo Humar <lhumar AT VIC.BIGPOND.NET DOT AU>
to file: 04/13/99 05:01 AM
pic13456.pcx)
Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
cc: (bcc: Bob Barden/AAL)
Subject: Fw: TCPFLUSH ERROR
>Hi Bob
>
>I received this note from our IBM support when we had some of our
problems.
>We also reduced the amount of nodes that backup at the same time to
relieve
>the server.
>
>TcpRead: Error 10054 is returned to ADSM from the NT TCP/IP stack.
>
> The mircosoft doc and errno.h files state that an error 10054 is a
>
> TCP/IP CONNECTION DROPPED.
>
Typically this happens when one of the TCP/IP stacks could not handle
the volume or the size of the incoming data and just disconnected the
TCP/IP session. So we are most likely looking at a networking
problem
rather than an ADSM problem.
This problem may be resolved by reducing the ADSM client
tcpwindowsize
and or tcpbuffersize to a size that the NT machine can handle.
Also look at the WIN NT TCP send/receive buffsize. Also look at the
Win NT mtu size to make sure that it is compatible/comaprable with the
the mtu size in the routers
Sorry I left the good bits out
>Cheers
>
>Leo Humar
>LCS Pty Ltd
>lhumar AT vic.bigpond.net DOT au
>
>-----Original Message-----
>From: Bob Barden <Bob_Barden AT AAL DOT ORG>
>To: ADSM-L AT VM.MARIST DOT EDU <ADSM-L AT VM.MARIST DOT EDU>
>Date: Tuesday, April 13, 1999 1:21 AM
>Subject: TCPFLUSH ERROR
>
>
>>I know I have seen similar reportings of this type of error, but I have
not
>>seen any comment on the reason for this type of error.
>>
>>Example: (during scheduled nightly backup window we
receive)
>> TcpFlush: Error 10054 sending data on Tcp/Ip socket 164.
>> sessSendVerb: Error sending Verb, rc: -50
>> TcpFlush: Error 10054 sending data on Tcp/Ip socket 164.
>> sessSendVerb: Error sending Verb, rc: -50
>> ANS1809E Session is lost; initializing session reopen procedure
>> ANS1809E Session is lost; initializing session reopen procedure
>> ANS1810E ADSM session has been reestablished.
>> (then this section of error messages will continue for a second
>>socket number and so on until the
>> backup has completed or the backup window has expired......)
>>
>>This is happening on all our ADSM NT clients regardless of NT service
pack
>>installed. The ADSM clients are at 3.1.0.1. Our ADSM server is MVS at
>>3.1.1.3. Our network is an ATM backbone.
>>
>>To date we have decreased our COMMTIMEOUT to 30 minutes and this has
reduce
>>the number of clients failing because of timeout.
>>
>>Any help to identify the cause of this situation would be most
>>appreciated......
>
|