Hi
        We have recently upgraded to ADSM V3 on our AIX servers, about 2 weeks
after the
upgrade the following problem started. While backing up some large files "that
appear to
be open" the backup abends. Note when we were running V2, if this occurred we
would
get a failure on the file and the backup would just continue with the next file.
        When the problem 1st started we were also having some network problems,
to fix the
network problems all switches were recycled on the backbone. After the switch
recycle
our backups stopped failing (for about 8 days and then began to fail again).
       I currently have about 10 different OS/2 clients out of about (120 OS/2
clients) that fail
(1 to 3 ) days, then run without problems for (1 to several days) and then fail
again. Note
while it appears to be the same clients they don't all fail on the same day.
      Below are the messages and below that is everything I have tried so far.
Any and all
help and suggestions welcome.

***********************************************   Messages
**********************************************************************

Server Messages:
    ANR0444W -- Protocol error on Session ###### for node _______ (OS/2) -
out-of-sequence verb
    ANR0484W -- Session ###### for node _________ terminated
                            -- Protocol violation detected
Client Messages:
    Error.log
       TCP Flush  Error 0 sending data on Tcp/IP socket 65535
        sessSendVerb: Error 0 sending verb, rc -50 from call to 'readRTn'
        connection failure
    MYSCHED.log
          While each client failed on a different file name
                    file sizes were large & similar ###,###,###
                    all files had an extension of .nsf
                    all failed in 2 minutes or less
                    all started failing between 19:50 & 20:20; however if after
the failure the
                          client is still in it's backup window it will
restart, this attempt usually
                          fails also.
                    all had one of the following message sets, on the failing
file
      ** Compression Data Grew **
      1st retry       ** Unsuccessful **
                     OR
      ** Unsuccessful **
       After the above message(s) is the failure message and backup results.

********************************** things attempted so far
************************************************

1: Run a manual incremental -- sometimes this will be successful, but most
times it also
          fails.
2: Looked at changing the number of retries, client Admin's have stated this
can't be allowed
3: ADSM support reports apar IC12138 indicates there is an error with
  the setsockopt call when the TCPBuffersize is larger than 32K
  They suggest setting it to 31K.
  While the default for TCPB on ADSM OS/2 V2 client code is 31K, I
  updated the DSM.OPT files, recycled the scheduler, and ran
  incrementals.
  Set some of the TCPB sizes to 16 and some to 31. Of the three
  incremental I tried today all three failed (scheduled backups
  will not run again until tonight.)

************************************ next ****************************

  What should I look at next, do, etc. !!!


WJNORRIS AT US.IBM DOT COM