🚀 go-pugleaf

In article <C1BMM5.84v@ccu.umanitoba.ca> umcarls9@ccu.umanitoba.ca (Charles Carlson) writes:
>In <16B5D10802.MACE@HDQCMS2H.UTSD.ATT.COM> MACE@HDQCMS2H.UTSD.ATT.COM writes:
>
->After applying the latest CSDs for TCPIP and NFS, we are experiencing data
->integrity problems with NFS.  When moving data between two OS/2 machines
->with NFS (or an OS/2 machine and a VM host running WLFS/VM) using copy
->or xcopy, we get occassional hangs.  When there is no hang, the copied
->file is not the same as the original (eg, ZIP or ZOO files fail CRC,
->files do not compare with a copy brought over on floppy).  No errors
->are reported and DIR commands appear to work correctly.  If you apply
->the CSD, be careful with your data.  NFS and TCP/IP at UN32132.
-
-I had (and reported to IBM) this problem several months ago.  
-How slow, or if fast, how loaded is your NFS server?  I was running Unix
-on a 4meg 386sx-16 as a server, and when using 8000 byte NFS packets, I would
-get corrupted data after a copy or xcopy.  I didn't have any problems at all
-when I copied to/from a 386DX-25.  Dropping the packet size to 4K fixed
-the problem.  I just upgraded my server, and was going to try 8000 byte
-packets again.  There goes that idea I guess.
-
-
--- 
-/ Charles Carlson  umcarls9@ccu.Umanitoba.CA  
-\ This .sig now available on a limited edition 2 record set, for 4 easy 
-/ installments of $19.95 each. To order, please have your Visa or Mastercard
-\ ready and phone: 1-800-555-1212, or send check or MO to 200-155 Carlton St.
-

  I was told by NFS tech support that the max buffer size
  can only be 7168, the reason being that IBM NFS uses part 
  of that buffer for some internal suff.

  Has anyone ever been successful with NFS and larger ( > 7168 ) 
  buffer size?

  So what is the story with the latest CSDs (Base: 32131, NFS:32132), 
  should they be avoided?

       Thanks -- alann

   alann lopes:        alopes@ucsd.edu       -- internet
   (619) 534-5438      ALOPES@UCSD           -- bitnet