Hi,
I've been experiencing some problems recovering large files to one of our SP
nodes, lots of:
Event: ANR8216W Error sending data on socket 12. Reason 32.~
Event: ANR0481W Session 11646 for node MIDSPE01 (AIX) terminated -
client did not respond within 60 seconds.~
Found out last night what the problems was. The ADSM server and SP switch
are too fast for the hard disk configuration. The ADSM server would send
about 500Mb of the file at about 35Mb/s, the client would buffer it up in
memory (512Mb) and the disks would go 100% busy as the client tries to write
the data. So, if you're seeing the above errors, it might well be an I/O
bottleneck.
Regards,
Colin Smith
|