"netstat -M" / nfs question

[zaxxon]

Hi all,

I am looking for an explanation of the output of "netstat -M".
Nor in the man wether in any publications I found is an explanation of the columns and which relating parameters are in relationship.

We got problems with some nfs-mounts sometimes on this AIX 5.3 box, where on the server following shows up in errpt:

nfs_server:  bad  sendreply

I also found this:

http://publib.boulder.ibm.com/infoc...ommadmn/doc/commadmndita/nfs_servmessages.htm

which led me to "netstat -M" and other points of interesst.

Our "netstat -M" shows following (only 1st 10 rows, showing lots of "Failed" already:

Cluster pool Statistics
  Cluster Size   Pool Size   Calls   Failed   Inuse   Max Outcount
          2048        2048 6075022        0    1025           2048
          2048        2048 114439560        0    1024           2048
          2048        2048 10438391        0    1024           2048
          2048        2048 277342473        0    1025           2048
           512         512 3189555   840746     512            522
          2048         512 18654401  1141388     512           2052
         16384         128     129        1     128            128
         32768          24      25        1      24             24

I have only limited possibilities to test anything since we have to wait some days until the nfs-problem shows up. Until then it runs smooth. Also I can't give much information about the client and behaviour on these shares until next week, since I am only trying to help in that case an being not the main administrator of this/these box(es).
nfsstat shows errors too, but I have no idea where to search further.

Thanks

laters
zaxxon

 

[zaxxon]

Just got additional info:
The 2nd box that got the shares mounted will not allow anything related with the nfs-FS'es from the box that is the nfs-server.
Only method for colleagues to handle this is rebooting the nfs-client machine.
They say it can be reproduced when 2 programs are writing into those shares (I guess especially when they write into the same FS). Sounds like a locking problem maybe, not sure.
They use nfs3 - was only able to see nfs3 statistics in nfsstat. No idea how to check it else.


laters
zaxxon

 

[zaxxon]

*bump*

laters
zaxxon