HA NIM Takeover Problem

[dial8d]

I have setup a NIM master, and NIM alternate master, AIX 5.3 TL5, and some clients using NIMSH communication.

NIM master works fine, talking to the clients ok. The problem i am having is when i try to activate the NIM alternate master. The command fails, and from looking at the /var/adm/ras/nimsh.log logfile, it seems to be due to the different cpu id of the alternate master.

ed Apr 25 13:13:19 2007 success: we got 1st write query is 0
Wed Apr 25 13:13:19 2007 success: we got 2nd write local id is 00C7EB7B4C00
Wed Apr 25 13:13:19 2007 success: we got 3rd write remote id is 0001ED1ED700
Wed Apr 25 13:13:19 2007 success: we got 4th write command is /usr/lpp/bos.sysmgt/nim/methods/c_nimpush "/usr/lpp/bos.
sysmgt/nim/methods/c_switch_master" "-anew_master=nis10prdbckcl02" "-aoriginator=nis10prdbckcl02" "-aasync=yes"
Wed Apr 25 13:13:19 2007 passing OpenSSL setting of 0
Wed Apr 25 13:13:19 2007 error: remote value passed, 0001ED1ED700, does not match environment value 0001ECF4D700

If i then re-establish the primary NIM master, it works fine, giving the output below ;

ed Apr 25 13:10:21 2007 success: we got 1st write query is 0
Wed Apr 25 13:10:21 2007 success: we got 2nd write local id is 00C7EB7B4C00
Wed Apr 25 13:10:21 2007 success: we got 3rd write remote id is 0001ECF4D700
Wed Apr 25 13:10:21 2007 success: we got 4th write command is /usr/lpp/bos.sysmgt/nim/methods/c_nimpush "/usr/lpp/bos.
sysmgt/nim/methods/c_switch_master" "-anew_master=nis10prdbckcl01" "-aoriginator=nis10prdbckcl01" "-aasync=yes"
Wed Apr 25 13:10:21 2007 passing OpenSSL setting of 0
Wed Apr 25 13:10:21 2007 sending ack to client
Wed Apr 25 13:10:21 2007 setting descriptors to include 2nd port
Wed Apr 25 13:10:21 2007 command to exec __ /usr/lpp/bos.sysmgt/nim/methods/c_nimpush "/usr/lpp/bos.sysmgt/nim/metho
ds/c_switch_master" "-anew_master=nis10prdbckcl01" "-aoriginator=nis10prdbckcl01" "-aasync=yes"


Ive tried everything i can think of so far. Unfortunately, i cant set the cpu id to be the same on the alternate master as the primary master.

Any idea's ?

 

[dial8d]

Ok, got the solution.

NIM HA takeover relies on RSH comms between the alternate NIM master and NIM clients. It doesnt work with NIMSH, which sucks.