Systems Director DNZPAX075E NIM update error, what to do?
nagger 100000MRSJ Visits (6724)
I suddenly could not get AIX updates to work via my NIM server. The problem number DNZPAX075E just means the NIM command failed for unknown reasons. Not very helpful when you have a ISD Server, NIM Server and a Target AIX Server and the problem could be anywhere. I looked around and found a few other people were stuck on this one. So now I worked it out - with the help excellent help of systems Director support.
1) Restart the Systems Director Server and the NIM Server - just good house keeping and check you have free disk space in the filesystems. Then switch on the tracing for this part of Systems Director. I was told to do this
cd /optIt worked fine. You can see this is now in effect on the GUI or you could have added it there:
2) On the Systems Director Server flush the logs to make reading them easi
/o3) Check the current Target Server AIX level with: oslevel -s and make sure you are actually trying a level above what it has!! It does report there was nothing to do in the regular log output but does not finsh in a "Complete with error" which can be confusing.
4) Make sure the NIM server is higher than the level you are trying - not NOT all AIX 7 levels are higher than the AIX levels - check the oslevel -s output fourth numb
710The 1228 means 2012 and week 28. It is this reason that I don't like having NIM on the systems Director server itself - i.e. having to upgrade the AIX that ISD is running on for NIM reasons.
5) Next try the operation to NIM install a small service pack update to create a fresh log entry.
6) Go look in the log directory and identify the last log file- this is a LWI area feature so the logs are here:
# ls -ltr /optObviously error-log-0.html is the current log file.
7) Now the tricky part and where I can only cover what my problem was. I hope you find something similar. View the log file:
I looked to the error number DNZPAX075E and then back tracked a bit (back up the file) and noticed it logged the NIM command it did not like as foll
COM8) I checked the NIM server was connected and Inventory fresh and the Security Configure Access looks like this or better (with ssh):
Probably should have been checked sooner. Also check you have nimsh running on the NIM server with:
startsrc -s nimsh
9) Cut and paste that command and run it on my NIM server directly:
/usr/sbin/niminit -a connect=nimsh -a name=bronze4 -a master=puple410) It fails - OK I have duplicated the problem. So I stare hard at that command and suddenly notice the letter "r" is missing.
The NIM master (the NIM server itself in a simple NIM set-up) is called purple4 not puple4. How could Systems Director get that wrong!!! I can't find puple4 on the Systems Director list of machines!!! So where is it picking up the duff hostname?
11) Ran the "hostname" command on the NIM server - bingo. It thinks it is called puple4.
I used a mksysb reinstall the NIM server to move from internal VIOS based SAS disks to V7000 LUNs to give myself a performance boost and less work (no multiple disks and AIX mirroring) and I must have introduced this typo in the hostname. Darn!
12) So next fix the hostname and "wow" Systems Director AIX upgrades start working as normal - phew!! Now I feel a bit guilty with all that cussing I did about systems Director going wrong! It turned out to be my fault. I guess Systems Director could report error a little better to allow for all the user errors it has to deal with!
13) Don't forget to switch off that tracing to the log file - it can slow down Systems Director in regular use - i.e. only switch it on when you need it.
I hope this is useful, have a good day, Nigel