We help IT Professionals succeed at work.

error while trying to create a spot from a mksysb after upgrade the NIM to a new TL

sminfo
sminfo asked
on
hi

I have always done SPOTs from MKsysb resources on NIM, but after upgrade the NIM to AIX 7.1TL1-SP1 I get this error:

[isga01s@nim:/var/tmp] sudo nim -o define -t spot -a server=master -a location=/nim/spot1 -a source=MKSYSB_testaix spottestaix

 Creating SPOT in "/nim/spot1" on machine "master" from "MKSYSB_testaix" ...

 Restoring files from BOS image.  This may take several minutes ...

0042-001 nim: processing error encountered on "master":
   0042-103 m_mkspot: unable to restore from MKSYSB_testaix

I have enough space on the filesystem on the nIM, I have tested to create the SPOT with different versions of mksysb (aix5.3, 6,.1 and 7.1) and still get this error.

Don't know where to check.
Comment
Watch Question

Most Valuable Expert 2013
Top Expert 2013

Commented:
Hi,

this can happen when the NIM master doesn't find a mandatory file in the mksysb source when creating the SPOT.

I read about such problems in conjunction with HFI support, but I don't think you're using such interfaces, or are you? Check with

lsnim -a "if_defined" master.

You should check the NIM log, maybe there's a hint:

alog -t nim -o


I'll search IBM support, maybe there's a known issue...

wmp




Author

Commented:
hi wmp...

[isga01s@bibmnim:/home/isga01s] sudo lsnim -a "if_defined" master
master:
   if_defined = chrp.64.ent
   if_defined = chrp.mp.ent

alog doesn't show me anything interesting..

umm.. I saw that when it aborts nim kept the spot folder (/nim/spot1/spottestaix)  with s lot of files inside... so, after 4 or 5 re-intents nim is finished the creation of the  spot, but not sure at all if the spot is correct or not..


Most Valuable Expert 2013
Top Expert 2013

Commented:
Check it -

nim -o check spottestaix

nim -o lppchk -a lppchk_flags="-c" spottestaix

nim -o lppchk -a lppchk_flags="-v" spottestaix

nim -o showres spottestaix

showlog -N spottestaix -t niminst

The latter will show the "install" protocol of the SPOT as opposed to the nim alog which shows kind of "creation" messages.

Author

Commented:
sudo nim -o check spottestaix --> does not return data

 sudo nim -o lppchk -a lppchk_flags="-c" spottestaix
/usr/bin/lppchk: No such file or directory
warning: warning: 0042-001 m_lppchk: processing error encountered on "master":
   0042-175 c_ckspot: An unexpected result was returned by the "/usr/bin/lppchk" command:

See the log file:
        /nim/spot1/spottestaix/usr/lpp/bos/inst_root/var/adm/ras/nim.lppchk
for details or use the "showlog" operation.

sudo nim -o lppchk -a lppchk_flags="-v" spottestaix
/usr/bin/lppchk: No such file or directory
warning: warning: 0042-001 m_lppchk: processing error encountered on "master":
   0042-175 c_ckspot: An unexpected result was returned by the "/usr/bin/lppchk" command:

See the log file:
        /nim/spot1/spottestaix/usr/lpp/bos/inst_root/var/adm/ras/nim.lppchk
for details or use the "showlog" operation.

 sudo nim -o showres spottestaix
0042-001 nim: processing error encountered on "master":
   0042-207 m_showres: Unable to allocate the spottestaix resource to master.

 sudo showlog -N spottestaix -t niminst --> sudo: showlog: command not found (is it nim -o showlog?)

I think the spot is not OK..

wmp, can you explain me the meaning of this command: lsnim -a "if_defined" master

thanks!!
Most Valuable Expert 2013
Top Expert 2013
Commented:
>> ... does not return data ..<<

That's good! No output, no error!

>> /usr/bin/lppchk: No such file or directory <<

That's really strange! I never saw an AIX system without "lppchk"!

What do you get with "lppchk -v" alone?

Could this be a sudo problem?

>>  Unable to allocate ... <<

Is the the SPOT directory NFS exported? If so that's maybe a remnant of some previour broken operation. Unexport it, and retry!

Of course it's "nim -o showlog ..." I'm really getting old ...

>> lsnim -a "if_defined" master <<

shows all network install interfaces known to the NIM master. These are the interfaces every SPOT needs drivers for, regardless of the actual client interface used.

I suggested issuing this command to check whether there is an HFI interface present. These interfaces need particular drivers, and there must be a bug in AIX 7.1 causing these drivers not always to be present.

HFI = Host Fabric Interface

Author

Commented:
ummm wmp, I think it;s because sudo.. I;m making some test now creating spots as root..

 nim -o showlog -n spottestaix -t niminst does not work:
nim: Not a recognized flag: n

usage:
        nim -o <operation> [-F] [-t <type>] [ -S <YYMMDDhhmm> ]

strange...
[root@bibmnim:/] ls -l /usr/bin/lppchk
-r-xr-xr-x    1 bin      bin          318502 Sep 28 18:55 /usr/bin/lppchk

lppchk -v does not return nothing

Most Valuable Expert 2013
Top Expert 2013
Commented:
Ok,

use just

nim -o showlog spottestaix


When running lppchk -v outside of NIM it checks the lpp packages on the local machine, and since your installation is OK there is no output. No output, no error.


I found the reason for the "nim -o lppchk ..." problem, however:

NIM mounts the SPOT and tries to use the lppchk binary found inside this SPOT.

So the "not found" error means that the lppchk binary is not present in the SPOT, and that's of course a clear indication that the SPOT is NOT OK!


Have I been right with my guess that the SPOT directory was NFS exported?


wmp

Author

Commented:
wmp, what cause the creation of the SPOT is SUDO. But I don't understand why, because my user has:

israel ALL=(ALL) NOPASSWD: ALL

So all commands are run as root. When I create \the SPOT from root I see no error, on the contrary as a normal user with sudo it aborts. Really strange.

And yes, the mksysb is on a NFS  filesystem ( where NIM has all mksysb of LPARs).

I also had the similar issue. But no problem with SUDO.
 After creating SPOT. when i try verifying with "nim -o lppchk -a lppchk_flags="-c" spot" i got the below error:
0042-175 c_ckspot: An unexpected result was returned by the "/usr/bin/lppchk" command:
Even i could do bos_inst. But at the image restoration time it was struck after the pocket count complete.