[PaulClayton]

Got two VIO servers connected to a P570 P6 environment.
When ever I try to configure dynamic resources, I get the above message.
VIO = FP11
HMC= 7.3.3 (plus mandatory patches.

Don't have the problem with a client LPAR.

I have run all th checks, network, etc, but can't get through the RMC and am now at a loss.

If I run "lspartition -dlpar",it shows that the VIO's are not active.

There is a document that IBM sent me, but it does not help.

Any other info I can supply I will.

[cdelgadop]

Hi

Well if IBM has already sent a doc it might happen some of the tips i 'm about to give are already done:

Remember communications between LPARs and HMC in order to allow for DLPAR operations are done using two networks, a public one where all your LPARs and HMC controlling them are connected (so you can ping either by name or ip address from each LPAR to the HMC and from HMC to each LPAR) and a private one where HMC is connected to each pSeries Frame. You need to be sure you have that infrastructure.

Then you need to be sure that the following subsystems are running:

root@ / >lssrc -a | grep rsct
ctrmc rsct 241784 active
IBM.CSMAgentRM rsct_rm 323742 active
IBM.HostRM rsct_rm 311462 active
IBM.DRM rsct_rm 336036 active
IBM.AuditRM rsct_rm 340134 active
IBM.ERRM rsct_rm 221348 active
IBM.ServiceRM rsct_rm 274586 active
IBM.LPRM rsct_rm 315626 active
ctcas rsct inoperative
root@bmintcan04 / >

If not please try starting them up.

Once you have checked below requirement and have checked all stuff on that IBM document you can use the command

/usr/sbin/rsct/bin/rmcctrl

to completely stop and delete all objects related to the
DLPAR tasks and start again.

Hope this helps

[PaulClayton]

Yep all running, and tried that as well. the IBM RM troubleshooting guide is a good one, but I somehow think, that I have an autentication issue between the VIO and HMC.

[bbeuchaw]

Hi Paul, I just had this same issue (same versions, h/w, etc.) and was wondering if you ever found a resolution to this? Thanks much!

Also, if anybody else had the same issue and found a resolution, I'd appreciate any info before opening a PMR w/IBM...

Thanks again...

[PaulClayton]

I have had this call open for months now and IBM have not been able to resolve it.

[bbeuchaw]

Well, I went ahead and opened a PMR, and this is what they came back with - the last 2 cmds are what worked for me (we had done an alt_disk_install of our VIO servers, so they all had the same cluster ID that the rmc stuff was looking at, and this reset them all back to unique ones - at least that's how I understand it). Hope this helps!

login to the vioserver

$ oem_env_setup ....... become root

# lslpp -l csm.client
this should return 1.7x.

ping the HMC IP address , not the Host name

# ping <HMC IP >

if this works ok. (unless ping is disabled)

then run these two commands to shutdown and restart the RMC
connections

# /usr/sbin/rsct/install/bin/recfgct
# /usr/sbin/rsct/bin/rmcctrl -p

Wait about 5 minutes and then try the DLPAR action again.
Let me know if that would resolve the issue.

[PaulClayton]

The problem has finally been resolved. The issue was with the MTU of the virtual IO server being set at 9000 and the HMC being 1500.
Interestingly this appears to have occured after the FixPack 10 update.
The RSCT daemons required resetting and restarting prior to this working properly.