[bagtache]

Hi all,

Can anyone help me to solve the following problem:

I have an AIX 5.3 box with all the rsct binary services installed but only four rsct services (from group rsct_rm) are active. Can anyone tell me how to define the other rsct services and put them in rsct_rm group ?

Many thanks and best regards.

[duke900ssd]

Most RSCT services are lazy and will only spring into life whan required.

What environment is the system in: LPAR, HMC, IVM, VIOS, power4, power5, power6, etc?

[bagtache]

Hi Duck,

The environment is P595/LPAR (AIX 5.3)/HMC V6.1.2. Actualy the problem is that the DLPAR cmds doesn't run and i'm getting a message saying that the HMC can't reach the RMC service despite the fact the all IP @ are reachable, the HMC can ping the LPAR (and vice versa), the binaries of the rsct services are installed...

There are only two things that seem abnormal:

-Some RSCT services which are necessary to a normal DLPAR functionning don't belong to the rsct_rm group

-The HMC can ping the FQDN of the lpar but not the short name and i can't find a way to update the /etc/hosts file of the HMC.


Any Ideas ?

[ross.mather]

Here is the debugging guide for the Dynamic LPARing:

Dynamic LPAR tips and checklists for RMC authentication and authorization

[bagtache]

Hi Ross,

Thank's for your answer, but i have already visited this link and it doesn't contain any new information for me and, unfortunatly, this informations of this site couldn't help me to understand why one of my HMCs is saying that it can't reach the RMC service !!!!


Many thanks and best regards.

[ross.mather]

If one HMC works and the other not. then usually the flag for Partition Communication isn't active on the relevant interface on the HMC.

[bagtache]

Hi again;

In fact the problem raises only with some LPARs : on the same frame (P595) DLPARs cmds work with some LPARs and don't work with some other LPARs and i can't find the common configuration part which explains that. The only thing i have noticed is that on some LPARs (were DLPAR cmds Work) i have the following services active:
IBM.ERRM rsct_rm 278714 active
IBM.HostRM rsct_rm 303286 active
IBM.DRM rsct_rm 311494 active
IBM.CSMAgentRM rsct_rm 291044 active
IBM.ServiceRM rsct_rm 336050 active
IBM.AuditRM rsct_rm 307402 active
IBM.LPRM rsct_rm 327892 active
ec2servb:/>lssrc -a | grep IBM
ec2servb:/>lssrc -g rsct
Subsystem Group PID Status
ctrmc rsct 344258 active
ctcas rsct 495680 active

and where the DLPAR fails, i just have the following:

pz6serv:/>lssrc -a | grep IBM
IBM.ServiceRM rsct_rm 458848 active
IBM.ERRM rsct_rm 577704 active
IBM.AuditRM rsct_rm 483362 active
IBM.HostRM rsct_rm 835632 active
pz6serv:/>lssrc -g rsct
Subsystem Group PID Status
ctrmc rsct 430224 active
ctcas rsct inoperative


Beside that, the HMC is able to ping all the lpars by IP @ and all the LPARs can ping the HMC.

Any ideas ????

Best regards.

[duke900ssd]

Did you clone the problem LPARS, or change their IP address or hostname after the install?

If so, did you regenerate the RSCT keys after the install / change?

/usr/sbin/rsct/bin/rmcctrl -z <- Stop rsct.
/usr/sbin/rsct/install/bin/recfgct <- regenerate the host specific config key, this also cycles the daemon.
/usr/sbin/rsct/bin/rmcctrl -p <- Allow rsct remote connection.

Then wait about quater of an hour for everything to sync up again and retry the operation.

This process also works for the HMC but I think you'll probably need root / hscpe access to run the commands on the HMC.

If there are duplicate keys in an RSCT environment all sorts of strange things happen. A common symptom is /var filling up on the HMC which then causes even more strangeness...

[bagtache]

Hi Duck,

Again, i alredy did the RSCT rebuild because i was aware about the fact that some LPARs were incorectely cloned, unfortunatly this didn't fix the problem !!!

something is disturbing me: when i execute recfgct to rebuild the configuration, any service, beloging to the group rsct_rm, other than the following:
IBM.ServiceRM
IBM.ERRM
IBM.AuditRM
IBM.HostRM

is removed !!!!

I know that some involved services (like DRM) are lazy services and are started only when they are needed, but they need first to be known to the configuration and this is the problem: if i use mkssys to add them (i mean DRM and alike) manually, they are removed as soon as reccfgt is launched !!!
My conclusion, is that something was missed and the config database of RSCT doesn't contain any reference to the services we need. So we have first to correct this, before investigating network problem. I mean, that when executing: lssrc -a | grep IBM , all the services must apear in the list even though some are not started knowing that they will be started later where necessary.

Any ideas ?

Best regards.

[ross.mather]

Can you check the following sets of lpps?

lslpp -Lc | grep csm
lslpp -Lc | grep rsct


There is something buzzing round my head that that has caused this problem for me before.