TruCluster
cancel
Showing results for 
Search instead for 
Did you mean: 

TruCluster - "shutdown -c" command

TruCluster - "shutdown -c" command

On running the "shutdown -c" command the following output is observed -

systemA#shutdown -c now

systemb-ics0: Connection timed out

shutdown: Bad rcmd to systemb-ics0

The .rhosts files, /etc/hosts files etc have been edited with entries for systemA and systema
shutdown: member systemB-ics0 is UP but does not respond to rsh commands

systema-ics0: Connection timed out

shutdown: Bad rcmd to systema-ics0

shutdown: member systemA-ics0 is UP but does not respond to rsh commands

systemA#

Does Tru64/TruCluster get confused with naming of systemA rather than systema?
8 REPLIES
Michael Schulte zur Sur
Honored Contributor

Re: TruCluster - "shutdown -c" command

Hi,

have a look into the binary errorlog.
What os version and patch kit are you running?

greetings,

Michael

Re: TruCluster - "shutdown -c" command

Michael,

Thanks for reply
O/S is Tru64 V5.1B
Patch Kit is 4

Regards,

Turlough
Johan Brusche
Honored Contributor

Re: TruCluster - "shutdown -c" command


Turlough,

You should have systema-ics0 and systemb-icso in /.rhosts or /etc/hosts.equiv, this to allow shutdown to issue remote commands to the other cluster member(s) via the cluster interconnect (mem-channel or lan)

If you do not have these settings, also "clu_check_config -v " will fail on several items.

Rgds,

__ Johan.

_JB_

Re: TruCluster - "shutdown -c" command

Michael,

Have entries for systemA and systema in /.rhosts and /etc/hosts files.

Turlough
Michael Schulte zur Sur
Honored Contributor

Re: TruCluster - "shutdown -c" command

Tourlough,

Johan was referring to the cluster inter connect which has its own ip address.
You could try a
rlogin systemB-ics0 from systemA
When you get a permission denied then
systemA-ics0 is missing in /etc/hosts.equiv on systemB.

greetings,

Michael

Re: TruCluster - "shutdown -c" command

Thanks Michael, Sorry Johan last reply should have been to you!

Have verified .rhosts has correct entries(same file for both nodes in the cluster).
ping works fine for systemA and systemA-ics0
but rsh only works for systemA NOT systemA-ics0.

On reboot(last night) discovered that kloadsrv is not starting whicj I think is also stopping gated running also, the kloadsrv error is

"kloadsrv: Error 0: kls_restore_state read state size failed, return 1"

Anyone got access to source code to determine what this is referring to? Have not been able to find any ref on web yet.

Turlough
Venkatesh BL
Honored Contributor

Re: TruCluster - "shutdown -c" command

In the error message,
kloadsrv: Error 0: kls_restore_state read state size failed, return 1

"read state size" is the error message of the failure.

So, the error was caused as kernel state size could not be read.

Re: TruCluster - "shutdown -c" command

Any ideas on why kernel state size error should occur?