Error: Unable to connect to configuration database.

[Paul Condren]

Hi.

I have a problem with a 2 node cluster, which has been happily running for month, until today!

Ive rebooted one server yesterday - priimary node and no problems, however rebooted the standby node this morning and now thre cluster wont form! The following error is displayed when a command such as cmruncl, cmquerycl etc is used.

###############################################
Internal error: Unable to open communications to configuration daemon: Not owner
Error: Unable to connect to configuration database.
Internal error: Unable to open communications to configuration daemon: No such file or directory
cmrunnode : Unable to determine nodes on the current cluster.
Either no cluster configuration file exists, the file is corrupted,
cmclconfd is unable to run, or user root on node ira70871
doesn't have access to view the configuration.
###############################################

all the hosts, .rhosts etc have both noded and the networks for each are fine - they can ping and connect to each other on all configured lans.

Any help apriciated. Things tried so far are more reboots and a reinstyall of serviceguard 11.16 but no joy.

I can delteconf oreapply the conf either :o(

[RAC_1]

Are uou doing it as root? Also how does cmclconfig binary (/etc/cmcluster/cmclconfig) look like on both nodes? what error messages do you get when you force cluster start? cmruncl -f

[Steven E. Protter]

Shalom,

Sounds like the binary cluster configuration file is corrupt on one or both nodes.

It resides in /etc/cmcluster on most systems.

You probably need to erase it and run cmapplyconf again.

SEP

[Paul Condren]

Hi.

Ive removel all entries from the hosts file apart from the vial ones for the server thamselves, and nslookup fo both the hosts names and the IP addresses all works fine. I cant see anything in the syslog thats not being put on screen.

Ive tried a cmquery to build a fresh cluster file and that fails too. as doese a delete, check or apply. Its all wrong!

The stangest error that we are guessing as the root cause is this one which is recieved with a check or apply conf:

############################################
ira70871 *CLONE* >cmcheckconf -v -C safe_mgt.ascii

Checking cluster file: safe_mgt.ascii
Checking nodes ... Done
Checking existing configuration ... Done
Internal error: Unable to open communications to configuration daemon: Not owner
Error: Unable to connect to configuration database.
Internal error: Unable to open communications to configuration daemon: No such file or directory
Gathering configuration information ... Done
Gathering configuration information ..
Gathering storage information ..
.....
Gathering Network Configuration ..... Done

Error: Unable to connect to node ira70871: No such file or directory
Error: Unable to connect to node ira70872: Not owner
Error: Volume Group /dev/vg03 is not found on any nodes in the cluster.
Error: Volume Group /dev/vg04 is not found on any nodes in the cluster.
Error: Volume Group /dev/vg06 is not found on any nodes in the cluster.
Error: Volume Group /dev/vg07 is not found on any nodes in the cluster.
Error: Volume Group /dev/vg10 is not found on any nodes in the cluster.
Error: Volume Group /dev/vg12 is not found on any nodes in the cluster.
Warning: Not probing node ira70871 as it is currently unreachable.
This may cause network partitions to be reported.
Warning: Not probing node ira70872 as it is currently unreachable.
This may cause network partitions to be reported.
cmcheckconf : Unable to reconcile configuration file safe_mgt.ascii
with discovered configuration information.
ira70871 *CLONE* >
############################################

the message about the volume group goes away if I activate all the shared vg's or use a -K option so thats not the cause. it the line "Error: Unable to connect to node ira70872: Not owner" that is wierd to me.

[rariasn]

Verify "cmclconfig" permission

#ll /etc/cmcluster/cmclconfig
-rw------- 1 root root 22044 Mar 13 12:18 /etc/cmcluster/cmclconfig
#ll -d /etc/cmcluster
dr-xr-xr-x 10 bin bin 8192 Aug 2 09:17 /etc/cmcluster

rgs,

[Denver Osborn]

enable logging and look in the syslog.log file.

vi /etc/inetd.conf (make a copy 1st), then append "-T5" to your hacl entries. Once saved, inetd -c and run some mcsg commands. Look in the systelog.log for debug output.

What changed? Any patches, etc...

-denver

[Paul Condren]

Thanks for the help. THe cause of this has been traced to a line in the inetd.conf file bing commented out! its the idend line.

Nor sure how/when this happended. Easy fix tho.

thanks.