From: Maglinger, Paul (PMAGLINGER@scvl.com)
Date: Mon Nov 08 2004 - 18:52:32 EST
My thanks to those who replied, but we ended up finding the answer on our
own. Really strange set of circumstances and am curious if others had
experienced anything similiar.
While waiting for Oracle tech support to call back, by accident we
discovered that although root could perform a clu_get_info without a
problem, logging in as any other user (including the Oracle account) would
give us this error on both nodes:
Hostname = CONFIGURATION_ERROR
Cluster interconnect IP name = CONFIGURATION_ERROR
Doing a search on Google found a document that suggested that the rc.config
file may be corrupted. When reviewing the file, it was discovered that the
permissions on the rc.config were 600 rather than 644. The only thing that
we can surmise is that when we used the X-Windows GUI and ran the network
configuration, that it somehow changed the permissions on the rc.config file
in the /cluster/members/{memb}/etc directory. We changed the file back to
the original permissions and now everything is working correctly. We ran an
allverify and that was the only file that had been changed. Has anyone seen
anything like this?
Original posting:
ES40 2 node cluster running Tru64 5.1B PK4. We are implementing a vlan in
the corp and needed to change the default gateway. When we changed the
default gateway setting, now we cannot get the gsd to start. When we try to
start gsd, we get the following message:
Problem initializing with clusterware, cannot start daemon.
Even when we changed the default gateway back, we cannot get the gsd to
start. Alert log messages that we are getting from Oracle are as follows:
< 8753:ORA-00447: fatal error in background process
< 8754:ORA-29701: unable to connect to Cluster Manager
< 8757:ORA-00444: background process "EMN0" failed while starting
< 8758:ORA-29701: unable to connect to Cluster Manager
< 8764:ORA-00447: fatal error in background process
< 8765:ORA-29701: unable to connect to Cluster Manager
< 8768:ORA-00444: background process "EMN0" failed while starting
< 8769:ORA-29701: unable to connect to Cluster Manager
Right now the database cannot be started. Any thoughts would be greatly
appreciated.
Paul Maglinger
Systems Administrator Shoe Carnival Inc.
(812)867-4674 pmaglinger@scvl.com
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:50:11 EDT