Crucial error messages from vxconfigd are not logged to /var/adm/messages

Article:TECH37311  |  Created: 2005-01-09  |  Updated: 2005-01-09  |  Article URL http://www.symantec.com/docs/TECH37311
Article Type
Technical Solution

Product(s)

Environment

Issue



Crucial error messages from vxconfigd are not logged to /var/adm/messages

Solution



If vxconfigd can't join a Cluster Volume Manager cluster, by default some crucial error messages are only sent to the console device. For example, in the case vxconfigd cannot find a disk which is supposed to be in a shared disk group, vxconfigd will send the following messages to the console:


VxVM vxconfigd NOTICE CVM_VOLD_CONFIG command received
VxVM vxconfigd NOTICE CVM_VOLD_CHANGE command received
VxVM vxconfigd NOTICE establishing cluster
VxVM vxconfigd WARNING slave: missing disk 1104366468.5014.jacaranda
VxVM vxconfigd WARNING cannot find disk 1104366468.5014.jacaranda
VxVM vxconfigd ERROR -1 returned from volcvm_establish
VxVM vxconfigd ERROR cluster_establish: (222) Cannot find disk on slave node
VxVM vxconfigd ERROR kernel_fail_join() : master_takeover is 0
VxVM vxconfigd NOTICE CVM_VOLD_STOP command received

If vxconfigd was restarted in a Telnet session, the above messages will be sent to that Telnet session. If that particular Telnet session has been closed, there will be no trace of the above messages.

If console display is not accessible, then the above messages from vxconfigd may not be noticed.

To log the above messages to /var/adm/message, vxconfigd should be started with the option -x syslog.  To restart vxconfigd with the new options, run the following command:

# vxconfigd -x syslog -k -m enable

Caution: Do not run the above command on a system where VERITAS Cluster Server (tm) is monitoring the disk group or volume resources (Cluster Volume Manager shared or not) or the system has joined a Cluster Volume Manager cluster.  Running the above the command may cause Cluster Server to bring down the service groups.  
Bring down all service groups depending on the Cluster Volume Manager service group, then the Cluster Volume Manager service group,  and freeze the system before running the above command.


Caution: When vxconfigd is running with -x syslog option, the crucial error messages will be logged through syslogd to /var/adm/messages:

/var/adm/messages:
-----------------------------
Feb  9 13:32:40 jacaranda vxvm:vxconfigd: [ID 456610 daemon.notice] CVM_VOLD_CONFIG command received
Feb  9 13:32:40 jacaranda vxvm:vxconfigd: [ID 699813 daemon.notice] CVM_VOLD_CHANGE command received
Feb  9 13:32:40 jacaranda vxvm:vxconfigd: [ID 322665 daemon.notice] establishing cluster
Feb  9 13:32:46 jacaranda vxvm:vxconfigd: [ID 874742 daemon.warning] slave: missing disk 1104366468.5014.jacaranda
Feb  9 13:32:46 jacaranda vxvm:vxconfigd: [ID 857411 daemon.warning] cannot find disk 1104366468.5014.jacaranda
Feb  9 13:32:46 jacaranda vxvm:vxconfigd: [ID 778436 daemon.error] -1 returned from volcvm_establish
Feb  9 13:32:46 jacaranda vxvm:vxconfigd: [ID 172191 daemon.error] cluster_establish: (222) Cannot find disk on slave node
Feb  9 13:32:46 jacaranda vxvm:vxconfigd: [ID 700632 daemon.error] kernel_fail_join() : master_takeover is 0
Feb  9 13:32:46 jacaranda vxvm:vxconfigd: [ID 100000 daemon.error]

You can make this vxconfigd option permanent by modifying the startup script.  In /etc/rcS.d/S25vxvm-sysboot:

Original File Contents:
--------------------------------
# comment-out or uncomment any of the following lines to enable or
# disable the corresponding feature in vxconfigd.
#opts="$opts -x syslog"                 # use syslog for console messages
#opts="$opts -x log"                    # messages to /var/vxvm/vxconfigd.log
#opts="$opts -x logfile=/foo/bar"       # specify an alternate log file
#opts="$opts -x timestamp"              # timestamp console messages

After -x syslog is enabled:
--------------------------------------
# comment-out or uncomment any of the following lines to enable or
# disable the corresponding feature in vxconfigd.
opts="$opts -x syslog"                 # use syslog for console messages
#opts="$opts -x log"                    # messages to /var/vxvm/vxconfigd.log
#opts="$opts -x logfile=/foo/bar"       # specify an alternate log file
#opts="$opts -x timestamp"              # timestamp console messages

When the system reboots, this vxconfigd option will be used.

In VERITAS Volume Manager (tm) 4.0 , -x syslog is used by default.





Legacy ID



275120


Article URL http://www.symantec.com/docs/TECH37311


Terms of use for this information are found in Legal Notices