The Application resource shows "MONITOR TIMEDOUT" on cluster nodes where the resource is not running.
Article ID: 100004095
Updated On:
Description
Error Message
From "hares -display " output:
# hares -display | egrep -i 'flags|state' Flags node0 Flags node1 |MONITOR TIMEDOUT| Flags node2 |MONITOR TIMEDOUT| State node0 ONLINE State node1 OFFLINE|MONITOR TIMEDOUT State node2 OFFLINE|MONITOR TIMEDOUT Cause
This invalid state may result from an incorrect 'types.cf' file.
When updating the SFHA or VCS product on cluster nodes the latest 'types.cf' is not automatically copied into "/etc/VRTSvcs/conf/config".
This is by design as the 'types.cf' file may be modified by the system owner(s).
Overwriting this file with installation / upgrade actions could result in important changes to the file being lost or reset to default.
Prior to overwriting the file with the newer reference copy, it is recommended to verify if any custom modifications were made to the file.
Assuming there are no modifications, copying the latest 'types.cf' file should be undertaken to resolve the issue.
If modifications to the file were made, note of the changes should be made so that they may be implemented into the newer file.
Resolution
Verify the file in use is an older version.
For example:
# ls -l /etc/VRTSvcs/conf/types.cf-r--r--r-- 1 root system 12384 Feb 24 2010 types.cf# ls -l /etc/VRTSvcs/conf/config/types.cf-rw------- 2 root system 12223 Aug 4 12:53 types.cf
The file in the "/etc/VRTSvcs/conf" directory is larger, and was placed during the last VCS patch update.
Therefore, it is recommended to update the file.
To update the 'types.cf' file on each node:
a) Stop HA.
# hastop -all -force
b) Stop VCS infrastructure on each node:
# vxfenconfig -U# gabconfig -U# lltconfig -U
c) Copy the types.cf file on each node:
# cp /etc/VRTSvcs/conf/types/cf /etc/VRTSvcs/conf/config/types.cf
d) Start VCS infrastructure on each node:
# lltconfig -c# gabconfig -c -n# <---- where # is the number of nodes in the cluster# vxfenconfig -c
e) Start HA on each node:
# hastart
The "MONITOR TIMEDOUT state should no longer be shown on the node(s) where the Application resource is not running.
Applies To
Solaris 10
SF 5.1RP1
VCS 5.1RP1
Issue/Introduction
