Messages are not written to the Veritas Cluster Server (VCS) engine_A.log

Article: 100021724
Last Published: 2022-02-01
Ratings: 0 1
Product(s): InfoScale & Storage Foundation

Problem

Messages are not written to the Veritas Cluster Server (VCS) engine_A.log. No error messages in /var/log/messages indicating that logging for VCS has stopped.

 

Solution

To verify the issue is occurring, the following commands are executed:
 
1) Verify the process ID (PID) of HAD (High Availability Daemon). For example:
 
#ps -ef | grep had
root    21248 21219  0 13:19 pts/2    00:00:00 grephad
root    30130     1  0 Jan20 ?        00:00:03/opt/VRTSvcs/bin/had
root    30138     1  0 Jan20 ?        00:00:00/opt/VRTSvcs/bin/hashadow
 
The PID from the above example is 30130.
 
2) Verify the current state of the link to the engine_A.log.
 
Example:
#ls -l /proc/30130/fd |grep engine_A.log
l-wx------1 root root 64 Jan 22 13:21 13 -> /var/VRTSvcs/log/engine_A.log~(deleted)
 
In the above example, it is shown that the engine_A.log file has been deleted. The file is no longer linked to a valid file or directory due to the deletion of the original file in the file system. The VCS HAD engine continues to write to the unreferenced inode that previously belonged to the engine_A.log file. Whenthe"had" process exits, this unclaimed space will be returned to the file system free space, and the inode will become available.
 
Once "had"is restarted the engine_A.log file will be recreated and logging will occur. To restart the "had" process and recreate the engine_A.log execute the following commands:
 
#hastop -local -force
#hastart
 

 

Was this content helpful?