Symptoms:
In a HA environment a false fail-over can be triggered by a client starting a vote to move to a new server.Cause:
If a MDS doesn't reply in a specific given time a vote might initiated. As the default HA_SMITH_INTERVAL is 5 seconds if not otherwise specified in certain environments fail-overs without reason can be seen.Fix:
- Create the file /usr/cvfs/config/ha_smith_interval and create the following line
- ha_smith_interval=<integer> (i.e. 15 - 30 is an acceptable value)
The file is read once when StorNext starts. The integer value for the HAmon timer interval is expressed in seconds. The value can range from 3 to 1000, and the default is 5 seconds. The timer must be set identically on both servers.
Notes:
The file is read once when StorNext starts. The integer value for the HAmon timer interval is expressed in seconds. The value can range from 3 to 1000, and the default is 5 seconds. The timer must be set identically on both servers.Affected:
StorNext versions higher 4.0
To do:
Changing (creating) the files on both servers and restart the cvfs service.
0 Comments