ORACLE 10g rac node auto restart troubleshooting

After upgrading the database cluster to 10.2.0.5, the dual-node server continuously restarts and queries the oracle oprocd process logs. The information is as follows:

Jul 03 08:16:34.702 | INF | monitoring started with timeout(1000), margin(500), skewTimeout(125)Jul 03 08:16:34.704 | INF | fatal mode startup, setting process to fatal mode

You can see that the oprocd process interval is set

timeout(1000), margin(500), skewTimeout(125)

No obvious error information is found when viewing other log information. Therefore, it is suspected that the time parameter of oprocd is improperly set.

Modify the parameters as follows:

#crsctl set css diagwait 13 -force

Note the following when modifying parameters:

When we set the diagwait parameter, we need to stop the cluster activity of all rac nodes. The stop sequence is as follows (Modifying diagwait in a cluster activity may damage ocr.):

1. Run the command as root on all nodes

#crsctl stop crs #<CRS_HOME>/bin/oprocd stop

2. Check that the process on all nodes has stopped.

#ps -ef |egrep "crsd.bin|ocssd.bin|evmd.bin|oprocd"

3. Modify the parameters. You can modify the parameters on only one node.

#crsctl set css diagwait 13 -force

4. view the parameter modification information.

#crsctl get css diagwait

5. Start the cluster node

#crsctl start crs#crsctl check crs

After modifying the parameters, the cluster service runs normally and has not restarted automatically.