After a hard reboot, there is a lock file left behind in
/opt/IMDataAggregator/apache-karaf-*
[root@lab apache-karaf-4.2.6]# pwd
/opt/IMDataAggregator/apache-karaf-4.2.6
[root@lab apache-karaf-4.2.6]# ls -al
total 132
-rw-r--r-- 1 root root 0 Feb 10 08:15 lock
-rw-r--r-- 1 root root 29 Feb 10 08:15 .lock
drwxrwxr-x 2 root root 6 Feb 10 08:15 MIBDepot
-rw-r--r-- 1 root root 1825 Jun 7 2019 NOTICE
-rw-r--r-- 1 root root 5838 Jun 7 2019 README.md
-rw-r--r-- 1 root root 57118 Jun 7 2019 RELEASE-NOTES.md
(truncated list for readability)
………….
And this file is causing the DA process to fail.
If we remove this file manually, the DA starts.
Release : 21.2.6+
Component : PM Data Aggregator
NOTE - This issue can be seen on the data aggregator (DA) as well as the data collectors (DC), please apply the fix to both.
NOTE - This issue can be seen on the data aggregator (DA) as well as the data collectors (DC), please apply the fix to both.
In 21.2.6 we upgraded to karaf 4.
This version allows for karaf HA containers and that is dependent on that lock file being present in $KARAF_HOME.
But, DX NetOps Performance Management does not use the HA karaf options, so this dependency can be disabled.
We do not use the lock file at all, it is part of HA karaf containers, which we do not actually use.
Let’s disable it:
Stop dadaemon
Delete $KARAF_HOME/lock file
cd /opt/IMDataAggregator/apache-karaf
rm -f lock (not .lock)
Delete $KARAF_HOME/data and $KARAF_HOME/deploy/local-jms-broker.xml
cd /opt/IMDataAggregator/apache-karaf
rm -rf data
rm -f deploy/local-jms-broker.xml
Copy $KARAF_HOME/etc/system.properties
cd /opt/IMDataAggregator/apache-karaf/etc
cp system.properties system.properties.original
Edit $KARAF_HOME/etc/system.properties file, and add a line with karaf.lock=false
At the bottom.
E.G.
# Added on instructions in Case #xxxx by Broadcom
karaf.lock=false
Start dadaemon
Note: these steps must be redone after an upgrade until the permanent fix is added.
Broadcom is reviewing a permanent fix, no eta at this point.