We have recently faced a Blocker at our end where KahaDB files did not get cleaned up automatically and queued up for couple of days.
Unfortunately we did not notice this and the shared disk size became 100% full.
- This was a master/slave architecture where both are using the same shared drive. Hence ActiveMQ was unresponsive at this situation.
- We have moved the kahadb location to a temporary storage and got the disk cleaned up and pointed it back. Now the journal files are getting cleaned up automatically.
<kahaDB directory="/data02/apphome/Broker" journalMaxFileLength="100mb" cleanupInterval="5000"/>