Hi
I’ve recently installed HDP2.3 on a new cluster and have hit an issue that brought the whole hbase down.
I’ve been using opentsdb for the past 2 months on HDP2.2 (hbase 0.98) and everything seems fine.
However yesterday after 4hours of usage all regionservers died and would not come up. Looking at the logs I’ve found a particular exception that led me to find this hbase bug because the logs were exactly the same: https://issues.apache.org/jira/browse/HBASE-13329
This bug has been fixed in hbase 1.1.2, but HDP2.3 only has 1.1.0.1. The trouble was that hbase would not come up and always giving the exception ArrayIndexOutOfBoundsException.
Because the cluster was a production cluster and I needed to have it up asap I went with the very dangerous route of downloading the hbase hdp2.3 source, applying the patch (changing the short to an int..), compiling hbase-commons.jar and uploading it to all the servers.
Thankfully that worked and everything seems to be working ok again (I still had my old cluster sitting idle in case this didn’t work).
Is there an estimate when hdp2.3 can be updated to 1.1.2?
Thanks