check_zookeeper_znode.pl (Advanced Nagios Plugins Collection)

Project Notes

check_zookeeper_znode.pl

Part of the Advanced Nagios Plugins Collection, download it here:

https://github.com/harisekhon/nagios-plugins

================================================================================
Some useful examples:
================================================================================

* Check we have an active HBase Root Master (this is an ephemeral node that will disappear if Master is down):

check_hbase_master_znode.pl -H
or
check_zookeeper_znode.pl -H -z /hbase/master --ephemeral
________________________________________________________________________________

* Check we have an HBase Root RegionServer assigned:

check_hbase_root_regionserver_znode.pl -H
or
check_zookeeper_znode.pl -H -z /hbase/root-region-server
________________________________________________________________________________

* Check there are no HBase unassigned regions (should be blank hence -d ""):

check_hbase_unassigned_regions_znode.pl -H
or
check_zookeeper_znode.pl -H -z /hbase/unassigned -d "" --no-child-znodes
________________________________________________________________________________

* Check there are HBase Backup Masters:

check_hbase_backup_masters_znode.pl -H
or
check_zookeeper_znode.pl -H -z /hbase/backup-masters
________________________________________________________________________________

* Check given SolrCloud server is alive and holding it's ephemeral znode:

check_solrcloud_server_znode.pl -H -z /solr/live_nodes/:8983_solr
or
check_zookeeper_znode.pl -H -z /solr/live_nodes/:8983_solr --null --ephemeral
________________________________________________________________________________

* Check HDFS NameNode HA ZKFC is working

check_hadoop_namenode_ha_zkfc_znode.pl -H <zookeepers>
or
check_zookeeper_znode.pl -H <zookeepers> -z /hadoop-ha/nameservice1/ActiveStandbyElectorLock --ephemeral
________________________________________________________________________________

* Check MapReduce v1 JobTracker HA ZKFC is working

check_hadoop_jobtracker_ha_zkfc_znode.pl -H <zookeepers>
or
check_zookeeper_znode.pl -H <zookeepers> -z /hadoop-ha/logicaljt/ActiveStandbyElectorLock --ephemeral

================================================================================

usage: check_zookeeper_znode.pl [ options ]

-H --host ZooKeeper node(s) to connect to, should be a comma separated list of ZooKeepers the same as are configured on the ZooKeeper servers themselves (node1:2181,node2:2181,node3:2181). It takes longer to connect to 3 ZooKeepers than just one of them (around 5 secs per ZooKeeper specified + (session-timeout x any offline ZooKeepers) so you will need to increase --timeout). Connection order is deterministic and will be tried in the order specified unless --random-conn-order
-P --port Port to connect to on ZooKeepers for any nodes not suffixed with :<port> (defaults to 2181)
-z --znode Znode to check exists. Useful for a variety of checks of ZooKeeper based services like HBase, SolrCloud, NameNode & JobTracker HA ZKFC
-d --data Check given znode contains specific data (optional). This is a partial substring match, for more control use --regex with anchors. Careful when specifying non-printing characters which may appear as ?, may need to use regex to work around them with ".+" to match any character
-r --regex Check given znode contains data matching this case insensitive regex (optional). Checked after --data
-n --null Do not check znode contents, use on null znodes such as SolrCloud /solr/live_nodes/<hostname>:8983_solr as the API segfaults when trying to retrieve data for these null znodes
-e --ephemeral Check given znode is ephemeral (optional)
--child-znodes Check given znode to has child znodes (optional)
--no-child-znodes Check given znode to not have child znodes (optional)
-u --user User to connect with (Not tested. YMMV. optional)
-p --password Password to connect with (Not tested. YMMV. optional)
-w --warning Warning threshold or ran:ge (inclusive) for znode age (optional)
-c --critical Critical threshold or ran:ge (inclusive) for znode age (optional)
--random-conn-order Randomize connection order
--session-timeout ZooKeeper session timeout in secs (default: 2). This determines how long to wait for connection to downed ZooKeepers and affects total execution time
-h --help Print description and usage options
-t --timeout Timeout in secs (default: 10)
-v --verbose Verbose mode (-v, -vv, -vvv ...)
-V --version Print version and exit