Applies to:
- Zenoss Resource Manager 5.1.1
Summary
Upgrading to Zenoss Resource Manager 5.1.1 can result in the TTL (time to live) HBASE data retention policy being set incorrectly. The TTL setting controls how long performance data is retained for targets within Resource Manager. Changing this value can be useful in various scenarios, for example:
- decrease the TTL value to save space and prevent low space issues
- increase the TTL value to retain more performance data although more storage space is used
NOTE: The TTL value is set in seconds.
Procedure
To set the or change the TTL value for HBASE, perform the following:
- Login to the Control Center host.
- Attach to the writer container:
serviced service attach writer
- Change to the opentsdb directory:
cd /opt/opentsdb
- Set the TTL value, using seconds:
./set-opentsdb-table-ttl.sh {seconds}
For example, to set the TTL to 90 days, use the value 7776000 seconds:
./set-opentsdb-table-ttl.sh 7776000
- Exit the container:
exit
Could we get steps on showing the current setting?
I agree it would be nice to see what its currently set to.
On Control Center as root I ran the following. A user with access to the Zenoss daemons can also be used. Depending on your setup, you may need to access writer from the host running the container.
# serviced service attach writer
# ./opt/hbase/bin/hbase shell
> list
substitute your current app id or tenant id for acz..908 below
> describe 'acz1mly60doe36ymkfgcj4908-tsdb'
In the output I see the following
, TTL => '7776000 SECONDS (90 DAYS)',