Cassandra Repair Intervals

Question

The Cassandra documentation recommends to run the nodetool repair for every GC seconds (10 days), but this nodetool repair command takes more time and resources. Hence to reduce time and resources I followed partitioner range repair mechanism (nodetool repair -pr) on each node of the data center as mentioned in Cassandra docs:

https://docs.datastax.com/en/cassandra/2.1/cassandra/operations/opsRepairNodesManualRepair.html

When I used nodetool repair -pr I could see the time taken is less as compared to nodetool repair.

Hence currently I am running nodetool repair -pr command on each node of all my data centers one by one.

I want to know can we run nodetool repair -pr command in all the nodes in parallel? (or) Should I run nodetool repair -pr command with a time difference between each nodes?

You can do both, but it all depends... if you don't need to access the cluster from your App.s meanwhile and can and will prioritize all resources to the repair for the duration. Otherwise I would recommend to run with suitable time gap preferable off peak usage hours. We have a 14 node v.2.1 cluster using vnodes, where we repair 2 nodes per night, thus over a 7 day period we get to repair all nodes with -pr. — Steffen Winther Sørensen, Jan 17 '17 at 10:05
possibly resource manage compaction IO with 'nodetool setcompactionthroughput X' to leave some IO capacity for application responses. — Steffen Winther Sørensen, Jan 17 '17 at 10:22
@SteffenWintherSørensen, thanks we have scheduled the nodetool repair -pr in parallel for all my Cassandra nodes, and it works good. — Shoban Sundar, Feb 09 '17 at 04:54
@ShobanSundar do you mean, you run 'repair -pr -par' on all cluster nodes concurrently or sequential within gc_grace? — Steffen Winther Sørensen, Feb 12 '17 at 22:36
@SteffenWintherSørensen, I am running "repair -pr" on all cluster nodes concurrently within gc_grace. — Shoban Sundar, Feb 13 '17 at 05:44

Cassandra Repair Intervals

0 Answers0