I am running Datastax Enterprise on a cluster of 3 nodes. They all work under the same equipment: 2 Intel Xeon 2.2 Ghz cores, 7 GB of RAM, 4 TB Raid-0
This should be enough to start a light-load cluster that stores less than 1 GB of data.
In most cases, everything is just fine, but it seems that sometimes the running tasks related to the repair service in OpsCenter sometimes get stuck; this causes instability in the node and increased load.
However, if the node reboots, the stuck tasks are not displayed and the download is again at a normal level.
Due to the fact that we do not have a lot of data in our cluster, we use the parameter min_repair_timedefined in opscenterd.confto delay the repair service so that it does not run too often.
It seems a little strange that the tasks that are designated as “Complete” and show 100% progress do not disappear, and yes, we waited an hour for them to leave, but they won’t; the only way we decided to solve is to restart the nodes.
Edit:
Here is the conclusion from nodetool compactionstats
Edit 2:
I work under Datastax Enterprise v. 4.6.0 with Cassandra v. 2.0.11.83
Edit 3:
This is inferred from dstatin node, which usually works
This is derived from dstaton node with a closed seal
Change 4:
Output from iostaton node with closed seal, see tall "iowait"
