AnsweredAssumed Answered

Upgraded nodes not used

Question asked by thealy on Dec 12, 2012
Latest reply on Dec 13, 2012 by Ted Dunning
Running M3 on 20 node cluster / v2.0.0

I recently upgraded the disks on two nodes. After several days, they are still at 0% utilization, even though data is being added constantly. maprcli dump balancerinfo shows them with a fullnesslevel of "Average", but the percentage is 0.

Balancer is enabled with these settings:
<code>
maprcli config load -json | grep "cldb.balancer"
   "cldb.balancer.disk.delta":"0",
   "cldb.balancer.disk.max.switches.in.nodes.percentage":"25",
   "cldb.balancer.disk.paused":"0",
   "cldb.balancer.disk.sleep.interval.sec":"120",
   "cldb.balancer.disk.threshold.percentage":"30",
   "cldb.balancer.logging":"0",
   "cldb.balancer.role.max.switches.in.nodes.percentage":"25",
   "cldb.balancer.role.paused":"0",
   "cldb.balancer.role.skip.container.active.sec":"600",
   "cldb.balancer.role.sleep.interval.sec":"900",
   "cldb.balancer.startup.interval.sec":"1800",
</code>

There are no active container moves. I also tried moving some systems to a different "rack", but that appeared to have no effect. Percentages on the nodes range from 0 to 17 percent full.

How can I get some data spread out to these two nodes?

Thanks

Outcomes