Hello, Eugene!

> 2) How come Non-heap memory usage is minimal?
"Non-heap memory" here it's JVM managed memory regions other then heap used
for internal JVM purposes (JIT compiler, method area, etc.), it's not a
memory used by Ignite to store data (information about this memory can be
obtained by data region metrics).

> 3) How can I tell how much memory the table is consuming?
AFAIK there is no such functionality in visor now. There is JMX metric
CacheGroupMetricsMXBean#getTotalAllocatedSize, which can help you, but
unfortunately there is the bug in current implementation of this metric
with persistent store enabled (ticket [1], already fixed, fix will be
available in Ignite 2.7) and there still no implementation of this metric
with persistence disabled (ticket [2]).

[1]: https://issues.apache.org/jira/browse/IGNITE-8515
[2]: https://issues.apache.org/jira/browse/IGNITE-8517


2018-08-17 18:37 GMT+03:00 Alexey Kuznetsov <[email protected]>:

> Hi!
>
> > 1) In Data region metrics, why is everything 0?
> Did you enable metrics?
>
> See:
>   DataRegionConfiguration dataRegionCfg = new DataRegionConfiguration();
>   dataRegionCfg.setMetricsEnabled(true);
>
> > 4) Total busy time is 15s, the upload took longer than that.
> This is actually time spend in compute engine see:
> org.apache.ignite.cluster.ClusterMetrics
>     /**
>      * Gets total time this node spent executing jobs.
>      * @return Total time this node spent executing jobs.
>      */
>     public long getTotalBusyTime();
>
> I hope 2) & 3) will answer some one who knows about it.
>
> --
> Alexey Kuznetsov
>
>

Reply via email to