[Problem] The shuffle server memory not release

[wfxxh]

uniffle version: 0.6.0, I deploy on k8s,but the shuffle server memory not release ,even my spark application is stop it does not release too my server config is below:

rss.rpc.server.port 19999 rss.jetty.http.port 19998 rss.rpc.executor.size 2000 rss.storage.type MEMORY_LOCALFILE_HDFS rss.coordinator.quorum 10.100.41.162:19999 rss.server.disk.capacity 50g rss.storage.basePath /home/data rss.server.flush.thread.alive 1 rss.server.flush.threadPool.size 10 rss.server.buffer.capacity 4g rss.server.read.buffer.capacity 2g rss.server.heartbeat.timeout 60000 rss.server.heartbeat.interval 10000 rss.rpc.message.max.size 1073741824 rss.server.preAllocation.expired 120000 rss.server.commit.timeout 600000 rss.server.app.expired.withoutHeartbeat 120000 rss.server.flush.cold.storage.threshold.size 128m

[jerqi]

JVM can occupy the memory although they don't process any data.

[wfxxh]

But when the memory is full, the shuffle server pod restart,this case my spark application faild

[jerqi]

Why do the shuffle server restart? There should be some information in the logs or stdout.

[wfxxh]

It is restart by k8s, reason is memory is too high.I think if memory release ,it will not be appear

[jerqi]

It is restart by k8s, reason is memory is too high.I think if memory release ,it will not be appear

Maybe we should give more memory to the pod.

[wfxxh]

It is 32G now,I can not give more

[jerqi]

You can adjust the parameter of memory in the bin/rss-env.sh and conf/server.conf.

[wfxxh]

XMX_SIZE ? it is 30G now.

[jerqi]

XMX_SIZE ? it is 30G now.

Could you reduce the value?

[wfxxh]

I have reduced it to 8G，but the pod restart too

[jerqi]

I have reduced it to 8G，but the pod restart too

Does the server restart because of the same reason? You give the pod 32G memory, XMX_SIZE is 8G, don't it?