Houliang Qi created IOTDB-1019:
----------------------------------
Summary: [Distributed] Memory control of the cluster engine
Key: IOTDB-1019
URL: https://issues.apache.org/jira/browse/IOTDB-1019
Project: Apache IoTDB
Issue Type: Improvement
Components: Core/Cluster, Core/Engine
Reporter: Houliang Qi
We are glad to see that robust memory control is provided in a stand-alone
engine to prevent OOM. However, this memory control only considers the memory
occupied by the single engine, and does not consider the memory occupied by the
distributed engine (such as the raft logs kept in the memory), which may cause
OOM in a distributed cluster.
There are two possible solutions to this problem:
1. Set a configuration parameter to control the maximum memory in the view of a
single engine (not the maximum memory of the JVM)
2. Consider the memory occupied by the cluster engine.
I think the first one would easier and helpful enough.
Any ideas?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
