Mikhail Efremov created IGNITE-28022:
----------------------------------------
Summary: Improper CMG group executor usage
Key: IGNITE-28022
URL: https://issues.apache.org/jira/browse/IGNITE-28022
Project: Ignite
Issue Type: Bug
Reporter: Mikhail Efremov
*Description*
In logs we may found the following logs:
{code:java}
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification
[grpId=metastorage_group, consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_24,
consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_6,
consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_15,
consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_3,
consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_17,
consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_16,
consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_23,
consistentId=node-2]
[INFO][%node-0%JRaft-FSMCaller-Disruptor-cmg_group_stripe_0-0][TopologyAwareRaftGroupService]
New peer will be sending a leader elected notification [grpId=21_part_5,
consistentId=node-2]
{code}
This indicates for wrong executor usage for futures chaining and should be
fixed.
*Motivation*
Wrong load on system group may be a reason for cluster bad performance or
instability.
*Definition of Done*
# There no usages of wrong disruptor treads for partitioning or MS tasks,
system groups should use proper separated executors.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
