Bharat Viswanadham created HDDS-3072:
----------------------------------------
Summary: SCM scrub pipeline should be started after coming out of
safe mode
Key: HDDS-3072
URL: https://issues.apache.org/jira/browse/HDDS-3072
Project: Hadoop Distributed Data Store
Issue Type: Bug
Reporter: Bharat Viswanadham
Assignee: Bharat Viswanadham
We should start scrubbing pipelines after SCM is out of safe mode.
Reasons to do this:
# Right now, we do scrub pipeline as part of triggerPipelineCreation, now when
we scrub pipelines in allocated state for more than
"ozone.scm.pipeline.allocated.timeout", we might close some pipelines and with
this, we might not be able to come out of safeMode. As in SafeModeRules, we get
pipeline count from pipelineDB during initialization.
Example scenario:
# Stop 3 Datanodes.
# Restart SCM.
# Start Datanode after 6 mts. We shall never come out of safe mode, as
pipeline in allocated state will meet scrubber time out condition.
To not to be in these kinds of scenarios, better thing to be done here is scrub
pipelines after SCM out of the safe mode
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
