[
https://issues.apache.org/jira/browse/HBASE-20745?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HBASE-20745:
--------------------------
Release Note: Log Master WAL Proc at INFO level so can tell where we
transition; will help debugging/figuring accounting. Also change
DEFAULT_RIT_CHORE_INTERVAL_MSEC from 5 to 60 seconds; makes it so we emit STUCK
RIT notice once a minute only rather than 12 times a minute (latter was causing
us to quickly roll-away the logging around problem 'events').
> Log when master proc wal rolls
> ------------------------------
>
> Key: HBASE-20745
> URL: https://issues.apache.org/jira/browse/HBASE-20745
> Project: HBase
> Issue Type: Sub-task
> Components: debugging
> Reporter: stack
> Assignee: stack
> Priority: Major
> Fix For: 2.0.2
>
> Attachments: HBASE-20745.master.001.patch
>
>
> Emit when we roll master proc WAL so can see when they happen. Want to
> correlate instances of corruption w/ events on Master. Currently hard to do
> on a server where log-level is INFO (default for many deploys).
> Also, we log STUCK regions every 5 seconds. If a bundle of regions get stuck,
> we can log so frequently, we roll away where the problem happened so lose the
> chance to debug. Let me fix that too....
> Need both debugging instances of parent issue.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
