[
https://issues.apache.org/jira/browse/HDDS-8063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17807847#comment-17807847
]
Ethan Rose commented on HDDS-8063:
----------------------------------
Unfortunately we don't know the root cause or have a way to repro the
inconsistency. We only merged a workaround in HDDS-7156 that mitigates the
issue when it occurs. On an existing cluster you can tell when the issue has
occurred when we see the log message {{No transaction found in container <id>
with pending delete block count <count>}}. This means the workaround in
HDDS-7156 was used to clear the pending delete block count. Your best bet to
repro on an existing cluster would probably to create and delete a large amount
of data.
> Container pending delete block count may be inaccurate
> ------------------------------------------------------
>
> Key: HDDS-8063
> URL: https://issues.apache.org/jira/browse/HDDS-8063
> Project: Apache Ozone
> Issue Type: Bug
> Reporter: Ethan Rose
> Assignee: Sadanand Shenoy
> Priority: Major
>
> In HDDS-7156, it was noticed that the pending delete block count for
> containers could be higher than its actual value, eventually getting to a
> point where the container had no delete transactions but positive pending
> delete block count. In that jira a workaround was merged so this wouldn't
> prevent blocks from being deleted. In this jira we should find the cause for
> the drift and fix it.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
