You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ozone.apache.org by "Lokesh Jain (Jira)" <ji...@apache.org> on 2020/11/19 15:01:00 UTC
[jira] [Commented] (HDDS-4356) SCM is flooded with useless
"Deleting blocks" messages
[ https://issues.apache.org/jira/browse/HDDS-4356?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17235532#comment-17235532 ]
Lokesh Jain commented on HDDS-4356:
-----------------------------------
In my cluster I can see the block and container ID. I am not sure why it is not showing the container and block ID here. I have created a PR to move the logging to debug level.
{code:java}
2020-11-12 00:19:59,733 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks conID: 653 locID: 105192278022701717 bcsId: 0
2020-11-12 00:19:59,733 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks conID: 651 locID: 105192293888660519 bcsId: 0
2020-11-12 00:19:59,733 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks conID: 685 locID: 105192404974048262 bcsId: 0
2020-11-12 00:19:59,733 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks conID: 748 locID: 105192593469751662 bcsId: 0
{code}
Related code.
{code:java}
LOG.info("Deleting blocks {}",
StringUtils.join(",", bg.getBlockIDList()));
{code}
This should ideally happen only if OM is sending an empty block list.
> SCM is flooded with useless "Deleting blocks" messages
> ------------------------------------------------------
>
> Key: HDDS-4356
> URL: https://issues.apache.org/jira/browse/HDDS-4356
> Project: Hadoop Distributed Data Store
> Issue Type: Improvement
> Components: SCM
> Affects Versions: 1.0.0
> Reporter: Wei-Chiu Chuang
> Assignee: Lokesh Jain
> Priority: Major
> Labels: pull-request-available
>
> Testing a 1.0.0 SCM.
> I'm seeing these messages flooding SCM log file when a dead DN is detected.
> {noformat}
> 2020-10-19 13:48:19,642 INFO org.apache.hadoop.hdds.scm.node.DeadNodeHandler: A dead datanode is detected. 9b27c38d-9104-491b-b76b-959dc9dd06a2{ip: 10.12.1.82, host: rhel12.ozone.local, networkLocation: /default, certSerialId: null}
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.server.SCMBlockProtocolServer: SCM is informed by OM to delete 1000 blocks
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,894 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,895 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,895 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> 2020-10-19 13:48:24,895 INFO org.apache.hadoop.hdds.scm.block.BlockManagerImpl: Deleting blocks
> {noformat}
> SCM deletes 1000 blocks max at a time, and the "Deleting blocks" message repeats for 1000 times. Worse, it doesn't give any useful information (at the very least, it should print the block id)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org