[
https://issues.apache.org/jira/browse/HDFS-13538?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18016950#comment-18016950
]
ASF GitHub Bot commented on HDFS-13538:
---------------------------------------
shkhrgpt opened a new pull request, #7918:
URL: https://github.com/apache/hadoop/pull/7918
<!--
Thanks for sending a pull request!
1. If this is your first time, please read our contributor guidelines:
https://cwiki.apache.org/confluence/display/HADOOP/How+To+Contribute
2. Make sure your PR title starts with JIRA issue id, e.g.,
'HADOOP-17799. Your PR title ...'.
-->
### Description of PR
DiskChecker was enhanced to do checking with some i/o in
[HADOOP-13738](https://issues.apache.org/jira/browse/HADOOP-13738). But this
was rolled back partially because of fsync issue seen in
[HADOOP-15450](https://issues.apache.org/jira/browse/HADOOP-15450) and the
problem of disk full being flagged as a check failure
([HDFS-13538](https://issues.apache.org/jira/browse/HDFS-13538)).
This PR tries to address
[HDFS-13538](https://issues.apache.org/jira/browse/HDFS-13538) and enable i/o
based disk checking ONLY for HDFS with a flag that can be turned on in dfs
configs.
### How was this patch tested?
```
[INFO] -------------------------------------------------------
[INFO] T E S T S
[INFO] -------------------------------------------------------
[INFO] Running org.apache.hadoop.util.TestDiskCheckerWithDiskIo
[INFO] Tests run: 6, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.129
s
> HDFS DiskChecker should handle disk full situation
> --------------------------------------------------
>
> Key: HDFS-13538
> URL: https://issues.apache.org/jira/browse/HDFS-13538
> Project: Hadoop HDFS
> Issue Type: Bug
> Reporter: Kihwal Lee
> Assignee: Arpit Agarwal
> Priority: Critical
> Labels: pull-request-available
> Attachments: HDFS-13538.01.patch
>
>
> Fix disk checker issues reported by [~kihwal] in HADOOP-13738:
> When space is low, the os returns ENOSPC. Instead simply stop writing, the
> drive is marked bad and replication happens. This make cluster-wide space
> problem worse. If the number of "failed" drives exceeds the DFIP limit, the
> datanode shuts down.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
