You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-issues@hadoop.apache.org by "Todd Lipcon (JIRA)" <ji...@apache.org> on 2010/03/21 09:11:27 UTC
[jira] Created: (HDFS-1054) Make sleep after failure in
nextBlockOutputStream smarter and configurable
Make sleep after failure in nextBlockOutputStream smarter and configurable
--------------------------------------------------------------------------
Key: HDFS-1054
URL: https://issues.apache.org/jira/browse/HDFS-1054
Project: Hadoop HDFS
Issue Type: Improvement
Components: hdfs client
Affects Versions: 0.22.0
Reporter: Todd Lipcon
Assignee: Todd Lipcon
If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-1054) Remove unnecessary sleep after
failure in nextBlockOutputStream
Posted by "Jakob Homan (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12860628#action_12860628 ]
Jakob Homan commented on HDFS-1054:
-----------------------------------
Changing to INFO should work fine.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Jakob Homan (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jakob Homan updated HDFS-1054:
------------------------------
Status: Open (was: Patch Available)
Canceling pending requested revision.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-1054) Remove unnecessary sleep after
failure in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12853632#action_12853632 ]
Todd Lipcon commented on HDFS-1054:
-----------------------------------
bq. -1 tests included
Not sure how I would test this - it's very difficult to test for the lack of a sleep :)
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-1054) Make sleep after failure in
nextBlockOutputStream smarter and configurable
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12847897#action_12847897 ]
Todd Lipcon commented on HDFS-1054:
-----------------------------------
FYI the situation where I'm running into this is an hbase edits stress test where I'm killing a DN that's local to one of the region servers. That region server immediately starts acting up because all file creates take an extra 6 seconds (it still thinks the local DN is up until the NN marks it down). In this case it's getting an immediate "Connection Refused" from the local DN anyway, and the second attempt always works fine since it makes it into the HDFS-630 excludedNodes list
> Make sleep after failure in nextBlockOutputStream smarter and configurable
> --------------------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Fix Version/s: 0.20-append
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.20-append, 0.20.3, 0.21.0, 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.20-append, 0.21.0
>
> Attachments: hdfs-1054-0.20-append.txt, hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-1054) Remove unnecessary sleep after
failure in nextBlockOutputStream
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12860771#action_12860771 ]
Hadoop QA commented on HDFS-1054:
---------------------------------
-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12442793/hdfs-1054.txt
against trunk revision 937855.
+1 @author. The patch does not contain any @author tags.
-1 tests included. The patch doesn't appear to include any new or modified tests.
Please justify why no new tests are needed for this patch.
Also please list what manual steps were performed to verify this patch.
+1 javadoc. The javadoc tool did not generate any warning messages.
+1 javac. The applied patch does not increase the total number of javac compiler warnings.
+1 findbugs. The patch does not introduce any new Findbugs warnings.
+1 release audit. The applied patch does not increase the total number of release audit warnings.
-1 core tests. The patch failed core unit tests.
-1 contrib tests. The patch failed contrib unit tests.
Test results: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h2.grid.sp2.yahoo.net/160/testReport/
Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h2.grid.sp2.yahoo.net/160/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Checkstyle results: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h2.grid.sp2.yahoo.net/160/artifact/trunk/build/test/checkstyle-errors.html
Console output: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h2.grid.sp2.yahoo.net/160/console
This message is automatically generated.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Reopened: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon reopened HDFS-1054:
-------------------------------
Reopening for commit to append branch
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.20-append
>
> Attachments: hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Make sleep after failure in
nextBlockOutputStream smarter and configurable
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Status: Patch Available (was: Open)
> Make sleep after failure in nextBlockOutputStream smarter and configurable
> --------------------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Attachment: hdfs-1054.txt
New patch addresses Jakob's review comments. Thanks!
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-1054) Remove unnecessary sleep after
failure in nextBlockOutputStream
Posted by "Jakob Homan (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12860625#action_12860625 ]
Jakob Homan commented on HDFS-1054:
-----------------------------------
Patch looks good except for unnecessary changes to log output (combining two log entries into one). We've not nailed down the log output format yet, but there's still no need to change it without need, in case someone is currently parsing its output.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Resolved: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "dhruba borthakur (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
dhruba borthakur resolved HDFS-1054.
------------------------------------
Fix Version/s: 0.21.0
(was: 0.20-append)
Resolution: Fixed
I committed it into the hadoop-0.20-append branch (as indicated by the affected version) but reverted the fix-version to the original value of 0.21.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.20-append, 0.20.3, 0.21.0, 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.21.0
>
> Attachments: hdfs-1054-0.20-append.txt, hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-1054) Remove unnecessary sleep after
failure in nextBlockOutputStream
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12853627#action_12853627 ]
Hadoop QA commented on HDFS-1054:
---------------------------------
-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12440729/hdfs-1054.txt
against trunk revision 930967.
+1 @author. The patch does not contain any @author tags.
-1 tests included. The patch doesn't appear to include any new or modified tests.
Please justify why no new tests are needed for this patch.
Also please list what manual steps were performed to verify this patch.
+1 javadoc. The javadoc tool did not generate any warning messages.
+1 javac. The applied patch does not increase the total number of javac compiler warnings.
+1 findbugs. The patch does not introduce any new Findbugs warnings.
+1 release audit. The applied patch does not increase the total number of release audit warnings.
+1 core tests. The patch passed core unit tests.
+1 contrib tests. The patch passed contrib unit tests.
Test results: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h5.grid.sp2.yahoo.net/298/testReport/
Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h5.grid.sp2.yahoo.net/298/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Checkstyle results: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h5.grid.sp2.yahoo.net/298/artifact/trunk/build/test/checkstyle-errors.html
Console output: http://hudson.zones.apache.org/hudson/job/Hdfs-Patch-h5.grid.sp2.yahoo.net/298/console
This message is automatically generated.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Status: Patch Available (was: Open)
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Jakob Homan (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jakob Homan updated HDFS-1054:
------------------------------
Status: Resolved (was: Patch Available)
Resolution: Fixed
Thanks for the revision. I can't reproduce the failed test. Contrib failure is unrelated. +1.
I've committed this. Thanks Todd. Resolving as fixed.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Attachment: hdfs-1054-0.20-append.txt
Attaching patch for 0.20-append branch. It's just slightly different than what went into trunk (did less cleanup to minimize change). No unit test but has been cluster tested for last several months with lots of failure testing.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.20-append
>
> Attachments: hdfs-1054-0.20-append.txt, hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "dhruba borthakur (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
dhruba borthakur updated HDFS-1054:
-----------------------------------
Affects Version/s: 0.20-append
0.20.3
0.21.0
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.20-append, 0.20.3, 0.21.0, 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.20-append
>
> Attachments: hdfs-1054-0.20-append.txt, hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HDFS-1054) Remove unnecessary sleep after
failure in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12860627#action_12860627 ]
Todd Lipcon commented on HDFS-1054:
-----------------------------------
I combined the log entries because I think the bad datanode should be INFO level. Would you prefer that I just change that message to INFO rather than combining them at the same time? Given that the abandonment is a direct result of the bad node, I think it makes more sense to have one message. Also, the node exclusion log is from HDFS-630 which hasn't been released yet, so I don't think we risk that anyone is parsing it.
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Fix Version/s: 0.20-append
Marking for 0.20-append
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.20-append
>
> Attachments: hdfs-1054.txt, hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Remove unnecessary sleep after failure
in nextBlockOutputStream
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Summary: Remove unnecessary sleep after failure in nextBlockOutputStream (was: Make sleep after failure in nextBlockOutputStream smarter and configurable)
> Remove unnecessary sleep after failure in nextBlockOutputStream
> ---------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HDFS-1054) Make sleep after failure in
nextBlockOutputStream smarter and configurable
Posted by "Todd Lipcon (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HDFS-1054?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Todd Lipcon updated HDFS-1054:
------------------------------
Attachment: hdfs-1054.txt
Here's a patch which removes the sleep altogether. I looked through the svn logs and this sleep has been there since Nutch, and I can't think of any particular reason why it's useful.
If someone can think of a time when it actually is a good idea to sleep, I'll upload a patch that makes it configurable, but otherwise I think we should just remove it.
> Make sleep after failure in nextBlockOutputStream smarter and configurable
> --------------------------------------------------------------------------
>
> Key: HDFS-1054
> URL: https://issues.apache.org/jira/browse/HDFS-1054
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs client
> Affects Versions: 0.22.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Attachments: hdfs-1054.txt
>
>
> If DFSOutputStream fails to create a pipeline, it currently sleeps 6 seconds before retrying. I don't see a great reason to wait at all, much less 6 seconds (especially now that HDFS-630 ensures that a retry won't go back to the bad node). We should at least make it configurable, and perhaps something like backoff makes some sense.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.