You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-dev@hadoop.apache.org by "Owen O'Malley (JIRA)" <ji...@apache.org> on 2009/05/08 09:11:45 UTC
[jira] Created: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
The length of the heartbeat cycle should be configurable.
---------------------------------------------------------
Key: HADOOP-5784
URL: https://issues.apache.org/jira/browse/HADOOP-5784
Project: Hadoop Core
Issue Type: Improvement
Components: mapred
Reporter: Owen O'Malley
Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amar Kamat (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12713435#action_12713435 ]
Amar Kamat commented on HADOOP-5784:
------------------------------------
Wondering which one is more intuitive, _number-of-heartbeats-per-sec_ or _heartbeat-interval_. The title says _heartbeat-interval_ should be configurable whereas the description states _number-of-heartbeats-per-sec_ should be configurable. I personally think _heartbeat-interval_ is easier to set and play around. Thoughts?
Regarding the test case, cant we spoof tasktracker status and invoke JobTracker.heartbeat() ? This way we can increment the tracker count and query the jobtracker for the current heartbeat interval? Thoughts?
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amareshwari Sriramadasu updated HADOOP-5784:
--------------------------------------------
Status: Patch Available (was: Open)
test-patch result:
{noformat}
[exec]
[exec] +1 overall.
[exec]
[exec] +1 @author. The patch does not contain any @author tags.
[exec]
[exec] +1 tests included. The patch appears to include 3 new or modified tests.
[exec]
[exec] +1 javadoc. The javadoc tool did not generate any warning messages.
[exec]
[exec] +1 javac. The applied patch does not increase the total number of javac compiler warnings.
[exec]
[exec] +1 findbugs. The patch does not introduce any new Findbugs warnings.
[exec]
[exec] +1 Eclipse classpath. The patch retains Eclipse classpath integrity.
[exec]
[exec] +1 release audit. The applied patch does not increase the total number of release audit warnings.
[exec]
{noformat}
ant test passed on my machine
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784-1.txt, patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12715065#action_12715065 ]
Amareshwari Sriramadasu commented on HADOOP-5784:
-------------------------------------------------
Test failures are not related to the patch. All tests passed on my machine
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784-1.txt, patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12711466#action_12711466 ]
Amareshwari Sriramadasu commented on HADOOP-5784:
-------------------------------------------------
Current heartbeat interval is set to {{clusterSize / 100}}, and minimum interval is capped at 3seconds.
It assumes that JT can process 100 heartbeats in a second. See http://issues.apache.org/jira/browse/HADOOP-1900?focusedCommentId=12542530&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#action_12542530
Now, if we make number of heartbeats that should arrive in a second configurable (with default value as 100) , heartbeat interval can be calculated as
{code}
heartbeatInterval = max((clusterSize / #heartbeats in a second), HEARTBEAT_INTERVAL_MIN) ;
{code}
Thoughts?
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12713487#action_12713487 ]
Hadoop QA commented on HADOOP-5784:
-----------------------------------
-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12408937/patch-5784.txt
against trunk revision 778994.
+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 tests are needed for 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 Eclipse classpath. The patch retains Eclipse classpath integrity.
-1 release audit. The applied patch generated 493 release audit warnings (more than the trunk's current 492 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/Hadoop-Patch-vesta.apache.org/410/testReport/
Release audit warnings: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch-vesta.apache.org/410/artifact/trunk/current/releaseAuditDiffWarnings.txt
Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch-vesta.apache.org/410/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Checkstyle results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch-vesta.apache.org/410/artifact/trunk/build/test/checkstyle-errors.html
Console output: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch-vesta.apache.org/410/console
This message is automatically generated.
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Owen O'Malley (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12713438#action_12713438 ]
Owen O'Malley commented on HADOOP-5784:
---------------------------------------
This looks good, but I wish there was a good way to set up a test case. I guess the best way would be to create a JobTracker and call the heartbeat method and observe the requested heartbeat interval.
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amareshwari Sriramadasu updated HADOOP-5784:
--------------------------------------------
Fix Version/s: 0.21.0
Status: Patch Available (was: Open)
test-patch result:
{noformat}
[exec] -1 overall.
[exec]
[exec] +1 @author. The patch does not contain any @author tags.
[exec]
[exec] -1 tests included. The patch doesn't appear to include any new or modified tests.
[exec] Please justify why no tests are needed for this patch.
[exec]
[exec] +1 javadoc. The javadoc tool did not generate any warning messages.
[exec]
[exec] +1 javac. The applied patch does not increase the total number of javac compiler warnings.
[exec]
[exec] +1 findbugs. The patch does not introduce any new Findbugs warnings.
[exec]
[exec] +1 Eclipse classpath. The patch retains Eclipse classpath integrity.
[exec]
[exec] +1 release audit. The applied patch does not increase the total number of release audit warnings.
{noformat}
It is difficult to write unit test for this.
Tested the patch by running sort on 500 nodes with mapred.heartbeats.in.second=200.
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Assigned: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amareshwari Sriramadasu reassigned HADOOP-5784:
-----------------------------------------------
Assignee: Amareshwari Sriramadasu
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amareshwari Sriramadasu updated HADOOP-5784:
--------------------------------------------
Status: Open (was: Patch Available)
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784-1.txt, patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amareshwari Sriramadasu updated HADOOP-5784:
--------------------------------------------
Attachment: patch-5784-1.txt
Patch updated with testcase.
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784-1.txt, patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12714512#action_12714512 ]
Hadoop QA commented on HADOOP-5784:
-----------------------------------
-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12409148/patch-5784-1.txt
against trunk revision 779944.
+1 @author. The patch does not contain any @author tags.
+1 tests included. The patch appears to include 3 new or modified tests.
+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 Eclipse classpath. The patch retains Eclipse classpath integrity.
+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/Hadoop-Patch-vesta.apache.org/426/testReport/
Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch-vesta.apache.org/426/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Checkstyle results: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch-vesta.apache.org/426/artifact/trunk/build/test/checkstyle-errors.html
Console output: http://hudson.zones.apache.org/hudson/job/Hadoop-Patch-vesta.apache.org/426/console
This message is automatically generated.
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784-1.txt, patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Amareshwari Sriramadasu (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Amareshwari Sriramadasu updated HADOOP-5784:
--------------------------------------------
Attachment: patch-5784.txt
Patch making the number of heartbeats that arrive JobTracker configurable.
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Attachments: patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Updated: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Devaraj Das (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Devaraj Das updated HADOOP-5784:
--------------------------------
Resolution: Fixed
Status: Resolved (was: Patch Available)
I just committed this. Thanks, Amareshwari!
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
> Assignee: Amareshwari Sriramadasu
> Fix For: 0.21.0
>
> Attachments: patch-5784-1.txt, patch-5784.txt
>
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (HADOOP-5784) The length of the heartbeat cycle
should be configurable.
Posted by "Steve Loughran (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/HADOOP-5784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12707966#action_12707966 ]
Steve Loughran commented on HADOOP-5784:
----------------------------------------
is the goal here to detect failures of data nodes, or to have an up to date track of which task trackers have capacity?
> The length of the heartbeat cycle should be configurable.
> ---------------------------------------------------------
>
> Key: HADOOP-5784
> URL: https://issues.apache.org/jira/browse/HADOOP-5784
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: Owen O'Malley
>
> Currently, the hearbeat cycle is set to (# nodes / 100) in seconds. This can be too long for clusters that need to run low latency jobs. We should make the number of heartbeats that should arrive a second configurable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.