ZooKeeper
  1. ZooKeeper
  2. ZOOKEEPER-608

Receipt of ACK from observer should not be logged as ERROR

    Details

    • Type: Improvement Improvement
    • Status: Closed
    • Priority: Critical Critical
    • Resolution: Fixed
    • Affects Version/s: 3.3.0
    • Fix Version/s: 3.3.0
    • Component/s: None
    • Labels:
      None
    • Hadoop Flags:
      Reviewed

      Description

      Observers - in general - don't send ACKs. But a couple of times they need to. Currently, these are all logged as an ERROR, which is wrong. They should at most be WARN (and this would probably be confusing to the user). INFO might be better.

      1. ZOOKEEPER-608.patch
        0.9 kB
        Henry Robinson

        Activity

        Transition Time In Source Status Execution Times Last Executer Last Execution Date
        Patch Available Patch Available Open Open
        1h 33m 1 Henry Robinson 16/Jan/10 00:34
        Open Open Patch Available Patch Available
        43d 22h 34m 2 Henry Robinson 16/Jan/10 00:35
        Patch Available Patch Available Resolved Resolved
        5d 1h 10m 1 Mahadev konar 21/Jan/10 01:45
        Resolved Resolved Closed Closed
        64d 15h 39m 1 Patrick Hunt 26/Mar/10 17:25
        Patrick Hunt made changes -
        Status Resolved [ 5 ] Closed [ 6 ]
        Hide
        Hudson added a comment -

        Integrated in ZooKeeper-trunk #674 (See http://hudson.zones.apache.org/hudson/job/ZooKeeper-trunk/674/)
        . Receipt of ACK from observer should not be logged as ERROR (henry via mahadev)

        Show
        Hudson added a comment - Integrated in ZooKeeper-trunk #674 (See http://hudson.zones.apache.org/hudson/job/ZooKeeper-trunk/674/ ) . Receipt of ACK from observer should not be logged as ERROR (henry via mahadev)
        Mahadev konar made changes -
        Status Patch Available [ 10002 ] Resolved [ 5 ]
        Hadoop Flags [Reviewed]
        Resolution Fixed [ 1 ]
        Hide
        Mahadev konar added a comment -

        I just committed this. thanks henry!

        Show
        Mahadev konar added a comment - I just committed this. thanks henry!
        Hide
        Mahadev konar added a comment -

        +1 patch looks good. I will go ahead and commit it.

        henry, please to open a jira for testobservers hang and please mark it for 3.3.

        Show
        Mahadev konar added a comment - +1 patch looks good. I will go ahead and commit it. henry, please to open a jira for testobservers hang and please mark it for 3.3.
        Hide
        Henry Robinson added a comment -

        Flavio -

        Observers send two ACKs each during the synchronisation protocol with the leader. These are legitimate responses so that the Leader can know that they're properly synced. We can potentially remove them, but it seems like more trouble than it's worth.

        I don't mind removing these debug messages, but I feel like they could be helpful if we do some Observers development in the future to verify any changes to the protocol. Happy to take them out though.

        We can open a JIRA for the testObservers hang, and I will take a look. I can't reproduce the failure but if it happens again, it'll be good to have somewhere to track it.

        Henry

        Show
        Henry Robinson added a comment - Flavio - Observers send two ACKs each during the synchronisation protocol with the leader. These are legitimate responses so that the Leader can know that they're properly synced. We can potentially remove them, but it seems like more trouble than it's worth. I don't mind removing these debug messages, but I feel like they could be helpful if we do some Observers development in the future to verify any changes to the protocol. Happy to take them out though. We can open a JIRA for the testObservers hang, and I will take a look. I can't reproduce the failure but if it happens again, it'll be good to have somewhere to track it. Henry
        Hide
        Flavio Junqueira added a comment -

        Pat, Henry, I'm confused. Observers really shouldn't send ACKs, and if they do we have a bug. If we are seeing many of these, don't we have an issue with the observer logic then? Now, the patch is clearly good, my question is if receiving ACKs from observers should be really at debug level.

        Also, I checked the hudson run that failed, and it seems that it timed out on testObserver. Henry is right in that it doesn't seem to be related to his patch, but I wonder if we should open a new jira or not. There isn't enough to work with other than "testObserver timed out oin hudson".

        Show
        Flavio Junqueira added a comment - Pat, Henry, I'm confused. Observers really shouldn't send ACKs, and if they do we have a bug. If we are seeing many of these, don't we have an issue with the observer logic then? Now, the patch is clearly good, my question is if receiving ACKs from observers should be really at debug level. Also, I checked the hudson run that failed, and it seems that it timed out on testObserver. Henry is right in that it doesn't seem to be related to his patch, but I wonder if we should open a new jira or not. There isn't enough to work with other than "testObserver timed out oin hudson".
        Hide
        Hadoop QA added a comment -

        -1 overall. Here are the results of testing the latest attachment
        http://issues.apache.org/jira/secure/attachment/12430453/ZOOKEEPER-608.patch
        against trunk revision 899383.

        +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 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/Zookeeper-Patch-h8.grid.sp2.yahoo.net/106/testReport/
        Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/106/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
        Console output: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/106/console

        This message is automatically generated.

        Show
        Hadoop QA added a comment - -1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12430453/ZOOKEEPER-608.patch against trunk revision 899383. +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 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/Zookeeper-Patch-h8.grid.sp2.yahoo.net/106/testReport/ Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/106/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html Console output: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/106/console This message is automatically generated.
        Henry Robinson made changes -
        Status Open [ 1 ] Patch Available [ 10002 ]
        Hide
        Henry Robinson added a comment -

        Surprised at this failure, not totally sure if it's related to this patch. Re-starting Hudson to find out,

        Show
        Henry Robinson added a comment - Surprised at this failure, not totally sure if it's related to this patch. Re-starting Hudson to find out,
        Henry Robinson made changes -
        Status Patch Available [ 10002 ] Open [ 1 ]
        Hide
        Hadoop QA added a comment -

        -1 overall. Here are the results of testing the latest attachment
        http://issues.apache.org/jira/secure/attachment/12430453/ZOOKEEPER-608.patch
        against trunk revision 899383.

        +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 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 passed contrib unit tests.

        Test results: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/105/testReport/
        Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/105/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
        Console output: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/105/console

        This message is automatically generated.

        Show
        Hadoop QA added a comment - -1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12430453/ZOOKEEPER-608.patch against trunk revision 899383. +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 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 passed contrib unit tests. Test results: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/105/testReport/ Findbugs warnings: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/105/artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html Console output: http://hudson.zones.apache.org/hudson/job/Zookeeper-Patch-h8.grid.sp2.yahoo.net/105/console This message is automatically generated.
        Henry Robinson made changes -
        Status Open [ 1 ] Patch Available [ 10002 ]
        Henry Robinson made changes -
        Attachment ZOOKEEPER-608.patch [ 12430453 ]
        Hide
        Henry Robinson added a comment -

        Patch attached - made the message debug only.

        Rationale: we don't protect against Byzantine failures, which is effectively what this tests for, so no point defensively logging an error (even if I made it only applicable to ACKs that are errors). Better to make it a debug so that developers who understand the messages can turn them on when fixing bugs.

        No tests in this one - but all tests in trunk pass.

        Show
        Henry Robinson added a comment - Patch attached - made the message debug only. Rationale: we don't protect against Byzantine failures, which is effectively what this tests for, so no point defensively logging an error (even if I made it only applicable to ACKs that are errors). Better to make it a debug so that developers who understand the messages can turn them on when fixing bugs. No tests in this one - but all tests in trunk pass.
        Patrick Hunt made changes -
        Field Original Value New Value
        Priority Minor [ 4 ] Critical [ 2 ]
        Hide
        Patrick Hunt added a comment -

        errors confuse/upset users. causes alot of support issues. Plz fix this.

        Show
        Patrick Hunt added a comment - errors confuse/upset users. causes alot of support issues. Plz fix this.
        Henry Robinson created issue -

          People

          • Assignee:
            Henry Robinson
            Reporter:
            Henry Robinson
          • Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development