FAQ
Hi all,

Recently 3 machines of my cluster stopped working with the following message

Shutting down. Incompatible version or revision.TaskTracker version '2.0.0-mr1-cdh4.1.2' and build '2.0.0-mr1-cdh4.1.2 from Unknown by jenkins source checksum b7f8aa714447372afd71f8fc3505b304' and JobTracker version '2.0.0-mr1-cdh4.1.3' and build '2.0.0-mr1-cdh4.1.3 from Unknown by jenkins source checksum 0222a5fdd4d9e3a2ea8f499b8a275db0 and hadoop.relaxed.worker.version.check is enabled and hadoop.skip.worker.version.check is not enabled


I did no updates of changed any configuration manually. They were OK in one
second, I restarted the cluster and in the other this happened.

In all machines that are working properly I have
in /usr/lib/hadoop-0.20-mapreduce the version with end "1.3".
In those 3 that stopped working I have in /usr/lib/hadoop-0.20-mapreduce
the version of jars ending with "1.2"

Just a remark: all machines have the same configuration and had hadoop
installed in the same day (at same time).

My ideas are:
a) try to update those 3 machines with problems;
b) Copy the 1.3 .jar files to these machines.

Anyone has any idea what caused this problem or how I can solve it?

Thanks in advance,

Best Regards,

Giovanni Comarela

Search Discussions

  • Harsh J at Feb 5, 2013 at 7:21 am
    It may have been done as a result of an automatic software repository
    update on your machines? CM and CDH both do not include any automatic,
    non-user-triggered update mechanisms so far.

    I'd say fix it via (a), its the simplest and most proper route, and
    you benefit from the critical bug fixes we've added to the refresh
    version (4.1.2 -> 4.1.3).

    On Tue, Feb 5, 2013 at 8:30 AM, Giovanni Comarela
    wrote:
    Hi all,

    Recently 3 machines of my cluster stopped working with the following message

    Shutting down. Incompatible version or revision.TaskTracker version
    '2.0.0-mr1-cdh4.1.2' and build '2.0.0-mr1-cdh4.1.2 from Unknown by jenkins
    source checksum b7f8aa714447372afd71f8fc3505b304' and JobTracker version
    '2.0.0-mr1-cdh4.1.3' and build '2.0.0-mr1-cdh4.1.3 from Unknown by jenkins
    source checksum 0222a5fdd4d9e3a2ea8f499b8a275db0 and
    hadoop.relaxed.worker.version.check is enabled and
    hadoop.skip.worker.version.check is not enabled


    I did no updates of changed any configuration manually. They were OK in one
    second, I restarted the cluster and in the other this happened.

    In all machines that are working properly I have in
    /usr/lib/hadoop-0.20-mapreduce the version with end "1.3".
    In those 3 that stopped working I have in /usr/lib/hadoop-0.20-mapreduce the
    version of jars ending with "1.2"

    Just a remark: all machines have the same configuration and had hadoop
    installed in the same day (at same time).

    My ideas are:
    a) try to update those 3 machines with problems;
    b) Copy the 1.3 .jar files to these machines.

    Anyone has any idea what caused this problem or how I can solve it?

    Thanks in advance,

    Best Regards,

    Giovanni Comarela


    --
    Harsh J
  • Giovanni at Feb 5, 2013 at 5:07 pm
    Hi,

    Thanks,

    Just to give a feedback. I ran Host Upgrade Wizard at Cloudera Manager and
    everything is working fine now.

    Thanks,

    Giovanni Comarela
    On Tue, Feb 5, 2013 at 2:21 AM, Harsh J wrote:

    It may have been done as a result of an automatic software repository
    update on your machines? CM and CDH both do not include any automatic,
    non-user-triggered update mechanisms so far.

    I'd say fix it via (a), its the simplest and most proper route, and
    you benefit from the critical bug fixes we've added to the refresh
    version (4.1.2 -> 4.1.3).

    On Tue, Feb 5, 2013 at 8:30 AM, Giovanni Comarela
    wrote:
    Hi all,

    Recently 3 machines of my cluster stopped working with the following message
    Shutting down. Incompatible version or revision.TaskTracker version
    '2.0.0-mr1-cdh4.1.2' and build '2.0.0-mr1-cdh4.1.2 from Unknown by jenkins
    source checksum b7f8aa714447372afd71f8fc3505b304' and JobTracker version
    '2.0.0-mr1-cdh4.1.3' and build '2.0.0-mr1-cdh4.1.3 from Unknown by jenkins
    source checksum 0222a5fdd4d9e3a2ea8f499b8a275db0 and
    hadoop.relaxed.worker.version.check is enabled and
    hadoop.skip.worker.version.check is not enabled


    I did no updates of changed any configuration manually. They were OK in one
    second, I restarted the cluster and in the other this happened.

    In all machines that are working properly I have in
    /usr/lib/hadoop-0.20-mapreduce the version with end "1.3".
    In those 3 that stopped working I have in /usr/lib/hadoop-0.20-mapreduce the
    version of jars ending with "1.2"

    Just a remark: all machines have the same configuration and had hadoop
    installed in the same day (at same time).

    My ideas are:
    a) try to update those 3 machines with problems;
    b) Copy the 1.3 .jar files to these machines.

    Anyone has any idea what caused this problem or how I can solve it?

    Thanks in advance,

    Best Regards,

    Giovanni Comarela


    --
    Harsh J

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupscm-users @
categorieshadoop
postedFeb 5, '13 at 3:00a
activeFeb 5, '13 at 5:07p
posts3
users2
websitecloudera.com
irc#hadoop

2 users in discussion

Giovanni: 2 posts Harsh J: 1 post

People

Translate

site design / logo © 2022 Grokbase