FAQ
Having issues with my upgraded Cloudera Manager box. Since upgrading to
4.6.1, I can't start any of the monitor services, serivce monitor, activity
monitor etc.
I went through upgrading just the manager, not the nodes. I was sidetracked
half way through the upgrade wizard, and it logged me out. When I logged
in, it said it was complete, but I think there may have been a step missed
form the upgrade process. The last step I recall was setting up the
postgres embedded databases for the activity monitor etc. This was the
point where it timed out and logged me out.
Is there a way to go through the wizard again, or will this cause more
issues?
Here is the error when I try to start any of the services:

1:41:33.838 PM INFO com.cloudera.cmon.firehose.Main

Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f16077bce0790d4)


Let me know if you need more entries from the logs.

Search Discussions

  • Jaykay at Jul 17, 2013 at 12:59 pm
    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since upgrading to
    4.6.1, I can't start any of the monitor services, serivce monitor, activity
    monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause more
    issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.Main

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f16077bce0790d4)


    Let me know if you need more entries from the logs.
  • Jaykay at Jul 17, 2013 at 1:22 pm
    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.Main

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since upgrading to
    4.6.1, I can't start any of the monitor services, serivce monitor, activity
    monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause more
    issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.Main

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f16077bce0790d4)


    Let me know if you need more entries from the logs.
  • Chris Leroy at Jul 18, 2013 at 2:54 pm
    It looks like your activity monitor -- at least -- is using the old 4.5.0
    version. That would lead to this error you're seeing -- where the activity
    monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be able to
    upgrade the cloudera-manager-daemons package directly to the new version.

    chris

    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay wrote:

    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.Main

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since upgrading to
    4.6.1, I can't start any of the monitor services, serivce monitor, activity
    monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause more
    issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Main

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607**7bce0790d4)


    Let me know if you need more entries from the logs.
  • Jaykay at Jul 18, 2013 at 3:03 pm
    Sounds possible Chris-how would I update the daemons?
    On Thursday, July 18, 2013 3:54:39 PM UTC+1, Chris Leroy wrote:

    It looks like your activity monitor -- at least -- is using the old 4.5.0
    version. That would lead to this error you're seeing -- where the activity
    monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be able to
    upgrade the cloudera-manager-daemons package directly to the new version.

    chris


    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay <johnk...@gmail.com <javascript:>>wrote:
    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.Main

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since upgrading to
    4.6.1, I can't start any of the monitor services, serivce monitor, activity
    monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause more
    issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Main

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607**7bce0790d4)


    Let me know if you need more entries from the logs.
  • Chris Leroy at Jul 19, 2013 at 4:08 am
    You can use whatever package management system is appropriate to your OS,
    for example, yum if you're on RHEL.

    chris

    On Thu, Jul 18, 2013 at 8:03 AM, Jaykay wrote:

    Sounds possible Chris-how would I update the daemons?

    On Thursday, July 18, 2013 3:54:39 PM UTC+1, Chris Leroy wrote:

    It looks like your activity monitor -- at least -- is using the old 4.5.0
    version. That would lead to this error you're seeing -- where the activity
    monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be able to
    upgrade the cloudera-manager-daemons package directly to the new version.

    chris

    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay wrote:

    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.**Main

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since upgrading
    to 4.6.1, I can't start any of the monitor services, serivce monitor,
    activity monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause more
    issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Mai**n

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607****7bce0790d4)


    Let me know if you need more entries from the logs.
  • Jaykay at Jul 23, 2013 at 12:02 pm
    Yum tells me No Packages marked for Update, so I think that rules it out?
    On Friday, July 19, 2013 5:08:18 AM UTC+1, Chris Leroy wrote:

    You can use whatever package management system is appropriate to your OS,
    for example, yum if you're on RHEL.

    chris


    On Thu, Jul 18, 2013 at 8:03 AM, Jaykay <johnk...@gmail.com <javascript:>>wrote:
    Sounds possible Chris-how would I update the daemons?

    On Thursday, July 18, 2013 3:54:39 PM UTC+1, Chris Leroy wrote:

    It looks like your activity monitor -- at least -- is using the old
    4.5.0 version. That would lead to this error you're seeing -- where the
    activity monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be able to
    upgrade the cloudera-manager-daemons package directly to the new version.

    chris

    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay wrote:

    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.**Main

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since upgrading
    to 4.6.1, I can't start any of the monitor services, serivce monitor,
    activity monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause
    more issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Mai**n

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607****7bce0790d4)


    Let me know if you need more entries from the logs.
  • Chris Leroy at Jul 24, 2013 at 3:13 am
    Jarkay,

    Can you try starting up one of the MGMT roles which is not starting up and
    send me the log file? Maybe zip it and send it just to me directly.

    chris

    On Tue, Jul 23, 2013 at 5:02 AM, Jaykay wrote:

    Yum tells me No Packages marked for Update, so I think that rules it out?

    On Friday, July 19, 2013 5:08:18 AM UTC+1, Chris Leroy wrote:

    You can use whatever package management system is appropriate to your OS,
    for example, yum if you're on RHEL.

    chris

    On Thu, Jul 18, 2013 at 8:03 AM, Jaykay wrote:

    Sounds possible Chris-how would I update the daemons?

    On Thursday, July 18, 2013 3:54:39 PM UTC+1, Chris Leroy wrote:

    It looks like your activity monitor -- at least -- is using the old
    4.5.0 version. That would lead to this error you're seeing -- where the
    activity monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be able
    to upgrade the cloudera-manager-daemons package directly to the new version.

    chris

    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay wrote:

    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.**Mai**n

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since upgrading
    to 4.6.1, I can't start any of the monitor services, serivce monitor,
    activity monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause
    more issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Mai****n

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607******7bce0790d4)


    Let me know if you need more entries from the logs.
  • Jaykay at Jul 25, 2013 at 10:03 am
    Hi Chris,
    I mailed you the logs-let me know if you got them.
    On Wednesday, July 24, 2013 4:13:52 AM UTC+1, Chris Leroy wrote:

    Jarkay,

    Can you try starting up one of the MGMT roles which is not starting up and
    send me the log file? Maybe zip it and send it just to me directly.

    chris


    On Tue, Jul 23, 2013 at 5:02 AM, Jaykay <johnk...@gmail.com <javascript:>>wrote:
    Yum tells me No Packages marked for Update, so I think that rules it out?

    On Friday, July 19, 2013 5:08:18 AM UTC+1, Chris Leroy wrote:

    You can use whatever package management system is appropriate to your
    OS, for example, yum if you're on RHEL.

    chris

    On Thu, Jul 18, 2013 at 8:03 AM, Jaykay wrote:

    Sounds possible Chris-how would I update the daemons?

    On Thursday, July 18, 2013 3:54:39 PM UTC+1, Chris Leroy wrote:

    It looks like your activity monitor -- at least -- is using the old
    4.5.0 version. That would lead to this error you're seeing -- where the
    activity monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be able
    to upgrade the cloudera-manager-daemons package directly to the new version.

    chris

    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay wrote:

    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.**Mai**n

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since
    upgrading to 4.6.1, I can't start any of the monitor services, serivce
    monitor, activity monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause
    more issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Mai****n

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607******7bce0790d4)


    Let me know if you need more entries from the logs.
  • Jaykay at Jul 31, 2013 at 9:09 am
    Hi Chris,

    This is still happening-can you let me know if you received the logs by
    email?

    John
    On Thursday, July 25, 2013 11:03:37 AM UTC+1, Jaykay wrote:

    Hi Chris,
    I mailed you the logs-let me know if you got them.
    On Wednesday, July 24, 2013 4:13:52 AM UTC+1, Chris Leroy wrote:

    Jarkay,

    Can you try starting up one of the MGMT roles which is not starting up
    and send me the log file? Maybe zip it and send it just to me directly.

    chris

    On Tue, Jul 23, 2013 at 5:02 AM, Jaykay wrote:

    Yum tells me No Packages marked for Update, so I think that rules it out?

    On Friday, July 19, 2013 5:08:18 AM UTC+1, Chris Leroy wrote:

    You can use whatever package management system is appropriate to your
    OS, for example, yum if you're on RHEL.

    chris

    On Thu, Jul 18, 2013 at 8:03 AM, Jaykay wrote:

    Sounds possible Chris-how would I update the daemons?

    On Thursday, July 18, 2013 3:54:39 PM UTC+1, Chris Leroy wrote:

    It looks like your activity monitor -- at least -- is using the old
    4.5.0 version. That would lead to this error you're seeing -- where the
    activity monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be able
    to upgrade the cloudera-manager-daemons package directly to the new version.

    chris

    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay wrote:

    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.**Mai**n

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since
    upgrading to 4.6.1, I can't start any of the monitor services, serivce
    monitor, activity monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause
    more issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Mai****n

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607******7bce0790d4)


    Let me know if you need more entries from the logs.
  • Chris Leroy at Jul 31, 2013 at 3:34 pm
    Jaykay,

    Apologies for the delay. I don't see the logs. I bet they got spam filtered
    out and then cleaned up. Mind trying to send them again?

    chris

    On Wed, Jul 31, 2013 at 2:09 AM, Jaykay wrote:

    Hi Chris,

    This is still happening-can you let me know if you received the logs by
    email?

    John

    On Thursday, July 25, 2013 11:03:37 AM UTC+1, Jaykay wrote:

    Hi Chris,
    I mailed you the logs-let me know if you got them.
    On Wednesday, July 24, 2013 4:13:52 AM UTC+1, Chris Leroy wrote:

    Jarkay,

    Can you try starting up one of the MGMT roles which is not starting up
    and send me the log file? Maybe zip it and send it just to me directly.

    chris

    On Tue, Jul 23, 2013 at 5:02 AM, Jaykay wrote:

    Yum tells me No Packages marked for Update, so I think that rules it
    out?

    On Friday, July 19, 2013 5:08:18 AM UTC+1, Chris Leroy wrote:

    You can use whatever package management system is appropriate to your
    OS, for example, yum if you're on RHEL.

    chris

    On Thu, Jul 18, 2013 at 8:03 AM, Jaykay wrote:

    Sounds possible Chris-how would I update the daemons?

    On Thursday, July 18, 2013 3:54:39 PM UTC+1, Chris Leroy wrote:

    It looks like your activity monitor -- at least -- is using the old
    4.5.0 version. That would lead to this error you're seeing -- where the
    activity monitor is not able to communicate with the SCM server at startup.

    I'm not sure how to re-run the upgrade wizard, but you should be
    able to upgrade the cloudera-manager-daemons package directly to the new
    version.

    chris

    On Wed, Jul 17, 2013 at 6:22 AM, Jaykay wrote:

    Sorry, I added the wrong line from the logs!
    Should be this one:

    2:21:25.463 PM ERROR com.cloudera.cmon.firehose.**Mai****n

    Could not retrieve scm descriptor for initial ts id cache initialization


    On Wednesday, July 17, 2013 1:59:03 PM UTC+1, Jaykay wrote:

    Actually, the Alert Publisher does start ok.
    On Wednesday, July 17, 2013 1:48:47 PM UTC+1, Jaykay wrote:

    Having issues with my upgraded Cloudera Manager box. Since
    upgrading to 4.6.1, I can't start any of the monitor services, serivce
    monitor, activity monitor etc.
    I went through upgrading just the manager, not the nodes. I was
    sidetracked half way through the upgrade wizard, and it logged me out. When
    I logged in, it said it was complete, but I think there may have been a
    step missed form the upgrade process. The last step I recall was setting up
    the postgres embedded databases for the activity monitor etc. This was the
    point where it timed out and logged me out.
    Is there a way to go through the wizard again, or will this cause
    more issues?
    Here is the error when I try to start any of the services:

    1:41:33.838 PM INFO com.cloudera.cmon.firehose.**Mai******n

    Starting Firehose. Args: [--pipeline-type, ACTIVITY_MONITORING_TREE, --mgmt-home, /usr/share/cmf], Version: 4.5.0 (#243 built by jenkins on 20130226-1821 git: 0c84b050ec50fc9cde9625085f1607********7bce0790d4)


    Let me know if you need more entries from the logs.

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupscm-users @
categorieshadoop
postedJul 17, '13 at 12:48p
activeJul 31, '13 at 3:34p
posts11
users2
websitecloudera.com
irc#hadoop

2 users in discussion

Jaykay: 7 posts Chris Leroy: 4 posts

People

Translate

site design / logo © 2022 Grokbase