FAQ
We had to reconfigure and reshoot all of our data nodes in a running
Cloudera CDH4 cluster (after stopping services of course) due to a hardware
setup problem. After the data node reconfiguration, it was not obvious how
to reinstall all of the Cloudera services on the newly reshot nodes.

We tried deleting and re-adding the cluster. Management services seem fine,
and our closer head node and name node are both still running (only the
data nodes were reshot). The Cloudera manager option to use parcels seems
to start OK, but each line in the status bar graph under the line "The
selected parcels are being downloaded and installed on all the hosts in the
cluster" seems just to cycle from 0% complete up to some number (typically
6% for Impala, up to 18% for Solr, about 9% for CDH) and then drop back to
0%.

What is going on in this display? Is it cycling through all of the nodes in
the cluster and showing us the installation progress on each node, one at a
time? Or is this cycling behavior indicating some failure and restart of
the download process?

We tried navigating to the "Parcels" field of the "hosts" display but it
just says "no cluster added", which apparently is the state until we can
add a cluster successfully.

The general question here is "what steps are necessary to recommission
Cloudera services on a cluster after replacing all of its data nodes"?
Secondarily, how can we add a cluster via the parcel method and what does
this odd looping behavior in the status progress bars mean?

To unsubscribe from this group and stop receiving emails from it, send an email to scm-users+unsubscribe@cloudera.org.

Search Discussions

  • Alan Sill at Jan 30, 2014 at 6:15 pm
    Answering my own question for the record:

    The cause of this was insufficient space on the disk partition used by the
    Cloudera SCM server. Clearing out space and stopping and starting the cloud
    era-scm-server process allowed restarting the downloads.

    On Thursday, January 30, 2014 11:35:49 AM UTC-6, Alan Sill wrote:

    We had to reconfigure and reshoot all of our data nodes in a running
    Cloudera CDH4 cluster (after stopping services of course) due to a hardware
    setup problem. After the data node reconfiguration, it was not obvious how
    to reinstall all of the Cloudera services on the newly reshot nodes.

    We tried deleting and re-adding the cluster. Management services seem
    fine, and our closer head node and name node are both still running (only
    the data nodes were reshot). The Cloudera manager option to use parcels
    seems to start OK, but each line in the status bar graph under the line "The
    selected parcels are being downloaded and installed on all the hosts in the
    cluster" seems just to cycle from 0% complete up to some number (typically
    6% for Impala, up to 18% for Solr, about 9% for CDH) and then drop back to
    0%.

    What is going on in this display? Is it cycling through all of the nodes
    in the cluster and showing us the installation progress on each node, one
    at a time? Or is this cycling behavior indicating some failure and restart
    of the download process?

    We tried navigating to the "Parcels" field of the "hosts" display but it
    just says "no cluster added", which apparently is the state until we can
    add a cluster successfully.

    The general question here is "what steps are necessary to recommission
    Cloudera services on a cluster after replacing all of its data nodes"?
    Secondarily, how can we add a cluster via the parcel method and what does
    this odd looping behavior in the status progress bars mean?
    To unsubscribe from this group and stop receiving emails from it, send an email to scm-users+unsubscribe@cloudera.org.

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupscm-users @
categorieshadoop
postedJan 30, '14 at 5:35p
activeJan 30, '14 at 6:15p
posts2
users1
websitecloudera.com
irc#hadoop

1 user in discussion

Alan Sill: 2 posts

People

Translate

site design / logo © 2022 Grokbase