Details

    • Type: Sub-task Sub-task
    • Status: Resolved
    • Priority: Major Major
    • Resolution: Won't Fix
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: scripts
    • Labels:
      None

      Description

      Finding the accumulo processes on a machine is a bit annoying because:
      1) All processes are called 'Main' in jps (other systems do this as well)
      2) long jvm-related argument lists make it difficult to find which pids are for particular daemons like the tserver, randomwalkers, etc.

      This patch adds a script to provide a concise listing suitable other scripting.

        Activity

        Hide
        Jonathan Hsieh added a comment -

        Attached patch applies to 1.4, 1.5 and trunk branches.

        Show
        Jonathan Hsieh added a comment - Attached patch applies to 1.4, 1.5 and trunk branches.
        Hide
        Josh Elser added a comment -

        Cool, thanks for the patch, Jonathan. Two questions:

        1) What does this bring to the table that `jps -m` doesn't?
        2) Might you also want to grep out the ssh command in case someone is using this with pssh or pdsh?

        Show
        Josh Elser added a comment - Cool, thanks for the patch, Jonathan. Two questions: 1) What does this bring to the table that `jps -m` doesn't? 2) Might you also want to grep out the ssh command in case someone is using this with pssh or pdsh?
        Hide
        Jonathan Hsieh added a comment -

        1) I wasn't aware of jps's -m option (or the other options) and this was useful where I've used accumulo on a cluster managed by Cloudera Manager. CM is guilty of launching processes called Main. ex:

        # jps -m
        16828 NameNode
        2166 HeadlampServer --db-conf-dir /var/run/cloudera-scm-agent/process/303-cloudera-mgmt-REPORTSMANAGER --db-conf-file headlamp.db.properties --mgmt-home /usr/share/cmf
        19939 war --httpPort=58080
        2067 NavigatorMain --conf-dir /var/run/cloudera-scm-agent/process/305-cloudera-mgmt-NAVIGATOR
        29923 Main
        26980 Main monitor --address jon-ahb2-1.ent.cloudera.com
        7560 DFSZKFailoverController
        2044 Main --pipeline-type ACTIVITY_MONITORING_TREE --mgmt-home /usr/share/cmf
        26888 Main gc --address jon-ahb2-1.ent.cloudera.com
        2139 AlertPublisher
        26721 Main master --address <node1>
        6993 QuorumPeerMain /var/run/cloudera-scm-agent/process/393-zookeeper-server/zoo.cfg
        27099 Main tracer --address <node1>
        31508 Jps -m
        2020 EventCatcherService
        2113 Main --pipeline-type SERVICE_MONITORING --mgmt-home /usr/share/cmf
        634 JobTracker
        2088 Main --pipeline-type HOST_MONITORING --mgmt-home /usr/share/cmf
        7610 JournalNode
        

        That said the longer jps -ml | grep accumulo works pretty well.

        # ajps
        26721 master --address <node1>
        26888 gc --address <node1>
        26980 monitor --address <node1>
        27099 tracer --address <node1>
        
        # jps -lm | grep accumulo
        26980 org.apache.accumulo.start.Main monitor --address <node1>
        26888 org.apache.accumulo.start.Main gc --address <node1>
        26721 org.apache.accumulo.start.Main master --address <node1>
        27099 org.apache.accumulo.start.Main tracer --address <node1>
        

        2) I've only used pssh with the accumulo code base and haven't used pdsh. (I use cssh for my small test clusters). I'll give pdsh a try and update it if you think this is worth pursuing further.

        Show
        Jonathan Hsieh added a comment - 1) I wasn't aware of jps's -m option (or the other options) and this was useful where I've used accumulo on a cluster managed by Cloudera Manager. CM is guilty of launching processes called Main. ex: # jps -m 16828 NameNode 2166 HeadlampServer --db-conf-dir / var /run/cloudera-scm-agent/process/303-cloudera-mgmt-REPORTSMANAGER --db-conf-file headlamp.db.properties --mgmt-home /usr/share/cmf 19939 war --httpPort=58080 2067 NavigatorMain --conf-dir / var /run/cloudera-scm-agent/process/305-cloudera-mgmt-NAVIGATOR 29923 Main 26980 Main monitor --address jon-ahb2-1.ent.cloudera.com 7560 DFSZKFailoverController 2044 Main --pipeline-type ACTIVITY_MONITORING_TREE --mgmt-home /usr/share/cmf 26888 Main gc --address jon-ahb2-1.ent.cloudera.com 2139 AlertPublisher 26721 Main master --address <node1> 6993 QuorumPeerMain / var /run/cloudera-scm-agent/process/393-zookeeper-server/zoo.cfg 27099 Main tracer --address <node1> 31508 Jps -m 2020 EventCatcherService 2113 Main --pipeline-type SERVICE_MONITORING --mgmt-home /usr/share/cmf 634 JobTracker 2088 Main --pipeline-type HOST_MONITORING --mgmt-home /usr/share/cmf 7610 JournalNode That said the longer jps -ml | grep accumulo works pretty well. # ajps 26721 master --address <node1> 26888 gc --address <node1> 26980 monitor --address <node1> 27099 tracer --address <node1> # jps -lm | grep accumulo 26980 org.apache.accumulo.start.Main monitor --address <node1> 26888 org.apache.accumulo.start.Main gc --address <node1> 26721 org.apache.accumulo.start.Main master --address <node1> 27099 org.apache.accumulo.start.Main tracer --address <node1> 2) I've only used pssh with the accumulo code base and haven't used pdsh. (I use cssh for my small test clusters). I'll give pdsh a try and update it if you think this is worth pursuing further.
        Hide
        Christopher Tubbs added a comment -

        I just do:

        alias accps='jps -lm | grep \ org\\.apache\\.accumulo'
        Show
        Christopher Tubbs added a comment - I just do: alias accps='jps -lm | grep \ org\\.apache\\.accumulo'
        Hide
        Jonathan Hsieh added a comment -

        Instead of a script, maybe it would be good to at least include the one-liner in the documentation?

        Show
        Jonathan Hsieh added a comment - Instead of a script, maybe it would be good to at least include the one-liner in the documentation?
        Hide
        Josh Elser added a comment -

        Instead of a script, maybe it would be good to at least include the one-liner in the documentation?

        Sounds good to me! Not sure if we have a great place for one-liner, ease-of-use things like this yet, but having a centralized place we can point people to sounds very useful. It will hopefully preserve the knowledge instead of it being one-off responses picked up by Google scraping mail-archives.a.o

        Show
        Josh Elser added a comment - Instead of a script, maybe it would be good to at least include the one-liner in the documentation? Sounds good to me! Not sure if we have a great place for one-liner, ease-of-use things like this yet, but having a centralized place we can point people to sounds very useful. It will hopefully preserve the knowledge instead of it being one-off responses picked up by Google scraping mail-archives.a.o
        Hide
        Jonathan Hsieh added a comment -

        HBase has a troubleshooting section in its manual – maybe something like that for accumulo? http://hbase.apache.org/book.html#trouble

        Show
        Jonathan Hsieh added a comment - HBase has a troubleshooting section in its manual – maybe something like that for accumulo? http://hbase.apache.org/book.html#trouble
        Hide
        Eric Newton added a comment -

        That's a good idea. There are a lot of utilities that could be explained there. I created ACCUMULO-1562.

        Show
        Eric Newton added a comment - That's a good idea. There are a lot of utilities that could be explained there. I created ACCUMULO-1562 .
        Hide
        Jonathan Hsieh added a comment -

        Will be documented instead in ACCUMULO-1562

        Show
        Jonathan Hsieh added a comment - Will be documented instead in ACCUMULO-1562

          People

          • Assignee:
            Unassigned
            Reporter:
            Jonathan Hsieh
          • Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development