DX Application Performance Management

Expand all | Collapse all

How to check for hung or corrupt EM's

Jump to Best Answer
  • 1.  How to check for hung or corrupt EM's

    Posted 11-04-2009 04:47 AM
    I am looking into using an external monitoring program to verify the health of an EM. I currently use a cron script and would like to know how other people do this.  I want to be able to determine 2 conditions:  1) That an EM is capable of allowing a login  2) That data can be retreived from an EM.  The current process I use through cron does the following:  A) Attempts to log in to an EM through a CLW.  B) CLW extracts data from the EM to a file  C) Logs out of the EM  D) Starts another script that looks to see if the original script is still running, indicating that the EM login has hung.  E) Checks the output of the data extract to verify that data can be extracted.  This is all easily doable through cron, but the problem come in when I have a third party program (Tivoli ITMA) that needs to check and see if the login has hung. Does anyone else do this and how do you do it?    Sample cron snippets:       cd /wily/moms/moma/lib
        /wily/moms/moma/jre/bin/java -Dhost=localhost -Dport=9999 -Duser=xx -Dpassword=yy -jar CLWorkstation.jar list agents matching \"Custom Metric Agent*\" > /logs/p
    rocesscheck/colla.status.log  This verifies the login and puts a data extract to a file    if [ `ps -aef|grep java|grep userxx|grep -v grep|wc -l` = 0 ]; then
                    status="One or more CLW scripts are HUNG on $CollHost."      
              fi  The above checks that the login hasn't hung and does a data extract.    Then tail the log file and look for "Exception" indicating that data cannot be extracted:  if [ `tail /logs/processcheck/colla.status.log|grep "Exception"|grep -cv grep` =
     1 ]; then

  • 2.  Re: How to check for hung or corrupt EM's

    Posted 11-05-2009 06:29 AM
    Hi Gerald,  I am checking with some folks here if they have some ideas.   I will let you know what I hear back.   Meanwhile maybe another user here will have something to offer.  Regards,Chris

  • 3.  Re: How to check for hung or corrupt EM's
    Best Answer

    Posted 11-05-2009 07:01 AM
    Have you considered using an EMEPAgent?   The EMEPAgent provides the following utilities:  ·                 EM log reader  ·                 EM GC verbose reader for Sun and IBM JVMs (might work for others)  ·                 Optional EM CPU Utilization if agent is on Solaris, NT, or AIX  .                 Optional EM process availability checker.  ·                 Optional EM machine disk utilities if EM is on Solaris or Windows.   May require separate Perl module installations.  ·                 Optional EM WebView HTTP availability.   May require separate Perl HTTP module installation.      

  • 4.  Re: How to check for hung or corrupt EM's

    Posted 11-05-2009 11:50 AM
    I think that I may have found my solution. The problem always is detecting the hung login, the rest is easy.  Since I now I need to check the EM from a remote machine, I'll just have to put a timer on the login script.

  • 5.  RE: Re: How to check for hung or corrupt EM's

    Posted 08-03-2010 12:35 PM
    Can you share the script.I am also looking for such type of script and helpful me in writing my script.

    Thanks in Advance.