Compatibility Problem between Citrix Memory Optimization and SAP GUI.

If you published your SAP GUI to the Citrix PS, than there is a chance you get an error message like this "CSapEditorCtrl::GetObject: Object x does not exist". This error caused by Incompatibility between the SAP GUI and Citrix Memory Optimization. To solve this error you must exclude the SAP GUI from Citrix Memory Optimization. To do that here the step
1. Deactivate Citrix Memory Optimization, to desctivate this feature, go to Citrix Management Console
Click Properties at your Farm Name, then Memory/CPU Utilization Management Disable the memory Optimization
2. Exclude the SAP GUI from Citrix Memory Optimization Process.
Open the register editor(regedit), see the "HKEY_LOCAL_MACHINE\SOFTWARE\Citrix\SFO\ProcessExclusionList", for example if yours
SAP GUI Located at "C:\Program Files\SAP\Front End\SAPGui", than add entry for DWORD which contain "C:\Program Files\SAP\Front End\SAPGui\*.*". Set the Value data to 0.
After that run the ADSCleaner.exe, this tools will list and delete the data streams that created by Citrix Memory Optimization.
After that restart this server this step only to make sure the changes that we made on the ProcessExclusionList registry. Than Activate Again the Citrix Memory Optimization.

Installing pnp ( PNP is not PerfParse ) at CentOs 5

Because nagios not come with tools to generate chart, so I need another tools to this job.
I found several tools for this task such as perfparse, NagiosGrapher, nagiosstat, and pnp.
After try to install some of this tools I choose to use pnp. I use pnp because it's easy to install.
To install pnp first read the documentation then proceed with the installation step

./configure --sysconfdir=/etc/nagios --datadir=/usr/share/nagios/pnp --bindir=/usr/bin --localstatedir=/var/nagios

you will see the configuration summary

make all && make install

Then you finish with the installation.

Now we gonna edit our configuration files.

For the first step edit the nagios configuration file, at my system it is the nagios.cfg, edit the following lines
Now edit the command definition file at my system it's called commands.cfg, edit the process-service-perfdata definition
at the command-line section set the value to /usr/bin/perl /usr/local/nagios/libexec/process_perfdata.pl.

here the process-service-perfdata section at my commands.cfg

According to pnp documentation process_perfdata.pl can not start under ePN(embeded Perl Nagois), that's why we must put the perl full path.

This type of configuration is default mode. There are others type of configuration mode, they are Bulk Mode and Bulk Mode with NPCD.

Create 2 other files serviceextinfo.cfg and hostextinfo.cfg. Both file name must match with the file name at Extended host/service info definitions section in the nagios configuration file (see above).

to create both files use this command "touch serviceextinfo.cfg hostextinfo.cfg"

Then edit both file.Here the content of my serviceextinfo.cfg (this will monitor the CPULoad at my windows machine)


The service_description must be the same at your service configuration file.

And for the hostextinfo,cfg

At my system I also edit the config.php, which the configuration file for pnp, I only change the temp directory to /tmp/nagios

$conf['temp'] = /tmp/nagios

Stop the nagios service, and then use the nagios command with -v parameter to check if there are no errors at your configuration file.
If everything is ok then star tthe nagios service then you are done.
To check if your installation success, see at the RRD direcotry whether the rrd and xml files generated on that place.
I got a problem with SELinux feature, that casue the pnp can not read it's RRD dabatabase directory. I still not figure out how to
fix this, so for now I disabled SELinux on my system.

Nagios monitoring tools...

Because the nature of my job as System Engineer, and also as Administrator for my division, which running 24x7 so I must know which computer that having sluggish performance, before the computer became useless in daily operational.
My division is a collaboration between Indonesia company and a Dutch based company..well they outsource some of their task to my division, including monitoring task, they use BMC Patrol for monitoring their customer environment.
BMC Patrol is a powerful tools,but is very expensive.Because I only need to monitoring my division environment, so I must use a cost free tools (because there is no way the management will permit me to buy a monitoring tools), and after googling i found a monitoring tools that called nagios.I like nagios because it a great tools with free of charge, and the most important it's open source.Nagios can monitor anything on your computer with the help of its plugin. There are lots of plugin for nagios, such as plugin to monitor mysql server, Citrix, Oracle, CPU, memory, etc. Well there are lots of plugin for nagios, so it's almost imposibble to mention then one by one. You can find nagios at their web www.nagios.org, there are documentation for that, and also you can use google to find help when you stuck with nagios.

But when you installing nagios from the source you must have lot of patient and read the documentation over and over again.
For me I use the rpm package from rpmforge, only need to run the yum command, then the nagios automatically installed on your system.
After you installed then you must configure the nagio configuration file, and what I like from linux is the configuration file use text, and not registry like the one you find a windows system.Because I use centos as the nagios server I use the guide from Max Hetrick at the centos wiki here the web http://wiki.centos.org/HowTos/Nagios. Because nagios not provide a graphic output so you need an add-on to generate chart, for this i use pnp(PNP is not PerfParse ), it's a great tools and quite easy to installed.

Because all the computer at my division ia a windows machine (except for the nagios server), i use the NSClient++ at the windows machine, and at the nagios server I use check_nt plugin. You can monitor all the counter you can find at windows performance counter. To see which counter available at your computer just use the perfmon snap-in. For start see this link http://nagios.sourceforge.net/docs/3_0/monitoring-windows.html.

Don't use Process Monitor at Windows 2000

Yesterday, I try ProcMon at one of my customer server, that server is a Windows 2000 SP4 with Update Rollup 1, so I think the ProcMon should be running smoothly on that server, but after I start that application and click the Agree button , the server restart it self... it seems the server got crash and it took more than 10 minutes so that server finish it boot process. Thank god I try that outside office hour, if i do that while the customer still work..you can guess what kind of email i will got here... There should be a warning at the Process Monitor web , that this tool can cause crash on windows 2000.... btw I use this tool because it is recommended by MS at their technet web.... well that a story of the day ,,,