Internal documentation only
This page has been marked as draft.
Running multiple instances of Windows netprobe using a single binary folder
It is possible to install a Netprobe on a shared drive and then run it on multiple nodes. However, there are a couple of things to be aware of if a Netprobe installation on a shared drive is going to be used by multiple nodes simultaneously:
-
If the Netprobes record their log output to file and they are writing their logs to the shared drive, then it is necessary to make sure that they each one is writing to a unique file. The Netprobes may crash if more than one Netprobe attempts to write to the same file. This is a problem that can be encountered, for example, if a shared script is being used to set up the environment and launch the Netprobe on a set of UNIX machines, and this script is setting LOG_FILENAME to the same file on the shared drive. In this example, one can work around the problem by changing the script to set it to something like
setenv LOG_FILENAME netprobe.out.`hostname` -
If the Message Tracker plugin is being used, this writes a file in the Netprobe directory which ensures that no data is lost if the probe stops. If two Netprobe instances are using the same shared drive installation and both using Message Tracker plugins, there is the potential for these both to using the same name for this file and therefore overwrite each other?s data. By default this file is called
. .cache, but the name can be set by using the Message Tracker configuration setting persistenceFile. If multiple Netprobe instances are being run which use the same shared drive installation, more than one of which is running a Message Tracker sampler, it should therefore be ensured that each Message Tracker sampler has the persistenceFile setting defined and set to a name unique to that Netprobe and sampler.