service_check.exe

I noticed poor performance on my box and checks that said they were down when they werent... looking through the process monitor I see there are 107 instances of service_check.exe running.

I am runnng in service mode. MX2 Enterprise.

Thanks

What version are you running 5.2.0?



The instances indicate an issue in terms of the check not being performed correctly.



In order to trap the error you need to run it the monitoring_service in interactive mode so that all output of the service is sent to the console.



When you have done it then you will probably see that it sends an error dialog box. Copy the message of that dialog box along with the line (wipe any username/passwords) of the serverscheck.conf file and post it as a reply

I have issues where checks on some of my servers fail but they are actually up... I cant force it to create all these instances it just does it from time to time.. also did you know the process id is service_check.e not .exe.. just an FYI.

I am running 5.2.0 MS Enterprise.

Thanks

Running console.. I get an error that there are too many monitor managers runnng and it says it is killing one.. but it looks like the service_check.exe just gets stuck...

SERVICES|X|TRINITY MSSQL Service|X|3|X|180|X||X||X||X|true|X||X||X||X||X||X||X|TRINITY<X>MSSQLSERVER<X>firstfleetadministrator<X>PASSWORD|X||X|98|X|no|X|180|X||X||X|On each status change|X|SQL Alert Team|X||X||X|Services Group|X|
SERVICES|X|GATES MSSQL Service|X|3|X|180|X||X||X||X|true|X||X||X||X||X||X||X|GATES<X>MSSQLSERVER<X>firstfleetadministrator<X>PASSWORD|X||X|98|X|yes|X|180|X||X||X|On each status change|X|SQL Alert Team|X||X||X|Services Group

I am not running 5.2.1 same issue.

When you get the warning that too many monitor_managers are running then this means that you are not running it in interactive mode and that there still is a process running in the background.



You need to stop ServersCheck first. Then either set the service settings to run in interactive mode (see help file) or double click on the monitoring_manager.



Here is a section in the Help File describing what you need to do:

http://www.serverscheck.com/help5/debugginginteractive.html





I assume that you have set your ServersCheck service to run under a Windows Administrator account when performing the service and other checks?

Right... this is what I did.. I stopped the service.. then I went into task manager and killed all the exes that were running.. your exes of course.

Then when I run monitor manager I get this warning... I do not have the service running. Yes the service is running under a domain admin account.

I followed your instructions on that page and my way... which is to stop the service.. kill the left over exes and then start monitor_manager..

I still have the error and when I add service checks to my check base they just pile up in the task manager.

When starting the monitoring_manager.exe from the command prompt (c:...) you should at least get some kind of an output. Please copy that output in a reply.

Ok I have a new box at the advice of Maarten.

I have clean install.. I am running monitor_manager.

I have about 40 ping checks that fail right now.. it seems that the check runs and then sits there waiting on the fail checks.. does it not run these things parallel?

So if it runs a ping check and it its first is down does it sit there until the down checks run on that one process?

Yes that is correct. It executes the down checks on that single process (not on the other processes).

That single process are you telling me that it is like this because I am using monitor manager or is the service based method too?

No the monitoring_manager dispatches the jobs to the monitoring_rule.exe (see the architecture diagram in the help file).



One you have the Professional version then you only have one instance of the monitoring_rule.exe running. With the Enterprise Edition you have multiple depending on your license.



This means that the monitoring_rule instance performing the check that is down is performing all trying before making releasing its job. Two solutions: reduce the number of retrials and the interval "when down" checking

My browser shows MX2 Enterprise

This means that you should have multiple instances of the monitoring_rule.exe running