Discover hosts by NT Service

hamoja · Post by **hamoja** » Wed Aug 11, 2004 6:46 pm

Is there a way to create multiple NetBios/UNC tests for serveral servers for a specific service running on NT rather than individually?

Example, If I have 200 servers running antivirus service I would like to have that service monitored and restarted upon a failure. Presently the service crashes whenever the Windows service parameter is used to restart the service upon first or subsequent failures. On the other hand, Ksoft restarts it just fine.

KS-Soft · Post by **KS-Soft** » Thu Aug 12, 2004 11:41 am

Not sure how you can monitor a service using "UNC" test. Probably you are talking about "Service" test?
If you need to create many similar tests, use Replicator: http://www.ks-soft.net/hostmon.eng/replicator/index.htm

Regards
Alex

hamoja · Post by **hamoja** » Fri Aug 13, 2004 6:06 am

Perfect for creating similar tests which are UNC based. Questions:

1. How can I create similar alerts for each test using replicator?

2. Why some of my 200 newly created tests status are unknown? if they are tested individually, status reply OK

KS-Soft · Post by **KS-Soft** » Fri Aug 13, 2004 11:19 am

2)
Could you please provide more information:
- What test method do you use? UNC or Service?
- HostMonitor version?
- Windows version?

Regards
Alex

KS-Soft · Post by **KS-Soft** » Fri Aug 13, 2004 11:22 am

1) How can I create similar alerts for each test using replicator?

I think you don't need replicator at all. You may use the same alert profile for all 200 tests bacause you may use macro variables: http://www.ks-soft.net/hostmon.eng/mfra ... .htm#macro
Variables represent parameters of the test that had triggered an action.

Regards
Alex

hamoja · Post by **hamoja** » Fri Aug 13, 2004 1:33 pm

I am using a service test although it appears that the tests connects to logonservers using UNC ('\\servername")..

I just upgraded to version 4.6 and am still seeing unknowns on a bunch of servers. Like I mentioned before, if unknown tests are ran individually, status changes for each to OK.

Is this a bug?

KS-Soft · Post by **KS-Soft** » Sun Aug 15, 2004 5:58 pm

I am using a service test although it appears that the tests connects to logonservers using UNC ('\\servername")..

Yes, of course.

I just upgraded to version 4.6 and am still seeing unknowns on a bunch of servers. Like I mentioned before, if unknown tests are ran individually, status changes for each to OK.
Is this a bug?

Its not HostMonitor's bug. May be timeout issue if you have highly loaded network or you monitor remote network using dial-up connection...
- what Windows do you have installed on local/remote systems?
- Service Pack?
- do you monitor local or remote network?
- as I understand you monitor a few services on many systems?

Regards
Alex

hamoja · Post by **hamoja** » Mon Aug 16, 2004 5:58 am

hamoja · Post by **hamoja** » Mon Aug 16, 2004 6:28 am

My network communication to core is switched 100MB full duplex. Program runs on a 2Ghz P4 w/512MB of RAM system on Windows 2000 SP4

- what Windows do you have installed on local/remote systems?
- Service Pack?
- do you monitor local or remote network?
- as I understand you monitor a few services on many systems?".

- Windows NT4/2000 /SP4 + latest hotfixes
- Both - Mostly local
- Yes, currently 80 systems being monitored for that particular service. I have at least 270 various tests so far

Also, which macro variable do I need to use for my action parameters "Restart service on" for each system service failure

KS-Soft · Post by **KS-Soft** » Mon Aug 16, 2004 11:40 am

Also, which macro variable do I need to use for my action parameters "Restart service on" for each system service failure

%ServiceComp% The name of the target computer
%ServiceName% The name of the service

- Windows NT4/2000 /SP4 + latest hotfixes

Problem relates to NT4 and 2000 systems? What system is installed on Domain Controller? May be some problems with authorization...

Regards
Alex

hamoja · Post by **hamoja** » Tue Aug 17, 2004 5:49 am

My enviroment is Windows 2000 w/AD. If there was a authentication or authorization problems, why is it that when each test is ran individually the results are anything other than unknown?

This is very puzzling.

KS-Soft · Post by **KS-Soft** » Tue Aug 17, 2004 12:13 pm

My enviroment is Windows 2000 w/AD. If there was a authentication or authorization problems, why is it that when each test is ran individually the results are anything other than unknown?

I don't know. Probably 50% of our customers use Service test and they do not have such problem. We did not experience such behavior either.
Windows Service Control Manager can process many simultaneous requests. May be you have some antivirus software that drops connections when system receives many requests at the same time? Any other non standard software?

Regards
Alex

hamoja · Post by **hamoja** » Tue Aug 17, 2004 4:03 pm

We did not experience such behavior either.

How many service tests are you running of the same type to validate that statement?. Simliar or more than?. I have created 300 more TCP/IP tests today without seeing these types of issues. Apparently the problem is in the service tests, also I need to create an additional 400 more tests for the same type of service.

Are you saying that the problem exists elsewhere besides the hostmonitor application?. My server meets more than the standard requirement for KS monitor and antivirus is not hindering existing TCP tests created today, so then what other suggestions you have?

Marcus · Post by **Marcus** » Wed Aug 18, 2004 9:19 am

My enviroment is Windows 2000 w/AD. If there was a authentication or authorization problems, why is it that when each test is ran individually the results are anything other than unknown?

Have you looked at the security logs of your target servers. Most likely you are using two different accounts. Unless you mean that selecting more than 1 test to perform has a different result than running eacht test one by one. In that case I have no clue

Unknown simply means it can not query the service for a status. So it must be an authentication / authorisation problem.

How many service tests are you running of the same type to validate that statement?.

We are running thousands of service tests (on hundreds of servers). No problems, as long as the account we use for HostMonitor (we are running HostMonitor as a service) is a local acount on the server (domain validation does not work on non domain controllers).

have created 300 more TCP/IP tests today without seeing these types of issues. Apparently the problem is in the service tests

This sounds like an authentication problem again. To be able to run the service test, you must have a valid account on the target server. TCP/IP tests don't know users......

hamoja · Post by **hamoja** » Wed Aug 18, 2004 6:07 pm

Have you looked at the security logs of your target servers. Most likely you are using two different accounts. Unless you mean that selecting more than 1 test to perform has a different result than running eacht test one by one. In that case I have no clue

All servers are member servers of a single domain. I am using a single administrative domian account for all tests, hostmonitor runs all test concurrently which results in "unknowns" for a handfull of servers that are members of the same domain and using the same NT account which is not specified per test. I haven't yet added the soon to be others providing that this application functions.

If these tests are ran one at a time, the results are OK, normally a problem with authentication from the same account usually follows by a failure in all tests, including ones indivdually ran. So the account is obviously not locked as the other tests are successful.

I would like to know who else is running over 400 service tests without these type issues and if possible post their server configuration. You claim hostmontor can monitor over 800 service tests but I am yet to see a successful post on this matter.