HostMonitor can't open dialogs, error in the event log

astewart · Post by **astewart** » Mon Jul 30, 2007 10:36 am

I encountered a bug this morning with HostMonitor. When I try to do any command that tries to open a dialog (add a test, modify a test, even open "About..." under the help menu) nothing happens and in the windows application event log an error similar to the following occured.

Event Type: Error
Event Source: hostmon.exe
Event Category: None
Event ID: 0
Date: 7/30/2007
Time: 9:31:37 AM
User: N/A
Computer: XXXX
Description:
The description for Event ID ( 0 ) in Source ( hostmon.exe ) cannot be found. The local computer may not have the necessary registry information or message DLL files to display messages from a remote computer. The following information is part of the event: Access violation at address 0042D47C in module 'hostmon.exe'. Read of address 3B8BF090.

I am using HostMonitor 6.80 running as a service on Windows 2000 SP4. Restarting the HostMonitor service has fixed the problem for now, but I thought I should report it anyhow.

-Alex

KS-Soft Europe · Post by **KS-Soft Europe** » Mon Jul 30, 2007 10:42 am

Thank you for your report.

Could you provide more information, please?
- Do you use ODBC logging or ODBC Query test method? If yes, what ODBC driver do you use?
- Do you have installed antivirus monitor? Personal firewall? Content monitoring software? Non-standard winsock components? Network packet analyzer?

It could be resource allocation error. Can you check resource usage for each process? You may use standard Windows Task Manager to check Handles, GDI and USER objects. What is the total resource usage on your system? How many Handles, GDI and USER objects are used by hostmon.exe process?

Regards,
Max

astewart · Post by **astewart** » Mon Jul 30, 2007 11:28 am

I do not use ODBC logging or the ODBC query test (although we have the Oracle client installed on the system, 10.2.0.3.0). We recently started using the "check MySQL server" test, and had to add the libmysql.dll to the server to get it to work. We are using the dll included with the 5.0.37 build of MySQL (I tried to use the DLL you recommended here: http://www.ks-soft.net/download/LIBMYSQL.DLL but it couldn't communicate with our installation).

Do you have installed antivirus monitor? No
Personal firewall? No
Content monitoring software? No
Non-standard winsock components? No
Network packet analyzer? No
Basically this machine is dedicated to running only HostMonitor, although it does have IIS 5.0 installed which is used to run ASP scripts whose output is checked with HostMonitor.

HostMonitor (after being restarted less than an hour ago)
# of tests: 1160
Mem Usage: 27Mb
Handles: 1,250
User Objects: 577
GDI Objects: 878
Threads: 22

System (uptime of 10 days)
Mem Usage: 205 Mb (512 Mb total on the system)
Handles: 10552
UserObjects: 1036
GDI Objects: 1573
Threads: 531
Processes: 42

Let me know if any other info would be helpful to you. I'll try to get the process/system data during the problem if the problem begins occuring again.

KS-Soft Europe · Post by **KS-Soft Europe** » Mon Jul 30, 2007 11:41 am

Could you send all *.hml, *.lst and *.ini files from HostMonitor's folder to support@ks-soft.net? Please, compress the files before sending, if possible.

Regards,
Max

astewart · Post by **astewart** » Mon Jul 30, 2007 11:52 am

Files sent. Subject is "HostMonitor config files as requested from bug thread 17881"

astewart · Post by **astewart** » Tue Jul 31, 2007 3:01 pm

After receiving your email I checked on our HostMonitor installation and it had entered the bad state again. Additionally, it seems when hostmonitor is in this state it can't send email notifications either (syslog is filled with "7/31/2007 12:02:27 AM SMTP client error: Access violation at address 0042D47C in module 'hostmon.exe'. Read of address 3B8BF090 ". The stats for HM and the system while HM is in a bad state:

HostMonitor (experiencing the problem, uptime just over a day)
Mem Usage: 13Mb
Handles: 3,300
User Objects: 348
GDI Objects: 297
Threads: 30

System (uptime of 11 days)
Mem Usage: 205 Mb (512 Mb total on the system)
Handles: 12500
UserObjects: 784
GDI Objects: 986
Threads: 533
Processes: 40

KS-Soft · Post by **KS-Soft** » Wed Aug 01, 2007 5:38 am

HostMonitor (experiencing the problem, uptime just over a day)
Mem Usage: 13Mb
Handles: 3,300
User Objects: 348
GDI Objects: 297
Threads: 30

Handles usage is a little above normal however its normal when HostMonitor performs a lot of tests. How many handles used by HostMonitor after 2-3 hours of work?
Other resources just fine.

System (uptime of 11 days)
Mem Usage: 205 Mb (512 Mb total on the system)
Handles: 12500

12,500 handles can be a problem for Windows 2000. Its better to find and terminate some unimportant applications

I encountered a bug this morning with HostMonitor.
...
We recently started using the "check MySQL server" test, and had to add the libmysql.dll to the server

Have you used version 6.80 for long time before this problem appeared? May be problem relates to libmysql.dll... Could you try to disable MySQL tests?

If you have used some old version then you updated to version 6.80 and problem happened within week or two, I would recommend to try this update http://www.ks-soft.net/download/hm680std.zip (this is version 6.80 compiled with old memory manager that was used since HM 0.10)

Regards
Alex

astewart · Post by **astewart** » Wed Aug 01, 2007 11:21 am

I'll try to get you a performance measurement after 3 hours of time next time we restart the HM service.

Unfortunately I don't have have many choices in reducing handle count, as I said the machine is a dedicated HostMonitor machine and has little else running. I'll check for a few extraneous services, but that is really all that is running.

We've been running HostMonitor 6.80 since May 21st. I don't believe we ever encounter this problem until recently (definately haven't been seeing the problem for months or I would have reverted back to 5.92, our previous version).

I'll try disabling just the MySQL tests and see what happens, and if it bugs again I'll try your exe. I am also running the 4 tests of HostMonitor Max requested: CPU, UNC, URL, and WMI. So far nothing has bugged out.

KS-Soft Europe · Post by **KS-Soft Europe** » Wed Aug 01, 2007 1:56 pm

astewart wrote:I'll try to get you a performance measurement after 3 hours of time next time we restart the HM service.

Could you enable "Counter Logs" for hostmon.exe process and send it to us after the crash? You may enable such log using "Performance" applet that is accesible from "Administrative Tools" windows menu.

astewart wrote:Unfortunately I don't have have many choices in reducing handle count, as I said the machine is a dedicated HostMonitor machine and has little else running. I'll check for a few extraneous services, but that is really all that is running.

Hm. So, we have the situation, when Hostmonitor uses 3,300 handles, but total handles usage is over 12,000. What exact application uses the most of handles?

astewart wrote:I am also running the 4 tests of HostMonitor Max requested: CPU, UNC, URL, and WMI. So far nothing has bugged out.

4 tests? I sent you the .zip file with 4 test lists with different test methods. Each tests list contains several hundred of the test. Could you verify that?

Regards,
Max

astewart · Post by **astewart** » Wed Aug 01, 2007 2:40 pm

I have enabled a counter log for each hostmonitor process (yesterday when I started the 4 test lists), the counter log is recording every stat under the "Process" performance object.

Highest handle usage in order: HostMonitor.exe, svchost.exe, services.exe, csrss.exe, lsass.exe, dllhost.exe, inetinfo.exe, svchost.exe, svchost.exe, etc. (all of the services just listed have 500+ handles). All just system processes, HostMonitor, and IIS5. One of our 2k machines with nothing running has 9000 handles, and our production IIS machines have 25,000. I don't think handles, at least at a system level, is really a factor here.

The 4 tests I referred to were each group of test methods which all have 500+ monitor tests in them, sorry if that was confusing.

astewart · Post by **astewart** » Thu Aug 02, 2007 7:13 pm

Our main HostMonitor service encounter the problem again (with the MySQL tests disabled). I noticed around 6pm, but I don't know exactly how long it was in a bad state but it couldn't have been more than a few hours. I have now switched to using the new exe you provided. None of the other hostmonitors have had a problem yet, but they are all still running.

I am sending you the perfmon log, but I didn't see anything that struck me as odd.

KS-Soft Europe · Post by **KS-Soft Europe** » Fri Aug 03, 2007 3:28 am

astewart wrote:I have now switched to using the new exe you provided.

Please, let us know how it works.

astewart wrote:None of the other hostmonitors have had a problem yet, but they are all still running.

Hm. Looks like we have to look for another culprit. I have sent you the lists with Ping, Perf Counter, Folder/File availability and Folder/File size tests. Could you test them in te same way? I think, you may cancel monitoring previous lists.

astewart wrote:I am sending you the perfmon log, but I didn't see anything that struck me as odd.

I agree with you that log is useless without the fact of the crash.

Regards,
Max

astewart · Post by **astewart** » Mon Aug 06, 2007 11:48 am

Just wanted to give an updated. I have been running the 4 new test lists (using the standard 6.8 build) and our normal HostMonitor (using the exe with the old memory manager) since Friday afternoon. No problems as of yet. I'll keep you posted.

KS-Soft · Post by **KS-Soft** » Mon Aug 06, 2007 3:18 pm

Thank you for the notice. Lets continue testing a little more...

Regards
Alex

astewart · Post by **astewart** » Mon Aug 13, 2007 11:38 am

Bad news

The HostMonitor using the new exe began having problems at around 8:27PM on Friday (PDT). None of the custom HostMonitor lists have experienced a problem. I am sending you the perf log of the process, but I checked it out myself and nothing weird happened appears to have happened around the time it started having problems.

I am sort of at a loss of what to do at this point.