Pricess check - TIMED OUT
Pricess check - TIMED OUT
Hi!
In my test list I have many tests according to the same pattern:
for each server 2 process check test, 1 ping test, 1 url test
I have about 45 folders representing servers organized this way.
Developing Host Monitor (adding new tests for new servers) everything was right, but when number of folders had reached about 35 Host Monitor has started having strange "waves" of timed out process check tests. Sometimes all folders are not marked in red (all tests have status OK) but several times a day comes "wave" that turns most folders in red (tests with status "timed out").
As I mensioned this corresponds only to process check tests.
I've tried to change values in
Options\Behavior\Don't start more than XX tests per seconds
also
Folder proporties\Specials\Non-simultaneously test execution
but wihout any results.
What can I do to get rid of it? Is there any possibility to change time out values for this kind of tests? Maybe I've made some mistake in global Host Monitor configuration.
HM: 5.86, 5,92, 6.00 beta (my problem occurs in all versions),
System: Windows 2000 Prof. PL + SP4
Regards,
Przemo
In my test list I have many tests according to the same pattern:
for each server 2 process check test, 1 ping test, 1 url test
I have about 45 folders representing servers organized this way.
Developing Host Monitor (adding new tests for new servers) everything was right, but when number of folders had reached about 35 Host Monitor has started having strange "waves" of timed out process check tests. Sometimes all folders are not marked in red (all tests have status OK) but several times a day comes "wave" that turns most folders in red (tests with status "timed out").
As I mensioned this corresponds only to process check tests.
I've tried to change values in
Options\Behavior\Don't start more than XX tests per seconds
also
Folder proporties\Specials\Non-simultaneously test execution
but wihout any results.
What can I do to get rid of it? Is there any possibility to change time out values for this kind of tests? Maybe I've made some mistake in global Host Monitor configuration.
HM: 5.86, 5,92, 6.00 beta (my problem occurs in all versions),
System: Windows 2000 Prof. PL + SP4
Regards,
Przemo
-
- Posts: 2832
- Joined: Tue May 16, 2006 4:41 am
- Contact:
-
- Posts: 2832
- Joined: Tue May 16, 2006 4:41 am
- Contact:
My configuration :
total CPU usage: 35-60% (Pentium4 2.8GHz , 1GB RAM)
hostmon.exe process:
CPU : 6-40% (top one)
Mem Usage : 25MB
VM size: 34 MB
Handles: 414 (more have about 10 processes)
GDI : 1887 (top one)
1127 test 10,2/sec
Now we are using external program to get all needed informations from Micorsof Servers via SNMP - no "time out" problems, hope my hardware/software configuration help solve this problem.
total CPU usage: 35-60% (Pentium4 2.8GHz , 1GB RAM)
hostmon.exe process:
CPU : 6-40% (top one)
Mem Usage : 25MB
VM size: 34 MB
Handles: 414 (more have about 10 processes)
GDI : 1887 (top one)
1127 test 10,2/sec
Now we are using external program to get all needed informations from Micorsof Servers via SNMP - no "time out" problems, hope my hardware/software configuration help solve this problem.
CPU usage: max 5%
memory available for system: 90 Mb
processes: 22
hostmon.exe:
memory used: 27 MB
GDI obj.: 164
handles 337
Generaly system doesn't seem to be overloaded.
HM runs as an application. There are no firewalls. All ports are opened (HM host all monitored servers are inside VPN network).
HM's "Estimate load" window:
total testes: 439 - 50,0/min
261 relations, no loops.
including:
process check 107 - 10,2/min
memory available for system: 90 Mb
processes: 22
hostmon.exe:
memory used: 27 MB
GDI obj.: 164
handles 337
Generaly system doesn't seem to be overloaded.
HM runs as an application. There are no firewalls. All ports are opened (HM host all monitored servers are inside VPN network).
HM's "Estimate load" window:
total testes: 439 - 50,0/min
261 relations, no loops.
including:
process check 107 - 10,2/min
That is impossibleI don't use any RMA (and I'd like not to).
Reply is: "Timed out".

If HostMonitor performs Process test directly (without using RMA), Reply field of the test may display number of instances or it is an empty string. It cannot show "Timed out"
So, probably you are using RMA?
Regards
Alex
For sure I don't use any Remote Monitoring Agents at all.
In every test's proporties window filed Test by has value Host Monitor.
Timed out reply is possible (rationally) because for process tests I can see status Checking.. for a long time. After that status turns in Unknown and reply in Timed out.
For me it's not a matter of network latency (WAN) becauce I use the same test for host placed locally (LAN) and results are the same.

In every test's proporties window filed Test by has value Host Monitor.
Timed out reply is possible (rationally) because for process tests I can see status Checking.. for a long time. After that status turns in Unknown and reply in Timed out.
For me it's not a matter of network latency (WAN) becauce I use the same test for host placed locally (LAN) and results are the same.
I was right, but I was wrong
Process test cannot show "Timed out" responce but there is special procedure that checks for "hung" test items. Ok, now I know what this means - Windows API function RegConnectRegistry does not respond within 15 minutes. Why? I don't know, will try to find answer in Microsoft manuals.
Regards
Alex

Process test cannot show "Timed out" responce but there is special procedure that checks for "hung" test items. Ok, now I know what this means - Windows API function RegConnectRegistry does not respond within 15 minutes. Why? I don't know, will try to find answer in Microsoft manuals.
Regards
Alex
I am afraid you will have the same problem even using HostMonitor version 1.00I have the same situation in all the versions (5.92 to 6.10)

Try to use WMI test to check processes, this test method uses different technology (not Performance Counters) and works much better.
Regards
Alex