NetXMS Support Forum

Hi,

The patched nxdbmgr worked fine. Afterwords the core service starts and there is no error messages in the xmslog.txt log file.
Thank you for the fix!

The only thing is that as the update initially failed I reverted the live server to 1.1.6. I cloned the VM and did the test on that machine.
Will there be a full install package that I could apply to the live server?

Kind regards,

Dan

Hi,

Here's the result:

nxdbmgr.exe upgrade -t
NetXMS Database Manager Version 1.1.9

Upgrading database...
Upgrading from version 244 to 245
>>> ALTER TABLE nodes ADD runtime_flags integer
>>> UPDATE nodes SET runtime_flags=0
>>> UPDATE actions SET rcpt_addr='' WHERE CAST(rcpt_addr AS nvarchar(4000))=N'#0
0'
>>> SELECT action_id,rcpt_addr FROM actions WHERE rcpt_addr LIKE '%#%'
>>> UPDATE actions SET email_subject='' WHERE CAST(email_subject AS nvarchar(400
0))=N'#00'
>>> SELECT action_id,email_subject FROM actions WHERE email_subject LIKE '%#%'
>>> UPDATE actions SET email_subject='%m' WHERE action_id=1
>>> UPDATE actions SET email_subject='%m' WHERE action_id=2
>>> UPDATE actions SET email_subject='%m' WHERE action_id=3
>>> UPDATE actions SET email_subject='%m' WHERE action_id=5
>>> UPDATE actions SET email_subject='%m' WHERE action_id=6
>>> UPDATE actions SET email_subject='%m' WHERE action_id=7
>>> UPDATE actions SET action_data='' WHERE CAST(action_data AS nvarchar(4000))=
N'#00'
>>> SELECT action_id,action_data FROM actions WHERE action_data LIKE '%#%'
>>> UPDATE actions SET action_data='Event Name: %n\nIp: %a\nEvenID: %i\nTimestam
p: %t\nEvenCode: %c\nSeverity: %s\nSeverity text: %S\nNetXms server version: %v\
nUser tag: %u\nCustom message: %M\nCurrent value: %<currentValue>\nValue text: %
[ServiceStateName]\nThreshold: %<thresholdValue>\nIs repeat: %<isRepeatedEvent>'
WHERE action_id=1

Hi,

Trying to upgrade from Windows 2008 x64 NetXMS 1.1.6. with SqlExpress to 1.1.9 using u full install.
The programs stops with:

nxdbmgr.exe has stopped working

Problem signature:
  Problem Event Name:   APPCRASH
  Application Name:   nxdbmgr.exe
  Application Version:   0.0.0.0
  Application Timestamp:   4f10559e
  Fault Module Name:   ntdll.dll
  Fault Module Version:   6.0.6002.18541
  Fault Module Timestamp:   4ec3e855
  Exception Code:   c0000005
  Exception Offset:   0000000000049669
  OS Version:   6.0.6002.2.2.0.400.8
  Locale ID:   2067
  Additional Information 1:   153d
  Additional Information 2:   65c33816e286fcda867b11e451b1909a
  Additional Information 3:   b6d7
  Additional Information 4:   f5b2f9f9f15205c69cb45696238ca7a6

Read our privacy statement:
  http://go.microsoft.com/fwlink/?linkid=50163&clcid=0x0409


running nxdbmgr check manually gives:
Your database has format version 244, this tool is compiled for version 245. Use "upgrade" command to upgrade your database first.

running then the upgrades result in the same error as above.

Any idea of how to proceed?

Kind regards,

Dan

Hi Victor,

Your solution is elegant and works like a charm. Tanks!

Kind regards,

Dan

Current I report the current value of the DCI using the %<currentValue>
For the ServiceCheck.HTTP the value can be 0 to 3.
I would rather have a textual description instead of the numerical value. I thought of using a the transformation script but this will change the type from int to string which cannot be graphed.
Another place where I could do this is in the filtering script where CUSTOM_MESSAGE can be set.

sub main()
{
CUSTOM_MESSAGE = "";

if($1->value == 2)
{
   CUSTOM_MESSAGE = "cannot connect to server";
}
else
if($1->value == 3)
{
   CUSTOM_MESSAGE = "content of does not match";
}

  return true;
}

How do I test the "value" within the filtering script ($1->value above does not work of course) ?
Or maybe there's a better solution?

Regards,

Dan

Hi Victor,

Thanks for the prompt reply.

I changed the following:
-   server configuration variable AgentCommandTimeout set to 5000;
-   added the following to nxagentd.conf:
        *portCheck
        Timeout = 4000
The above solved the issue – thank you. So no bug, but is this by design? My settings before the change (I believe them to be the default) where 2 seconds on the server and 30 second on the portCheck agent. Do these not cause issues for other users?

Kind regards,
Dan

Hi

Our Linux hosting server hanged recently (extremely rare), however I did not receive any alarms for the ServiceCheck.HTTP that was configured for it.

Today I looked further into it... I've set a ServiceCheck.HTTP to monitor a web site (apache) on my machine. When I stop the web daemon, the DCI threshold triggers as expected. However, if I disconnect the server from the network no triggering occurs. Is this by design or is this a bug?

Regards,

Dan

Hi Victor,

Thanks for the example. Although it surly works I find it to be too error-prone, at least for me it is :-)
I'm glad I'm not misusing anything, so I'll stick to my script ;-)

Regards,

Dan

Hi Victor,
I'm afraid it doesn't solve my problem. Having a one shot to test the website is not good enough, I need to test that page every minute starting on a specific date. To my best knowledge cron does not offer this functionality.

I ended up adding a DCI Transformation Script to filter the event for a given date range:

sub main()
{
if(
        localtime(time())->year == 2012
       
     && localtime(time())->mon >= 0
     && localtime(time())->mday >= 1
     
     && localtime(time())->mon <= 0
     && localtime(time())->mday =< 15
  )
  {
   return $1;
  }

  return 0;
}

In this example for all tests within 1st and 15th of January 2012 the real result is returned ($1).
For anything outside that period, success (=0) is always returned disregarding the actual result.
Alternatively I could have added something similar to the Filtering Script of the Event Processing Policy, however that would require me to create a new Event and Processing Policy for each date rage I want to test on. Using the Transformation Script  for Filtering may be misusing it, but it does the job. I guess that adding a Filtering Script feature before the Transformation Script is an overkill.

Regards,

Dan

Hi Victor,

The content of this website changes at a specific date, I want the monitoring system to verify that the new content shows at that time.
If the test runs before the due date it will fail, so I need a way to start monitoring at a specific date/time.

Regards,

Dan

Hi,

I'm trying to set a website monitoring where the content changes at a specific date/time.
Otherwise said, I need to stop monitoring the current content at some moment and the start monitoring other content from that moment forwards.

What would be the most efficient way to achieve this?

I was thinking of creating two DCI's and set them when they should start/stop monitoring, can the cron syntax of the custom Schedule property of a DCI handle this?
Otherwise scripting will surly get things done, but is it really needed?

Regards,

Dan

Hi,

In the Event Browser I Find SYS_AGENT_UNREACHEABLE for all servers that were shut downed.
Once the servers were restarted another SYS_AGENT_UNREACHEABLE was received, followed by SYS_AGENT_OK for the servers without issues.
I'm afraid the events on the Agent servers were overwritten since 22 November, as fas as I can remember there were no event to be seen.
For some reason the agent starts but cannot be detected after reboot by the server. Restarting the agent service later on solves the problem. Strange.

Regards,

Dan

PS. I had no rule, just added one ;-)

Hi lindeamon,

Thanks for the tips. The trap OID's returned by following your advice were identical to those shown in the Event Browser.
Digging further on the net for the .4.11.93.0.99 trap I've discovered "Time Server Monitor" from Meinberg is probably the culprit as someone else using a different monitoring software was getting the same wrong OID's. I guess 'Time Server Monitors" emits different traps than their own MIB file states.
Since NetXMS seems to accept the wrong OID and triggers on it I'll get used to live with it 

Regards,

Dan

Hi,

After adding a new MIB using nxmibc  -d "c:\Program Files\NetXMS\var\mibs" -z
I moved netxms.mib to the mibs folder and restarted the core service.
I proceeded to add a new trap mapping .1.3.6.1.4.1.5597.50.1 (I could select it with the MIB browser which I guess proves the above steps succeeded)
Finally a went on and added an event template and Event Processing Policy.

When I trigger the SNMP trap I receive the following (error) in the Event Browser:
SNMP_UNMATCHED_TRAP SNMP trap received: .4.11.93.0.99 (parameters: 1.3.6.1.4.1.5597.50.1.99 == ...)
I get the same message even when I map to 1.3.6.1.4.1.5597.50.1.99
So apparently the trap Is perceived as a parameter, what am I doing wrong?

Dan

Hi,

Due to electricity work on the premises I had to shut down a couple of servers for the night.
After rebooting  the servers the Status DCI of these servers (Win 2003 and n2003 R2) became green, however no Agent DCI's readings were returned  

A status poll gave the following error:
[22/11/11] Poll request accepted
[22/11/11] Starting status poll for node WebToPrint1
[22/11/11] Checking NetXMS agent connectivity
[22/11/11] NetXMS agent unreachable
[22/11/11]       Current interface status is NORMAL
[22/11/11]    Starting status poll on interface Lan1
[22/11/11]       Starting ICMP ping
[22/11/11]       Interface is NORMAL for 67 polls (1 poll required for status change)
[22/11/11]       Interface status after poll is NORMAL
[22/11/11]    Finished status poll on interface Lan1
[22/11/11] Node is connected
[22/11/11] Finished status poll for node WebToPrint1
[22/11/11] Node status after poll is NORMAL
[22/11/11] **** Poll completed successfully ****

The event log on the Agent servers did not show any error. Restarting the agent service solved the problem.

• Is this an bug that NetXMS Agents do not start correctly upon server restart?
  
• Is it expected behavior for the NetXMS server not to report any problem if an Agent stops reporting while the Status is green, or is a separate Agent test needed?
  

Regards,

Dan