NetXMS Support Forum

Hi!

Looks like a bug in firmware. I've seen somewhat similar problem with some HP ProCurve switches - after lot of SNMP requests they stop responding to SNMP for a few minutes. What you can do is to disable all poll types except status poll for that switch (in properties on "Polling" page). If this will not help, then you have no other option as to disable status poll as well (or you can disable SNMP and just use ping to see if device is still online). Unfortunately there are no option to set polling interval individually for device (but looks like it has to be implemented).

Best regards,
Victor

Hi!

It works exactly as you suppose - system tries all listed community strings. However, SNMP configuration for already added devices is checked only during configuration polls (once per hour by default), and new discovery cycle can take even more time, so newly added community strings may not be used immediately.

Best regards,
Victor

Filter should apply to both active and passive discovery. Can you please send me a screenshot of your discovery configuration?

Best regards,
Victor

Hi!

I'll take a look at this bug. However, moving database with MS SQL export/import should work. What exact error you got when you try to connect database on new server?

Best regards,
Victor

Hi!

It's definitely a bug in the server which leaves some nodes unattached to neither subnet nor container object. I have one fix possibly related to this in 1.2.2 - may be it helps. If not, I'll have a chance to connect IBM blade to NetXMS in a couple of weeks, then I'll be able to do necessary tests.

Best regards,
Victor

Quote from: lweidig on June 30, 2012, 11:37:28 PM
1. While most discovered devices are ending up under the correct subnets under the "Entire Network" section, we have a number that were just dumped into the root of the tree.  What causes this and how do we get them into the Entire Network tree.

Yes, it's known problem. It should be fixed in release 1.2.2.

Quote from: lweidig on June 30, 2012, 11:37:28 PM
2. Can you explain how address filters work?  We are hoping to filter out the addresses out of the network we have links to and/or routes but it seems to add them either way.

Address filters are inclusive, not exclusive - you can specify what address ranges/subnets to include. Also you should enable filtering in general by setting filter to "Automatically generated..." and check "Accept nodes within given range". If you want to exclude some addresses from discovery, you have to create custom filtering script.

Quote from: lweidig on June 30, 2012, 11:37:28 PM
3.  What is the best way suggested to apply templates and classify newly discovered nodes in the Infrastructure Services tree? 

You can place objects in Infrastructure Services tree in whatever order you want. It's hard to suggest anything, it usually depends on your organization structure, preferences, etc. You can setup rules for automatically placing nodes into containers - see "Automatic Bind Rules" tab in container's properties.

Quote from: lweidig on June 30, 2012, 11:37:28 PM
4.  Some discovered sections end up with a components section which can have a lot of additional information.  Is there a list somewhere of types of devices it "knows" about?  Can we add additional detection easily?

"Components" tab shown for any SNMP device which supports ENTITY MIB.

Quote from: lweidig on June 30, 2012, 11:37:28 PM
5.  Is there any updated documentation?  Seems to not be in sync with product we are using (1.2.1) and even what I have found seems to not list all of the functions available, discuss what subagents are, ...  Seems there may be a lot of hidden power waiting to be unleashed

We are now in progress to move existing documentation into wiki and completing it. I suggest to look into wiki: http://wiki.netxms.org.

Best regards,
Victor

Hi!

I'll consult with the person who knows more about Oracle then I do

Btw, is it possible to reference your organization on our users page (https://www.netxms.org/users/)?

Best regards,
Victor

В 1.2.2 получится думаю. Главное - не забыть

P.S. Изменения делаются в агенте.

Hi!

Current version of web interface connects to server on localhost, so if you install it on separate machine, it cannot connect. It is possible to make web UI to connect to remote NetXMS server, but you have to change property file packed deep inside war file:

1. find file WEB-INF/plugins/org.netxms.webui.core_1.2.1.jar inside .war file;
2. inside this jar, find file org/netxms/webui/core/nxmc.properties;
3. Change "server" option to IP address of your NetXMS server.

Deploy modified war file and web ui should be connecting to your NetXMS server.

Best regards,
Victor

Hi!

Log viewer return results in pages. Try to press navigation buttons on toolbar - it should show you next pages.

Best regards,
Victor

Can you show me please output of

ldd /usr/lib/libnxddr_mysql.so

?

Best regards,
Victor

Hi!

It's a known bug. You either have to install libssl-dev package and re-run configure, so encryption will be enabled, or patch session.cpp so line 6466 looks like:


   msg.SetId(request->GetId());


Best regards,
Victor

Hi!

You can use templates: http://wiki.netxms.org/wiki/UM:Data_Collection#Templates.

Best regards,
Victor

Hi!

You don't need to use situations here, and you didn't add actual check for service. You should go to data collection for node where you want to monitor spooler service, and add new parameter with source NetXMS agent and name System.ServiceState(Spooler). Then you should add threshold for that parameter (on Thresholds tab) with condition "last value not equal 0" and set activation event to SPOOLER_DOWN. Then it should work.

Best regards,
Victor

Hi!

You should specify database id when reading Oracle-specific parameters. In your config database id is DB1, so you should specify parameters as

Oracle.DBInfo.IsReachable(DB1)
Oracle.DBInfo.Name(DB1)

and so on. List of available parameters can be found here: http://wiki.netxms.org/wiki/Oracle_Monitoring_Subagent.

Best regards,
Victor