NetXMS Support Forum

Hello!

Having NetXMS agents on monitored systems is recommended, but not required. If you only interested to now if particular host is alive or not (responds to ping) and/or that particular services (like ssh or http) are alive, you don't need agent on that host. But if you wish to get more information (like CPU usage or status of particular process) you will need an agent on monitored system (you may also use SNMP agent and collect data via SNMP).

Best regards,
Victor

Hello!

All timestamps used in NetXMS internally stored in UTC and converted to local clock when shown. As far as UTC times on client machine and server are identical everything should be ok. If you don't see graphs correctly, possibly UTC time on your server or client is incorrect. One example of possible incorrect time settings:

1. My timezone is EET, which is GMT + 2 hours.
2. On my workstation, I set timezone to EET, but on server in the same location - GMT.
3. On both workstation and server I set time 14:00. Looks like both shows time correctly, but internally workstation has UTC time 12:00 (because of timezone settings), but server has UTC time 14:00. If I will connect with NetXMS console to server I'll see unexpected timestamps.

Best solution is to use NTP for time synchronization (it will always set your machine clock to correct UTC time) on both server and client.

Best regards,
Victor

Hello!

This is an expected behavior. Main idea is that all IDs should be unique,  so system just keeps first unused id and increments it after every creation. There are two reasons why we don't reuse freed IDs:
1. If, due to error, reference to deleted object will remain somewhere, it will later be threated as an error because ID is invalid. But, if we reuse the id, later it will be impossible to determine if this was undeleted reference to old object or valid reference to new one.
2. Optimization of code - server don't need to scan for free ID - not a big issue however.

It should not be a problem, because we have 32-bit IDs, so we have more than 4 billions of available IDs. I hardly imagine monitoring system where such number of id's can be exhausted in reasonable period of time. Even if system's administrator will create and delete new trap record every second, it takes more than 126 years to exhaust ID space. Also, if such problem ever arises, it will not be hard to write special database maintenance tool which will "compact" id space.

Best regards,
Victor

Hello!

Ensure that winperf subagent is loading. You should have line in your config like this:

SubAgent = winperf.nsm

You can check list of currently loaded subagents by selecting node object in console, then Tools->Info->Agent->Subagent List.

Best regards,
Victor

Hello!

I hope that I fix all nxhttpd build problems now. New cvs snapshot will be created this night (21 Jun 4:15 EET) and should include all changes. You can try to build nxhttpd from this snapshot.

Best regards,
Victor

Hi!

It's a bug in configure, but after looking at config.log seems that workaround for you is different: try to locate libgd.so and add containing directory to /etc/ld.so.conf and then run ldconfig. For example, if libgd.so located in /usr/local/lib, just add a line /usr/local/lib to /etc/ld.so.conf.

Best regards,
Victor

Sure - dump-at-netxms.org (replace -at- with @).

Best regards,
Victor

Hello!

It's a bug in database initialization script. You can download fixed one at https://www.netxms.org/download/patches/003/dbinit_sqlite.sql and replace existing one in <install_dir>\lib\sql. Then, run nxconfig.exe to complete server's configuration.

We will also release fixed version shortly.

Best regards,
Victor

Hello!

I see only one solution - to use push DCI. You can define DCIs on web server's node object, set their source to "Push Agent", and then on management server (or any other machine) run your custom checks by cron and use nxpush command line utility to push new values to DCIs on web server's node object.

I also have a plans to implement "internal push" operation which will allow automatic copy of values from one DCI to another, but for current release the only way is through nxpush utility.

Best regards,
Victor

Hi!

Yes, content of config.log will be very helpful.

Best regards,
Victor

Hello!

It's a bug in FreeBSD subagent. Try to do the following: add the following lines to the end of src/agent/subagents/freebsd/freebsd.cpp:

extern "C" BOOL __NxSubAgentGetIfList(NETXMS_VALUES_LIST *pValue)
{
  return H_NetIfList("Net.InterfaceList", NULL, pValue) == SYSINFO_RC_SUCCESS;
}
extern "C" BOOL __NxSubAgentGetArpCache(NETXMS_VALUES_LIST *pValue)
{
  return H_NetArpCache("Net.ArpCache", NULL, pValue) == SYSINFO_RC_SUCCESS;
}

(it's a missing entry points caused message Cannot loads platform subagent "/usr/local/netxms/lib/libnsm_freebsd.so": Subagent doesn't provide any usable parameters) and recompile the agent. This should help.

Best regards,
Victor

Hello!

It's not possible now. But idea is very good, so I definitely will implement it.

Best regards,
Victor

Hello all!

NetXMS 0.2.17 is out. Changes from previous version:

- Implemented alarm timeouts
- ODBCQUERY subagent improved
- Web interface improved
- Object browser in Windows console improved
- Fixed issues: #122, #128, #130, #136, #138, #141, #156, #157, #158,
                #159, #161, #162

Best regards,
Victor

Hello!

I've never try to do anything with Windows Internal Database, but as far as I understand, it's just a variant of MSDE - so it may be possible to run NetXMS on it using SQL server driver (or via ODBC). You have to try.

Best regards,
Victor

Hello!

Looks like your MIB file refers to another MIB file which you don't include in your compilation. Look at the IMPORTS section in your MIB file - it should contain something like

IMPORTS
  .. some text ...
  BITS
  .. some text ..
  FROM another_mib_module

Try to find this additional module and include it in compilation. You can also post your MIB file here, then I'll take a look at it.

Best regards,
Victor