NetXMS Support Forum

I forgot to put it into installer. x64 version attached.

Best regards,
Victor

Try to execute manually query

UPDATE metadata SET var_value='322' WHERE var_name='SchemaVersion';

then netxmsd should start.

Best regards,
Victor

Oops, typo in PostgreSQL query Run nxdbmgr with -X option and the run manually the following queries:

ALTER TABLE users ALTER COLUMN system_access TYPE bigint;
ALTER TABLE user_groups ALTER COLUMN system_access TYPE bigint;

Best regards,
Victor

Quote from: Marco Incalcaterra on June 13, 2014, 12:49:06 AM
seems a good idea, I tried with an UPS but as far as I can see the node B will inherit the interfaces of the node A, can be avoided?

you need devemu subagent for that. This is how config for emulated device on my PC looks like:


MasterServers = 0.0.0.0/0
LogFile = C:\DevEmu\ATM001\log\nxagentd.log
ListenPort = 9001
EnableSubagentAutoload = no
SubAgent = devemu.nsm

*DEVEMU
IpAddress = 192.168.84.11
IpNetMask = 255.255.255.0
MacAddress = 551010000001
InterfaceName = Local Area Connection


Best regards,
Victor

Можно сделать такой configuration poll hook скрипт:


foreach(i : GetNodeInterfaces($node))
{
   if ((i->ifType == 24) && !AddrInSubnet(i->ipAddr, "127.0.0.1", "255.0.0.0"))
   {
      PostEvent($node, "BAD_LOOPBACK_CONFIG", null, i->name);
   }
}


Он будет посылать событие с именем BAD_LOOPBACK_CONFIG на каждый найденный интерфейс с типом 24 и адресом не из 127.0.0.0/8. Естественно, событие должно быть сначала создано. В %1 попадет имя интерфейса.

Not yet. Hopefully will be documented until 1.2.15 release.

Best regards,
Victor

Hi!

You can try to do SNMP walk from MIB Explorer view on different parts of MIB tree - but most likely you'll have to find some additional information about your devices - often data you get via SNMP is meaningless without description.

Best regards,
Victor

Происходит это так: NetXMS исключает loopback интерфейсы из топологии (что вроде как правильно). Loopback интерфейс определяется по адресу (из 127.0.0.0/8) или по типу (24 - software loopback). При дисковери проверяется, есть ли уже найденный IP адрес в системе. Поскольку loopback интерфейсы исключены из топологии, адрес не находится - создается новая нода. И так далее. Если таких устройств немного - можно добавить их адреса в фильтр discovery. Какой правильное решение - пока не знаю, возможно добавить конфигурационный параметр, который разрешит учет loopback в топологии если у них адреса не из 127.0.0.0/8?

Hi!

There are tutorial on creating alarms and sending notifications: http://wiki.netxms.org/wiki/How_to_send_e-mail_notification_when_disk_space_is_low - just replace NetXMS agent DCI with appropriate SNMP OID.

Best regards,
Victor

Hi!

Your best bet is to create simply script with nxshell (http://wiki.netxms.org/wiki/Using_nxshell_to_automate_bulk_operations) to read list of addresses and create node objects. If you are not good at Python - just ask, I'm quite sure that somebody will create such script.

Best regards,
Victor

Hi!

Change your DCI data type to String - this should help.

Best regards,
Victor

Now I understood Console don't allow SNMP walk if MIB selection dialog opened from template - this is because server does not know on which node it should do SNMP walk. It probably could ask for a node object to walk on instead of giving an error - I would add it as feature request.

Best regards,
Victor

Hi!

It's a new feature intended for monitoring devices without IP address and/or network connectivity. For example, you have serial attached UPS that you want to represent in monitoring as separate node. Then your configuration could be following:

1. connect UPS to some machine with network adapter (node A), setup agent (usually separate instance) and configure UPS subagent as usual
2. configure DEVEMU subagent if needed (it can return fake interface list)
3. create node for UPS (node B), set it's IP address to 0.0.0.0
4. set node B primary IP address to that of node A and check "This is address of remote management node"

After that all requests to NetXMS agent on node B will be actually directed to agent of node A. It is different from just defining node A as proxy node for DCIs on node B because server will handle all information provided by agent (interface list for example) as coming from node B. This gives you ability to use agent for status polls of non-networked box.

This option was created for two purposes - 1) to monitor serial attached controllers - we setup number of agents on a PC, each agent communicating with one controller and impersonating it for NetXMS; and 2) for simulating multiple hosts on one machine without need to create VM for each.

Best regards
Victor

Hi!

I'm finally built an intermediate release. It is available here: https://www.netxms.org/download/1.2.15-M1/.

Best regards,
Victor

Hi!

Can it be that status polling is disabled for that server? Otherwise looks like a bug. If you have debugging level 5 or higher enabled, you can look in log if there was status polls for that server and how they end up.

Best regards,
Victor