NetXMS Support Forum

Upgrades are incremental, so you can safely go directly to the latest version, "nxdbmgr upgrade" will handle all steps correctly.
Should be safe, but I'd backup first anyway.

EDIT: I tried upgrade from empty 1.2.17 database and found a small bug with the upgrade. Either wait for next release, or apply https://github.com/netxms/netxms/commit/b30dfa64644c639d177f4367bc6de13ff4708311 and rebuild.

Another option will be upgrading to some intermediate version, but I'm not sure which will work correctly (2.x something should work, but it will be trial and error).

Hello.

You need to upgrade Java to version 11 or newer.

Since 3.7 minimal java version is 11

You can download latest mapping here: http://standards-oui.ieee.org/oui.txt

Then parse it and save in the mapping table (in this example - using nxshell):

#!/usr/bin/env nxshell

TABLE_NAME='MAC Address'

# find mapping table, create if missing
table_id = -1
for t in session.listMappingTables():
    if t.name == TABLE_NAME:
        table_id = t.id
        break
if table_id == -1:
    table_id = session.createMappingTable(TABLE_NAME, 'Lookup table for MAC -> Vendor', 0)

target_table = session.getMappingTable(table_id)

with open('oui.txt', 'r') as f:
    target_table.data.clear()
    for line in f.readlines():
        if "(hex)" in line:
            (mac, _, name) = line.strip().split('\t')
            mac = mac.split(' ')[0].replace('-', ':')
            target_table.data.add(MappingTableEntry(mac, name, None))

session.updateMappingTable(target_table)

Then you can lookup with map() function (in this example - find interface which have primary IP address):


for (i : $node->interfaces) {
if (i->ipAddr == $node->ipAddr) {
return map("MAC Address", (substr(i->macAddr, 0, 8)));
}
}

Result: Apple, Inc.


On one of the nodes with agent (let's call it "node1") add external parameter:

ExternalParameter=ResolveMAC(*):curl -s "https://api.macvendors.com/$1"

Then add DCI in the template:
Parameter=ResolveMAC(%{node_mac})
Source Node=node1

But I'd use mapping tables instead of calling remote service.

Configuration -> Image Library

По умолчанию права наследуются от родителя (если это принудительно не отключить на контейнере). Именно так выданы права админам - на Infrastructure Services есть единственное правило для группы Admins.

Актуальные права пользователя высчитываются наложеним:
1) отнаследованные права на объекте
2) права на объекте для групп, в которых есть пользователь
3) права на объекте для пользователя

например можно сделать структуру (для примера есть user1, user2 и они в group1):
Container1 - тут дать read права для group1, наследование по умолчанию включено
+ Node1 - пустой access control
+ Node2 - тут в access control добавить user1 вообще без прав

по результатам:
user1 будет видеть Container1 и обе ноды
user2 будет видеть Container1 и только Node1 (read будет отнаследован от Container1, но потом перекрыт правилом на самой ноде)

права на объекты надо на самих объектах, права в редакторе юзеров - для глобальных вещей (типа достапа к логам)

Which version you are currently using?

There are multiple ways to do that, I'd probably use instance in action name:

1) configure 2 actions on the agent: StartA, StartB
2) Create 2 DCIs - Process.Count(A) and Process.Count(B) with threshold "< 1" on both (or any other way how you detect that process is down), set instance to "A" and "B" respectively (in thresholds tab, will be used in action name) and some activation event (can be the same).
3) handle this event in EPP with "Execute command on remote node via agent" action, as agent action name put something "Start%<instance>"

Sample config:


LogFile=c:\netxms\logs\nxagentd.log

[EXT:XYZ]
LogFile=c:\netxms\logs\nxagentd.xyx.log

Action=StartApp:c:\...\app.exe


Core agent (which runs as service) will read main section, and second instance of the agent will read EXT: section instead.
Second instance should be started in the user's session with additional command line arguments: -G EXT:XYZ -H
-G should match EXT: section in the config file and -H will hide agent's console window from user.

You can add agent to autostart using reg command (or in any other way):
reg add HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Run /v NetXMSExternalAgent /t REG_SZ /d "C:\NetXMS\bin\nxagentd.exe -c C:\NetXMS\etc\nxagentd.conf -G EXT:XYZ -H" /f

Один из вариантов - вытащить на карту не коммутатор, а его интерфейс (двойной клик на ноде в дереве откроет интерфейсы, если они еще не загружены), после чего поменять в пропертях интерфейса иконку для карты и имя.
Второй вариант - сделать ноду с IP=0.0.0.0, на ней DCI (скриптовый или с трансформационным скриптом) с включенной опцией "Use this DCI for node status calculation" и возвращать значение статуса (0-6), который и и будет менять статус этой ноды.

Check filtering script in event processing policy rules #3 and #5. You call nonexisting function there in line 3 of the script.

I assume that by "terminate event" you mean "terminate alarm".
Short answer is no - once terminated, alarm will remain only in alarm log. However, you can repeat threshold activation events (you can set that in threshold editor) and therefore recreate alarms, if threshold is still active. Everything else will not be repeated.

Another way - revoke "terminate alarms" permission from user (leave only "Resolve"), then terminate with separate user (or even another person after verification that issue was solved).

yum install libcurl-devel
потом ./configure без --without-curl