NetXMS Support Forum

Also, in 3.5 there's now an ability to skip action, if a particular timer is running. So your condition should now fit into 2 EPP rules.

В 3.5 в свойствах действий добавлена проверка - если существует таймер, то не выполнять действие. Теперь можно обойтись двумя правилами - в правиле node up проверять, если таймер, заведенный в node_down еще не истек, то не слать оповещение про node up.

Тогда ищите во View -> Alarm Log, фильтруя по времени создания.

It was just fixed, will be available in next patch release.

https://www.netxms.org/documentation/adminguide/reporting.html#installation-manual

Там нужно правильно конфиг файлы написать, потому что если в них что-то не то, то толковой обработки ошибок нет.

Похоже, что он доапгрейдил базу до 30.18 и на этом остановился. 30.18 соответствует версии 3.0-M0, можно попробовать поставить ее и там сделать nxdbmgr check.

Еще стоит сделать селект ThreadPool.Poller.BaseSize из config, есть ли вообще такая запись в этой таблице и что в ней.

Yes, it could be possible to create two interfaces manually for these nodes. If an interface is created manually, it will not be deleted (there's a flag about that not visible in UI).

As far as I know there's currently no built-in logic for interface failover. With some NXSL scripting it should be possible to set up a script DCI that is checking interface status and produces an alarm when both interfaces are down (and there could be some additional logic to avoid the alarm when these laptops are officially shut down).

But if an IP address changes for an interface, there will be a problem anyway.
SNMP could help, but thin clients could also not support it.

In general there's currently a flaw in the logic, it should get improved. I've created an issue for this: https://track.radensolutions.com/issue/NX-1952

Эти алармы должны найтись во View->Alarm browser

Но для этого нужно зайти под юзером system (по умолючанию этот юзер отключен, его нужно включить в Configuration->User manager)

Он все еще в состоянии "когда-нибудь" - вещь хорошая, но на разработку нужно порядочно времени. Единственный шанс, чтоб он быстро появился - если кто-нибудь закажет разработку за деньги.

Была проблема с тем, что инсталлятор NetXMS не ставил в систему какой-то из C++ runtime и драйвер постгреса из-за этого не работал. Это починено, но видимо в какой-то из следующих версий, чем 3.3.323
Если можете, попробуйте проапгрейдится.

Hi!

Did you try to connect from desktop client, not the web one? Did you suceed?

We did some testing on this.

If the node has NetXMS agent, everything works properly - domain name is resolved into IP, we connect to agent on this IP, agent replies.
In this case agent will report that the node has two interfaces, one will be UP, the other DOWN. In interface configuration you would need to set expected status to IGNORE, otherwise one of the interfaces will always produce an alarm on the node.

The other case is when node does not has NetXMS agent. In this case we use ICMP checks, which take IP address from interface. The thing is that in current version IP address on interface is not updated on status poll.

Do I understand correctly that you do not have NetXMS agent on these laptops?

We will need to test this internally. It should work this way, if dns name gets properly updated.

There's also an alternative approach to use tunnel connections - in this case nodes are connecting to server and are identified by their unique identifier.

Please share the dump file with the help of any online service, e.g. wetransfer.com (choose Get transfer link there in properties to get a link to the file)

The reasons for that could be two things - either slow database performance or locks in server process.

Please set LongRunningQueryThreshold server configuration parameter to, e.g. 10000. If queries that take longer then 10 second would happen, they will be written to server log file.

If not responding events are happening, but no log running queries are logged, then it's probably locks in server process.
If your server is on Linux and can catch the moment, when this event happens, try running this script:
https://github.com/netxms/netxms/blob/master/tools/capture_netxmsd_threads.sh

gdb should be installed on the system to run this script. Script should be run 3-4 times with interval 30-60 seconds. It will create files with debug information in /tmp. Please share these files.

It's also possible to run this script as action for not responding event, but care should be taken not to run this script too often, as it's quite heavy temporarily pauses all netxms processes. So the action should have a script that stores last launch time in, e.g. persistent storage and checks that some time passed since that time.