NetXMS Support Forum

Service check is performing the checks via NetXMS agent that is installed on NetXMS server. The agent should have portcheck subagent enabled.

But overall service check functionality is old and is not being developed. It's recommended to use netsvc subagent, which is based on curl. It provides Service.Check(URL[, regex]) agent parameter. Here's some information: https://www.netxms.org/documentation/adminguide/service-monitoring.html#netsvc-configuration

Ah, the same problem with desktop client also. So it's a different problem, then initially in this topic.

Is netxmsd process actually running at the moment, when you are trying to connect?
On what OS netxms server is running?

If it's linux, try to capture threads at the moment, when you are unable to connect. To do this, run this script: https://raw.githubusercontent.com/netxms/netxms/master/tools/capture_netxmsd_threads.sh
This script requires gdb to be installed on the system. Run the script 3 times with 10-20 seconds interval. It will produce files in /tmp. Please share these files.

Clearing cookies in your browser should help. Or you can incognito mode or another browser.

К примеру, мысобираем агентскую таблицу по параметру Agent.SubAgents. Сам этот табличный DCI назван "Loaded subagents". И там какие-то такие данные:
| NAME                | VERSION             | FILE                |
+---------------------+---------------------+---------------------+
| Linux               | 4.0.484             | linux.nsm           |
| PORTCHECK           | 4.0.484             | portcheck.nsm       |
| DBQUERY             | 4.0.484             | dbquery.nsm         |
| FILEMGR             | 4.0.484             | filemgr.nsm         |
| PING                | 4.0.484             | ping.nsm            |
| SSH                 | 4.0.484             | ssh.nsm             |
| VMGR                | 4.0.484             | vmgr.nsm            |

И мы хотим иметь DCI, в котором будет версия субагента PING. Делаем в script library скрипт:
table=GetDCIValueByDescription($node, "Loaded subagents");
if (table!=NULL)
{
  col=1;
  for (i=0; i < table->rowCount; i++)
  {
    if (table->get(i, 0) == "PING")
    {
      return table->get(i, 1);
    }
  }
}
return null;

Делаем на ноде новый DCI, выбираем тип Script, в поле parameter пишем имя под которым мы сохранили скрипт. Этот скрипт будет доставать из нашего табличного DCI значение из 1-й колонки (колонки считаются от 0) из той строки, где в 0-й колонке написано "PING".
Но сейчас у нас прямо в скрипте прописано, что именно брать, а хотелось бы, чтоб скрипт был универсальным, а мы бы ему сообщали параметрами. что именно нам нужно.

//  $1 - Table name
//  $2 - Column name from where to return the value
//  $3 - Column name where to search
//  $4 - Value to search

table=GetDCIValueByDescription($node, $1);

if (table!=NULL)
{
  col_return = table->getColumnIndex($2);
  col_search = table->getColumnIndex($3);
  if (col_search >=0 and col_return >= 0)
  {
    for (i=0; i < table->rowCount; i++)
    {
      if (table->get(i, col_search) == $4)
      {
        return table->get(i, col_return);
      }
    }
  }
}
return null;
Параметр нашего скриптового DCI теперь должне быть таким: GetTableCell(Loaded subagents, VERSION, NAME, PING).

.metadata is hidden folder, so it wont be shown by ls command. One of the options is to run:
cat ~/.nxmc/data/.metadata/.log

As for the server log file, it's necessary to increase debug level for  client.* and userdb debug tags. You can do this with the help of nxadm command:
nxadm -c "debug client.* 6"
nxadm -c "debug userdb 6"
and you can check current debug levels with
nxadm -c "debug"

after setting debug levels, wait until you get console disconnect and share complete server log file.

On what operating system do you run the management console?

Upgrade both, just to be on the safe side.

Try clearing cookies in your browser

Was it the same with a previous version of NetXMS, or you did not use it before?
This may happen, if your computer is switching from network to network, e.g. from ethernet to wifi. Also try pinging netxms server from your desktop to check that there are no connectivity issues.

There are thought to improve the protocol between management console and the server, one day this should be done, but not very soon.
Also, you can deploy web interface on the server. Communication with the web interface might work better in non-stable network.

Стоит помониторить память, потребляемую процессом netxmsd (агентский параметр Process.VMSize) и вообще свободную память. Один из вариантов - oomkiller - штука в линуксе, которая убивает процессы, если в системе заканчивается.
Еще можно запустить netxmsd не через systemctl, а в интерактивном режиме - это когда он лог в консоль пишет
netxmsd -D 5
цифра после D - уровень отладки. При этом netxmsd завершится если отключиться от консоли, нужно или держать соединение, или воспользоваться screen. После вылета посмотреть, может быть в консоли будет какая-то информация.

When NetXMS server starts, it makes a record in the database to mark that it's locked. This is to avoid another copy of NetXMS server to start on the same database. NetXMS server should remove this mark when it's being terminated, but sometimes, it does not happen.
You can unlock the database by issuing
nxdbmgr unlock

Something like this happens if lauching the installer directly from Firefox after download. Try running downloaded exe from windows explorer.

From first glance looks like it's pinging the interface and gets reply:
[19.11.2020 08:41:09]    Starting status poll on interface lte1
[19.11.2020 08:41:09]       Current interface status is NORMAL
[19.11.2020 08:41:09]       Starting ICMP ping
[19.11.2020 08:41:09]       Interface is NORMAL for 3763 polls (1 poll required for status change)

If you open Object Details for lte1 interface, does it has correct IP address? Could it be, that server can ping that address by some other route?

Try collecting server log at debug level 7 for the duration of status poll.

The DCIs you mention are not collected by SNMP. These are generated by NetXMS server while it's pinging nodes (see https://www.netxms.org/documentation/adminguide/icmp-ping.html#icmp-response-statistic-collection).

There should be some other reason why this router is not showing as down. Could be the correlation mechanism.
1) Show the output of status poll for this node.
2) Try temporarily increasing debug level to 6 or 7 for the server for the moment when this node goes from to down and show server log.

Please give some more details. Is there NetXMS agent or SNMP on that node?
If there's not, it's the default behavior of NetXMS - it will be pinging the node and showing that it's down.
If yes, would be interesting to know the use-case for that configuration.