NetXMS Support Forum

Hi!

All steps you do are correct. Can you please post screenshot of your table configuration? Also, is it possible to provide me with SNMP access to some test Mikrotik device so I can test if there are any bug withing NetXMS?

Best regards,
Victor

Тогда ничего сделать нельзя. Я подумаю как можно добавить переменные в запросы в DBQuery.

Странно, сервер должен кешировать engine boots и engine time и повторять запрос при получении "not in time window" ошибки. Есть возможнось собрать дамп SNMP пакетов при возникновении этой проблемы?

Такой возможности нет. Можем добавить в feature requests

Я поправил работу с сессиями в dbquery - надеюсь поможет. В 1.2.14 будут эти изменения.

Есть несколько путей - смотря что надо получить:

1. Можно обрабатывать события SYS_NODE_DOWN, SYS_NODE_UNREACHABLE, SYS_AGENT_UNREACHABLE, SYS_SNMP_UNREACHABLE и слать оповещения например.

2. Можно сделать "internal" DCI и в transformation script получать состояние ноды через $node->runtimeFlags. Например, проверить что нода полностью недоступна можно так:

if ($node->runtimeFlags & 0x0004)
{
   // node is unreachable
}

Полный список флагов из runtimeFlags:

QUEUED_FOR_STATUS_POLL     0x0001
QUEUED_FOR_CONFIG_POLL     0x0002
UNREACHABLE                0x0004
AGENT_UNREACHABLE          0x0008
SNMP_UNREACHABLE           0x0010
QUEUED_FOR_DISCOVERY_POLL  0x0020
FORCE_STATUS_POLL          0x0040
FORCE_CONFIGURATION_POLL   0x0080
QUEUED_FOR_ROUTE_POLL      0x0100
CPSNMP_UNREACHABLE         0x0200
RECHECK_CAPABILITIES       0x0400
POLLING_DISABLED           0x0800
CONFIGURATION_POLL_PASSED  0x1000
QUEUED_FOR_TOPOLOGY_POLL   0x2000
DELETE_IN_PROGRESS         0x4000
NETWORK_PATH_PROBLEM       0x8000

Да, по SNMP. Можно с консоли запустить Poll -> Topology и посмотреть что там происходит.

Сейчас можно поставить threshold на data collection error - он сработает если запрос параметра вернет ошибку. В 1.2.14 я добавил macro expansion еще и в аргумент pattern в параметрах File.Size и File.Count - теперь можно будет сделать параметр вида

File.Count(/path/to/files,%Y-%m-%d)

и сравнивать его с 0.

Hi!

Do you mean automatic agent installation on remote machine or just automatic addition of new machines with agents to monitoring database? Latter is possible with automatic discovery - you can turn on automatic discovery and set filter to only accept nodes with agent installed. Then you can use automatic binding scripts on containers to place those new nodes into object tree. You can also send notification when new node added by adding appropriate action for SYS_NODE_ADDED event.

Best regards,
Victor

Hi!

You can create internal type DCI (with source set to "Internal") and Dummy parameter - it will just generate 0. Then in transformation script you can use SNMPGetValue function to get actual value of required OID. This function will return null if OID does not exist, so you can check it and return some meaningful value in that case.

Best regards,
Victor

Hi!

I would suggest the following scheme:

Create events (of course you can name them any way you want):

TRAP_RAW (normal severity)
TRAP_INFO (normal severity)
TRAP_WARNING (warning severity)
TRAP_CRITICAL (critical severity)
TRAP_CLOSE (normal severity)

Create SNMP trap mapping for trap .1.3.6.1.4.1.23986.1.4.1 with the following varbind mapping:
#2 - .1.3.6.1.4.1.23986.1.2.3.1.9.200083
#3 - .1.3.6.1.4.1.23986.1.3.4.1.6.2368
#4 - .1.3.6.1.4.1.23986.1.3.4.1.8.2368  (this is needed if you want to pass message as parameter to event)

Create server action that will call the following NXSL script from library:


if ($event->parameters[2] == 2)
{
   switch($event->parameters[3])
   {
      case 3:
         evt = "TRAP_CRITICAL";
         break;
      case 2:
         evt = "TRAP_WARNING";
         break;
      default:
         evt = "TRAP_INFO";
         break;
   }
   PostEvent($node, evt, null, $event->parameters[4]);  // pass message as first parameter to generated event
}
else
{
   PostEvent($node, "TRAP_CLOSE");
}


Then create the following rules in event processing policy:

1.

event is TRAP_RAW

action: call script and stop processing

2.

event is TRAP_INFO, TRAP_WARNING, or TRAP_CRITICAL

action: create alarm with some key

3.

event is TRAP_CLOSE

action: terminate alarm with key used in rule #2

Please note that message extracted from trap will be available as %1 in message template for generated events and alarms.

Best regards,
Victor

Can you please compare MD5 hash of files netxms.mib on server and on client? Problem in 1.2.11 was caused by error during file transfer  -we have to check if now it is the same issue or not.

Best regards,
Victor

Hi!

I suppose you mean 1.2.13 - 1.2.12 already archived and not referenced directly on download page. I'll prepare 1.2.14 release in a next few days - it should fix crashes discovered in 1.2.12 and 1.2.13. Then 1.2.13 will be moved to archive.

Best regards,
Victor

Try to do configuration poll for the node with changed IP address - will it resolve primary host name to new IP?

Best regards,
Victor

Hard to suggest anything except trying to upgrade to latest Java 7 JRE or maybe try Java 6.

Best regards,
Victor