NetXMS Support Forum

Quote from: inspirit on March 04, 2013, 01:16:46 PM
Заметил в логе netxms следующие записи:

Code Select Expand


[04-Mar-2013 15:12:06.078] 9 network device drivers loaded
[04-Mar-2013 15:12:06.088] Built-in objects created
[04-Mar-2013 15:12:06.088] Loading container categories...
[04-Mar-2013 15:12:06.088] Apply template thread started
[04-Mar-2013 15:12:06.090] Loading built-in object properties...
[b][04-Mar-2013 15:12:06.110] NetObj::loadCommonProperties() failed for object Policies [5] class=16
[04-Mar-2013 15:12:06.114] NetObj::loadCommonProperties() failed for object Network Maps [6] class=19
[04-Mar-2013 15:12:06.117] NetObj::loadCommonProperties() failed for object Dashboards [7] class=22
[04-Mar-2013 15:12:06.120] NetObj::loadCommonProperties() failed for object Reports [8] class=24[/b]
[04-Mar-2013 15:12:06.131] Loading conditions...


На работе вроде бы не сказывается. Хотелось бы узнать - что это, и можно ли исправить, или забить.

Это нормально. Обьекты с ID меньше 10 - ето встроенные обьекты, которые есть всегда. Если они не менялись, то записей о них в базе может и не быть, они со значениями по умолчанию создаются.

Hi!

So far I plan 1.2.7 to be released between March 15 and March 20.

Best regards,
Victor

Problem with this MIB is that according to standard object names must start with lowercase letter, while all notification objects in this MIB starts with uppercase letter. Corrected MIB attached.

Best regards,
Victor

Value 10000 means "unreachable".

Best regards,
Victor

You have to options here:

1. Start using expected port state "ignore" for end users' ports instead of unmanaging them. Then you can use manage/unmanage on switch level without problems (additional benefit here is that you still be able to see link state of a port, but it's status will always be normal).

2. Add temporary rule in the beginning of event processing policy which will stop processing of all events from given node.

Best regards,
Victor

Hi!

Yes, you can right-click on node and select "unmanage". Then node will be put to "unmanaged" state and excluded from monitoring. To bring it back, select "manage" from context menu.

Best regards,
Victor

Hi!

Just fixed second and third bug, thanks for reporting! First one will take a bit more to fix.

Best regards,
Victor

Hi!

Both functions will require additional development. It is technically possible to create UI element filters in Eclipse RCP (underlying platform of client application) as additional plugins.

Best regards,
Victor

HP E-series supported already, driver's name is PROCURVE. If you have E-series switch for which generic driver is selected, please send me it's SNMP object id.

For creating driver for Shoretel switches, I need MIBs for them and result of SNMP walk of .1.3.6.1.4.1. Ideally also remote access to the device so I can test driver.

Best regards,
Victor

Hi!

There are few possible options. Option 1 is following:

1. Create additional event, ASA_SYSLOG_MESSAGE for example, and put varbind .1.3.6.1.4.1.9.9.41.1.2.3.1.5.0 into parameter #2, as you have done already with ASA_FAILOVER event.
2. Create script in script library, which will post ASA_FAILOVER event if it got ASA_SYSLOG_MESSAGE with "Switching to ACTIVE" text:


if ($event->parameters[2] match ".*Switching to ACTIVE.*")
{
   PostEvent($node, "ASA_FAILOVER", null, $event->parameters[2]);
}


3. Create new action of type "execute NXSL script", and specify name of the script you just craeted
4. Add rule to event processing policy, which will execute action created on step 3 for ASA_SYSLOG_MESSAGE event.

That's it - you will have ASA_FAILOVER events for each syslog message trap with "Switching to ACTIVE" text in it. Please note that in my script text will be passed to ASA_FAILOVER event as parameter #1.

If you don't want ASA_SYSLOG_MESSAGE events to clutter event log, you can uncheck "Write to log" option in event's configuration.


Alternative approach could be to not use SNMP traps for this, but feed syslog records from ASA directly to NetXMS server (using syslog protocol), and create a parser for incoming syslog records.

Best regards,
Victor

Hi!

Looks like application hangs or crashes at initialization. Can you check log files under web ui installation directory? Also, try to restart web ui service and access directly http://netxmsserver/nxmc, without accessing root.

Best regards,
Victor

Hi!

Driver for H3C (HP A series) switches is not ready yet. I have access to some HP A55xx devices, so driver will be provided in one of next releases.

Best regards,
Victor

Нашел баг в сортировке (для алфавитной сортировки использовалось поле "Name in legend" вместо заголовка графика). Следующая версия консоли будет сортировть правильно.

Quote from: inspirit on February 28, 2013, 11:46:49 AM
- можно ли как то контролировать рост записей в Mysql таблице snmp_trap_log? (housekeeper c ней не работает?)

Это баг. Добавил очистку старых записей из snmp_trap_log (контролируется новым параметром SNMPTrapLogRetentionTime). В 1.2.7 будет включено.

Quote from: inspirit on February 28, 2013, 11:46:49 AM
- Можно ли сделать что бы при выборе ноды в ява-консоли, после вкладки Ports был бы раздел Topology (Как в Legacy консоли раньше было, теперь оттуда тоже убрали).
-Можно ли менять местами (или вообще выбирать только нужные) столбцы в разделе Interfaces для Ноды (было бы удобно)

Поставлю в feature requests (и то и другое).

Надо заменить файлы node.cpp и snmp.cpp в каталоге src/server/core на приложенные.