NetXMS Support Forum

Quote from: blazarov on November 06, 2015, 01:34:21 PM
I just still don't understand what happens when some interface is deleted and instance discovery runs. Are the DCI with their history completely deleted? Is there a way to control that - for example leave - delete them one month after disappearing?

currently no - DCIs deleted instantly when instance discovery poll founds that instance do not exist anymore. Feel free to register a feature request - it sounds reasonable and should be relatively easy to implement.

Best regards,
Victor

Можно для интерфейсов сделать дополнительную иконку (скажем в правом верхнем углу) для обозначение operational state. Добавил в трекер: https://dev.raden.solutions/issues/974.

Quote from: hsvt on November 05, 2015, 10:47:44 PM
У меня для тех у кого ignore (абоненты, например) эвент не генерируется, состояние нормал, Иконки в разделе Ports - красные. Другой вопрос, если бывает, нужно поставить на мониторинг абонентский порт с какой нибудь плавающей проблемой, а у тебя по Hook::ConfigurationPoll всё равно будет ставиться Ignore. Как вот тут лучше сделать?

Можно проверять custom attribute интерфейса и не ставить ему ignore если атрибут выставлен в какое-то специальное значение.

Quote from: hsvt on November 05, 2015, 10:47:44 PM
Ну и с почтой и эвентом SYS_NODE_DOWN тоже не понятно, надо чтобы от одних устройств были только алармы, а от других и алармы и письма еще...

А как эти устройства различаются? Самое простое - положить все устройства где нужна почта в отдельный контейнер, и в EPP отсылку писем фильтровать еще и по контейнеру. Другой вариант - использовать custom attributes для отметки устройств, по которым надо почту слать.

You have to specify agent address in command line as well (before action name).

Best regards,
Victor

Hi,

can you please repeat this test with debug level set to 6? Also, are you using SNMPv3 for sending traps?

Best regards,
Victor

Hi,

then you can use transformation script on table DCI to find top five values and delete the rest.

Best regards,
Victor

Hi,

instead of terminating (probably non-existent) alarm for non-core systems you can just use "stop processing" (set by selecting checkbox on Actions property page for the rule). But otherwise looks correct.

Best regards,
Victor

Hi,

are you using desktop or web client?
As for geo maps, you can setup local tile server (https://switch2osm.org/serving-tiles/manually-building-a-tile-server-12-04/) and chenge server configuration parameter TileServerURL to point to your local tile server.

Best regards,
Victor

Hi,

seems to be a bug. I've registered it in our bug tracker: https://dev.raden.solutions/issues/973

Best regards,
Victor

Hi,

below is an example of the script that walks all parents of a node and collects values of custom attribute "contacts" into single string separated by semicolons:


global contacts = "";

foreach(o : GetObjectParents($node))
{
add_contacts(o);
}

println "Contacts: " . contacts;

sub add_contacts(curr)
{
c = GetCustomAttribute(curr, "contacts");
if (c != null)
{
   if (length(contacts) > 0)
contacts = contacts . ";" . c;
else
contacts = c;
}

foreach(o : GetObjectParents(curr))
{
add_contacts(o);
}
}


The only drawback is that it will not detect duplicate values. Duplicates can be eliminated like this:


global contacts = "";
global presense = %{ };

foreach(o : GetObjectParents($node))
{
add_contacts(o);
}

println "Contacts: " . contacts;

sub add_contacts(curr)
{
c = GetCustomAttribute(curr, "contacts");
if ((c != null) && (presense[c] == null))
{
   if (length(contacts) > 0)
contacts = contacts . ";" . c;
else
contacts = c;
presense[c] = true;
}

foreach(o : GetObjectParents(curr))
{
add_contacts(o);
}
}


Note that this script version requires server 2.0-RC1 or higher as it uses hash maps.

Best regards,
Victor

Hi,

from mane sessions stuck in CMD_GET_OBJECTS most likely there is a deadlock in a server. The only way to recover from that situation is to restart the server. I would suggest to upgrade to 2.0-RC2 or wait for 2.0 release - we have fixed some deadlocks in 2.0 branch so there are good chance problem will gone.

Best regards,
Victor

Hi,

how instance discovery is configured? Are you sure that you have correct ifIndex in $1 in filter? I just checked WNMP walk part of the script on my system (using execute server script option on node object) and it works fine - returns all IP addresses on my router.

Server will delete DCIs created by instance discovery if instance will gone.

Best regards,
Victor

Hi,

seems to be packaging problem again. I've fixed package definition.

As for the second problem, location of compiled MIB file (netxms.mib) was changed in RC2 - you have to move it to the root of server's data directory (set by DataDirectory option in netxmsd.conf), default is <installation prefix>/var/lib/netxms

Best regards,
Victor

да, ветки это webui/webapp и src/java/netxms-eclipse. Для веба используем Mars, RAP 2.3 или 3.0 (должно собираться и с тем и с тем). Релиз пока делаем на 2.3, поскольку 3.0 требует Servlet API 3.0, а некоторые сервера (Weblogic 11 например) его не поддерживают.