Open for Voting

New verb for Orion.Nodes (SDK)

Hi there

I'd like to see a new verb implemented that would allow for SDK trigger of node rediscover immediately after adding a node via SDK.  Its native already to the web gui and would be ideal for the SDK.

Without this verb network admins forced to set the NextRediscovery property to a low value, i.e datetime.datetime.now() + datetime.timedelta(minutes = 2) and then wait for the next rediscovery interval to come around (default 30 minutes) prior to being able to download configs in NCM.  This is due to the sysObjectID being unknown and NCM being unable to match the device to a device template.

A work around to this is to implement an SNMP get for the sysObjectID prior to setting Orion.Nodes properties and adding the device via SDK, however I believe this is a feature that should be native to the SDK as it already exists in the GUI.

Another workaround which I am implementing now is set the low NextRediscovery property and then schedule the config download task in Celery/Redis for 60 minutes out after adding the node.  This however could have an adverse effect on workers and queues if there are a lot of nodes being imported say on a major upgrade or re-install.

Please create a new verb RediscoverNow in Orion.Nodes to give us more flexibility.

Thanks

Donald

  • my scripts that add nodes already have to grab the sysobjectID of the node (from a CMDB) -- the reason is that you have to programmatically ad the correct pollers for the node (e.g. the CPU, Memory, LLDP, routing tables, bridge tables, etc) and the only way to do that is to get the sysobjectID.

    If you want to solarwinds to do this then it might be argued that the correct way to do this is to do a network-discovery of the node and then parse the results to add the node from the discovery (as this will also add all of the correct node pollers and give you the interfaces and volumes that you might need to monitor)