casonline

I'd like to think that - however that only works for node's who's IP address is already in the system as a monitored node. It does not account for the situation where your flow packets are sourced from a known node in the system, but from an IP address different than that which is being monitored as the node IP. If your…

What is the purpose of the NCM polling service ? if we use NPM to monitor/poll do we need the NCM poller service ? sounds like rediscovery can still take place independantly ? I'm having an issue where the NCM polling service is causing my SQL server cpu and load to spike heavily, so I shut it off. Need to know the impact…

any update on this ? can we now use a/d groups in NPM 10 / NCM 6 ?

A follow up to this - any ideas ? Can we tweak NPM to recognize the SFLOW agent IP as a "known node"

doesn't this leave a ton of orphaned records in other tables linked to those nodes' nodeID ? I did this manually and it seems to not delete the interfaces associated with the node out of the interfaces table

has proxy support been added yet ?

Brandon - Chad here... I have a particular need that allows for validating status of a node via snmp, and not just ICMP.. i.e during normal status polling (UP/DOWN), if we lose icmp to a node, but still have snmp access to that node, the node should still be considered up... Do we have that ability today ? 

Just as a valid test – I’ll shutdown the NTA service to see if I experience the same results. I’ve been told that we are just hammering our DB. The question is, what is the baseline for these types of metrics, under low, medium, and heavy loads. Does anyone have something we can reference ?

I realize you can do that on Cisco, but this isn't Cisco :) There is no setting for sflow/netflow source-IP on Foundry, however you can set the Agent IP field in the sflow packet. I believe I sent you a packet cap before based on this thread... if you don't have it, I can try to send another. You'll have to shoot me your…

does anyone have any ideas on this issue ?

sure - what's your email address ?

Has this been resolved yet ? thanks

I'd like to see the ability to do custom inventory for things OTHER THAN Cisco. For example, I'd like to be able to inventory a node, say a foundry switch, and get the serial number. This would operate sort of like custom pollers. you could relate custom inventory elements to types of devices/vendors. This would be huge to…

Is there any update to this ? We really need the ability to do SNMP informs to get away from agent-based solutions. I'm assuming the answer is NO, until SW gets off reliance on the Microsoft SNMP service for trap recption ? SW really needs to integrate their own SNMP service !!

You could front end by a load balancer. Replicate web site info and use standby pollers as well. Then all syslog and traps would only go to the active host.

I'm having the same issue... reboot and all comes back - now it's stopped again as I just found out to my dismay... the stability of the system is being questioned by mgmt now... not good.. Does SW have any ideas on this yet ? I'll call in a ticket referencing this other ticket as well...

I'll add to that... I have similar requirements/issues... Some may be by design, or may be bugs...dunno. I’m running into an issue where I’ve got a group/user set to have node management rights, and I’ve setup an account limitation to limit that is based on a group, say Vmware, which I’ve set as a dynamic group for any h/w…

I'm having the same problem on v9.2 sp2... any answer to this ? I'm trying to do traffic reports on a larg number of interfaces... > 900

I've got the same issue... checked my devices and they are not configured for netflow/sflow...