Donald_Francis

You can sure do that. We actually do it ourselves But even with QOS if your circuit gets saturated enough you will drop traffic anyway, it would just be some of the last traffic to be dropped.

So I never had it happen again after those few times. I never found the reason why. I think the whole issue is that SNMP goes directly to the CPU and fries the box in the process. I will say this however. I jumped on SUP-6E's as soon as they came out and I have a lot of them out there now. If you have one you know they are…

This is something several of us have complained about. I consider it a huge issue and one that bars me from using disvoery. I have upward of 1500 devices so manually doing anything is impossible. The discovery needs to track hostname or serial or something to notice that there is a duplicate and to only list it once. If…

I would open a ticket, I have had a lot of poller issues as of late myself.

As far as I know and have seen TFTP is only UDP 69.

I would think that is a no. That is just my opinion however.

Just my 2 cents here, but I think you would need an Orion instance per VRF. Even if Orion could handle duplicates (I do not think it can) how would even the Server that it is runnnig know how to get to one VRF or another just from a routing and switching perspective. Unless you have some kind of NAT going I do not think it…

Without something like NCM to check the config I really do not know of another to do this outside of the one you mentioned. 

We actually monitor about 150 netflow sources. Most are T1s or T1 multilinks but also in there are 2 full OC3s, 3 100MB links and 1 1gb link. Our setup handles that well but I can say that netflow is the most intensive part of Orion. We actually had to get CBQOS turned off in the DB because it was running itself crazy but…

I do not think you can edit the one on the right but in the web settings you can add your own easily.

Interface = not up

Click on "Custimze Page" in the upper left hand corner and just delete the resource on that page.

Since you will have 3 pollers (I would get 3 additional pollers and let NPM no poll anything) I assume you have a lot of elements being monitored. In that case I would make all 4 boxes dual quad cores with 6+ GB of RAM. That should be able to handle a good load without having any trouble with it. Going to lower hardware…

I too am waiting for it to come out. From what I hear it should be within a week or 2.