Nodes showing packet loss and/or down status when there is no packet loss in reality

filipv

Since couple of hours ago SOME of the nodes in my monitored networks are showing severe packet loss or down status (100% packet loss). Those readings all begun at exatcly the same time.

I logged on to the polling server and tried pinging those nodes: zero packet loss. Services are obviously running, traffic graphs all look as they should, no complaints from the customers.

Yet, Solarwinds says that many (not all!) nodes are down or have huge packet loss.

So, it's not a networking issue from the polling server to the rest of the network because, as I said already, when I ping those devices from the polling server itself there is no packet loss.

I tried restarting all Solarwinds services. The problem still remains.

So, I'm copletely baffled: In what scenarios NPM could report packet loss when there isn't any packet loss?

Thank you for your thoughts!

FormerMember

We only see this when the device is CheckPoint/Nokia HA firewall pair.  Can't recall the exact details, but its related to asymmetric routing between the pair and the return ping packet being dropped I believe.

We still don't know why we can ping from the SW box itself and everything appear fine, but when SW polls it gets 50%-100% loss.  It must be something to do with how the packet is generated.