Problems in RNRP
Hello, i have a problem with RNRP. Computers lose connection due to suspicion of loop or high broadcast. Where can I change rnrp parameters or is there information about it? I need to change broadcast parameters or amount of traffic in packages.
Thanks
Voted best answer
CHANGING PARAMETERS WONT FIX YOUR PROBLEM.
Those connections are being lost for a reason and it isnt because some "tuning" parameters in RNRP are wrong.
- Check that Receive Side Scaling is disabled in EVERY network card. This is very important !!
- Check that Rapid Spanning Tree is disabled in your switches.
- Make sure you dont have any other features that could be considered as providing "network redundancy" by switching packets in a network card or switch.
- The address range indicates you are using implicit addressing. Check that nobody has set explicit address parameters.
- Double Right Click the RNRP taskbar icon to get the diagnostics window and run the tests. Look for configuration errors in each node. Correct the errors and restart RNRP in any affected nodes.
- If the previous are checked and correct, then its quite likely you have a real network issue .... faulty cable, loose connection, faulty switch, bad fibre termination etc. Try to examine where the faults occur and if a particular group of stations are affected that use common network hardware.
Answers
Hi,
RNRP’s Loop Detection and Protection enabled by default from 5.0 SP2 RevDand later,
let me know which version you are using?
For Preventing Network loops Please check the following
Avoid connecting switches in a physical loop
Don’t use Auto MDI/MDI-X for ports connected to PCs
Don’t mix cross cable with straight cables
Filtering options in switches (Ingress filtering, bandwidth storm)
Media redundancy protocols RSTP (Rapid Spanning Tree Protocol) Activate in switches if they will be a part of a loop
Very good input from Rob Lyon!
i just want to add one more thing:
VMware seems to be in use.
At some point in time and certain version of ESXi and Microsoft Windows there were known issues with the (default) VMXNETx virtual NIC drivers of ESXi. In this scenario, the use of E1000 was promoted.
Add new comment