Re: [j-nsp] iBGP convergence time

2007-02-17 Thread Harry Reynolds
Hey Richard, I can't comment on longer convergence times, but can perhaps clarify the 8.2 INH feature. JNI started using RE based INHs back in the 5.x time-frame, for the reasons you describe. When a NH change occurs the RE can update a shared INH to forwarding-next-hop mapping, regardless of how

Re: [j-nsp] How to trace fpc/cfeb CPU usage?

2007-02-17 Thread Jakub Jermak
Hi, On Sat, Feb 17, 2007 at 09:17:23AM +0100, Joerg Staedele / Trusted Network GmbH wrote: > I dont have a graphical view of the cfeb cpu usage yet (is it possible via > snmp?) Just download mib-jnx-chassis.txt for your JUNOS version and it's: JUNIPER-MIB::jnxOperatingCPU.6.1.0.0 for cfeb cpu

Re: [j-nsp] How to trace fpc/cfeb CPU usage?

2007-02-17 Thread Joerg Staedele / Trusted Network GmbH
On Sat, Feb 17, 2007 at 12:10:11AM +0100, Rafa? Szarecki wrote: > Hi all, > > FEB/CFEB CPU is responsible for delivery information between PFE (asics) and > RE. > > If your m7i is doing packet sampling (netflow/jflow, or to file)? This is > one of very often reasons of hi CPU usage. > > This CPU