[j-nsp] KRT Queue issue (was: Re: bfd = busted failure detection :)

[Richard A Steenbergen]

On Wed, Jan 06, 2010 at 09:26:03AM -0700, David Ball wrote:
> Ran into this again on a t640 running a 9.2SR release a couple days
> ago.  KRT queue was 2-3k operations queued yesterday, is over 52k
> operations queued now.  An RPD core dump was apparently the trigger
> the LAST time this happened to us, but no core dump this time, so it's
> starting to look a little more like the apparently random occurance
> the rest of you have been experiencing.  It's with ATAC.

We had an RPD core dump on us while upgrading to 9.3R4 on an MX960
several weeks ago, no progress on the case thus far, but that was the
first time we've seen something like that happen. I've been replicating
the slow route installation issue for Juniper on a few different boxes,
but we still haven't gathered enough information to track down the
cause. On the first example case, almost the entire routing table took
over 18 minutes to install on 9.5R3. As soon as we figure it out, I'll 
let you guys know. :)

-- 
Richard A Steenbergen <ras at e-gerbil.net>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)