[j-nsp] EX4200 9.4R2.9 process crash on previously valid config

Richard A Steenbergen ras at e-gerbil.net
Thu Apr 2 21:14:28 EDT 2009


On Thu, Apr 02, 2009 at 04:25:14PM -0700, Robert Raszuk wrote:
> > Honestly, I think you should just give up now and accept the fact that
> > Juniper is the new Cisco.
> 
> Excuse me ? Cisco in vast majority of new products is way much better 
> now. Yes historically there was an issue with IOS, but AFAIK that has 
> been also fixed now.

Dare I ask who you're defending here? In all honesty, with respect to
the quality and time/attention that goes into the software development
and QA process of new software, JUNOS has gone massively down hill
lately. Anyone who expects that a bug fix R1->R2 build is not likely to
introduce some massive new issue in something that was previously
working fine clearly hasn't been deploying JUNOS over the last year.

My advice to Jeff was that it is absolutely unrealistic to expect to be
able to deploy new code without testing it on your configuration in the
lab first, especially on the EX (which is still a very immature
platform). Maybe this is just the natural result of complexity, as JUNOS
tries to cope with being every thing to every one on every platform
without forking the code, but the point is that it isn't the same
"simple and just works" JUNOS he's dealt with in the past.

And on the subject of config validation, I would feel slightly better
about the whole thing if the issue I pointed out 2 years ago (that the
duplicate IP check went from an error to a warning, allowing a typo to
commit and bring down both interfaces) had ever been addressed, but I
just checked 9.4R2 and guess what it hasn't. I also gave several other
examples where config validation on the EX let you configure things that
weren't supported and which killed the box in the process, so it's
clearly something you need to watch out for on this platform.

> PS: On the topic of the future of EX-series I will at this point refrain 
> from commenting on the public list :).

EX makes me sad... I really want to like the thing, but I think Juniper
has just completely missed the boat from a product management
perspective. Another good idea ruined by trying to focus exclusively on
Enterprise. :( The software has gotten significantly less bad over the
last year though, I'm still holding out hope that it will be a good box
some day. 

-- 
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)


More information about the juniper-nsp mailing list