Mailing List Archive

On Tue, Feb 11, 2003 at 07:37:06AM +0800, Jonathan Tse wrote:
> Is there anyone using it in live traffic with reasonable load? I would like
> to hear your comment on the stablity of Junos 5.6.

We've used it on an M5 that runs gigE, ATM OC3c, tunnel PIC, OSPF,
~50 BGP peers, IP multicast (running as the RP in a PIM domain), and
load on the gig side has risen to a couple hundred megs at least
(with filters/policiers slowing it down on the OC3c side) without
any major issues.

The one potential problem you'll want to be careful with is the
following:

PSN-2003-01-065
Memory leak during IP address resolution in JUNOS software Release 5.6R1

Note, this pertains only to 5.6 code built before January 17, 2003.

John

Hi John,

Thanks. What is the latest version now (with the patch number)?

regards,
Jonathan.

----- Original Message -----
From: "John Kristoff" <jtk@depaul.edu>
To: <juniper-nsp@puck.nether.net>
Sent: Tuesday, February 11, 2003 7:47 AM
Subject: Re: [j-nsp] Junos 5.6 any field trial?


> On Tue, Feb 11, 2003 at 07:37:06AM +0800, Jonathan Tse wrote:
> > Is there anyone using it in live traffic with reasonable load? I would
like
> > to hear your comment on the stablity of Junos 5.6.
>
> We've used it on an M5 that runs gigE, ATM OC3c, tunnel PIC, OSPF,
> ~50 BGP peers, IP multicast (running as the RP in a PIM domain), and
> load on the gig side has risen to a couple hundred megs at least
> (with filters/policiers slowing it down on the OC3c side) without
> any major issues.
>
> The one potential problem you'll want to be careful with is the
> following:
>
> PSN-2003-01-065
> Memory leak during IP address resolution in JUNOS software Release 5.6R1
>
> Note, this pertains only to 5.6 code built before January 17, 2003.
>
> John
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>

On 2/10/03 6:47 PM, "John Kristoff" <jtk@depaul.edu> wrote:

> The one potential problem you'll want to be careful with is the
> following:
>
> PSN-2003-01-065
> Memory leak during IP address resolution in JUNOS software Release 5.6R1
>
> Note, this pertains only to 5.6 code built before January 17, 2003.
>
> John

Have you actually seen this happen?

Do you know what the exact situation that leads to this is?

--Phil

For a couple of weeks now, we've been running 5.6R1.3 (this is the most
current version AFAIK) with a dozen BGP sessions, and OC-3c and OC-12c and
GigE interfaces on an M20. Smooth as can be.

We don't do a lot of the weirder stuff, though, so there could be bugs
lurking somewhere, you never know. In my experience, it's always better
than Cisco.

Kevin

> Is there anyone using it in live traffic with reasonable load? I would like
> to hear your comment on the stablity of Junos 5.6.
>
> thank you!
>
> regards,
> jonathan.
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>

hi,

thanks for all the replied!

regards,
Jonathan.

----- Original Message -----
From: <sigma@smx.pair.com>
To: <juniper-nsp@puck.nether.net>
Sent: Tuesday, February 11, 2003 12:37 PM
Subject: Re: [j-nsp] Junos 5.6 any field trial?


>
> For a couple of weeks now, we've been running 5.6R1.3 (this is the most
> current version AFAIK) with a dozen BGP sessions, and OC-3c and OC-12c and
> GigE interfaces on an M20. Smooth as can be.
>
> We don't do a lot of the weirder stuff, though, so there could be bugs
> lurking somewhere, you never know. In my experience, it's always better
> than Cisco.
>
> Kevin
>
> > Is there anyone using it in live traffic with reasonable load? I would
like
> > to hear your comment on the stablity of Junos 5.6.
> >
> > thank you!
> >
> > regards,
> > jonathan.
> >
> > _______________________________________________
> > juniper-nsp mailing list juniper-nsp@puck.nether.net
> > http://puck.nether.net/mailman/listinfo/juniper-nsp
> >
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>

On Mon, Feb 10, 2003 at 08:34:39PM -0500, Phil Rosenthal wrote:
> Have you actually seen this happen?
> Do you know what the exact situation that leads to this is?

I have not see any problems in my situation. The reason being that
the gig interface connected to a subnet that includes only one other
host and traffic between the two was continuous. Therefore, the
ARP table is always kept current, avoiding new ARP requests.

Not looking at the Juniper doc, I believe each new ARP entry eats
16 bytes of memory. Given enough ARPs, you can guess some likely
scenarios.

John