Re: PM3 stopping?

Tom Samplonius (tom@sdf.com)
Tue, 8 Jul 1997 08:23:24 -0700 (PDT)

On Tue, 8 Jul 1997, S. Van Steen wrote:

> > On Mon, 7 Jul 1997, Jacob H. Suter wrote:
> >
> > > I've found this trick... ComOS apparently has a hole when it
> > > comes to large ICMP packets.. Filter all ICMP packets to the ethernet
> > > interface of the PM from both dialups and the ethernet side of things.
> > > It'll save yourself some administration hassle (and screw up traceroutes -
> > > but Livingston's already do that to 90% of the installed system base of
> > > the world anyways).
>
> Indeed, as this is what a PM3 says after someone does a "ping -s 12000" for
> about a minute:
>
> isdn_dpktout: **** NO BUFFS ****
> isdn_dpktout: **** NO BUFFS ****
> isdn_dpktout: **** NO BUFFS ****
>
> pm> ver
> Livingston PortMaster PM-3 ComOS 3.5c6
> System uptime is 1 minutes
> pm> sh mem
> System memory 4194304 bytes - 3037868 used, 1156436 available
> 96:1 1152:1 128:0 640:1 80:1 208:1 112:3 272:1 160:2 16:3 48:0 32:12
> System nbufs 2800 - 37 used, 2763 available
>
> The ICMP echo requests will eat *ALL* your nbufs but will not be answered :-(
>
> --Steven.

Read release notes for b20.

Tom