> > And *what* stuck W1 bug? The last W1 bug I was aware of was fixed several
> > revs ago, at 3.3.1. And everyone I know who had reported the bug with
> > the earlier code reported it fixed in the new code.
> >
> > Are you running recent code?
>
> I had a three month old pm2er30 running 3.3.3 (latest I could find) and it
> gets stuck spitting out abort errors about 4 times a day. I have had a linux
> box sitting next to it constantly monitoring the w1 error counters and reset
> the port once one increases rapidly. Telco replaced the entire copper feed (!)
> and has run every conceivable test including BERT, BPV, all kinds of traffic
> patterns, ... we got ONE SINGLE BIT ERROR in 48 hours ! I even installed a
> $5000 UPS unit cause "it could be the power" *sigh* I've got sites running
> 200+ days flawless without UPS. Even replaced the csu/dsu, v.35 cable, comos
> (tried 3.0.4, 3.1.4, 3.3.1, 3.3.1c1, 3.3.2, 3.3.2c1, 3.3.3 and I would have
> tried others if I had the images), portmaster, the hub and on the other side
> the cisco2503 including its v.35 cable and its IOS version (btw: each time it
> happens it nicely reports line up, protocol down) and hell, even the modems
> "maybe one is malfunctioning and causing some weird interference" all to
> no avail ...
>
> Until, in an act of sheer desperation I installed an 18 month old portmaster
> running comos 3.1.4 (yeah right, old code) it has been there for 5 days, saw
> exactly 1 abort and 0/1 frame errors and the port hasn't been stuck yet.
> I suppose I could own 2 portmasters with exactly the same hardware problem
> but the main issue is: if a linux box sitting next to it can deal with it then
> CERTAINLY something can be built in the comos to fix it as well (support has
> a very long mail from me describing exactly how it manifests itself which
> they answerd "have telco run tests") Mind you, my problem ain't solved yet,
> although I have not yet tried to recreate the problem on another site the
> moment I need to connect something to the W1 port of those two boxes remains
> one I await with fear.
>
> --Steven.
>
> PS: Even with all its problems I still prefer portmaster above cisco2511
> but this is getting a very subjective opinion ...
>
I talked to a Livingston Rep at the Interop in Las Vegas about this problem
earlier this year and was told that it was known internally that the PM2eR
couldn't handle a full T1 load. This momentary lockup problem started after
we upgraded our frame relay to T1 port speed. We gave up on that problem
after a while and replaced the W1 with a Cisco 2501 for a router. Needless to
say, the lock up problems have disappeared.
Turnando Fuad
NSNet