NETROM query

[<DEVANS%COLOLASP.BITNET@CORNELLC.cit.cornell.edu>]

I observed the following behaviour several times over the weekend. It
does not seem reasonable, but is it correct?

The situation:
My system wants to send its first IP packet to WS5N, a system a long
way away, but accessible over the NETROM. Execution of "netrom route
info ws5n" shows there to be two possible nearest neighbours from my
station, N5OP (a NOS pseudo-netrom node) and FNL, a real-live NETROM
box. The route quality to N5OP is slightly higher.

What happens:
My system starts sending SABMs to N5OP. This is as I would expect.
However, it turns out that N5OP has just taken his box off the air,
so no UAs come back from him. Eventually, after a number of retries,
my SABMs cease. AND THAT'S ALL. The alternative route, through
FNL, is never tried.

I saw this happen several times over the weekend. I do not know if the
failure to try the second route occurs under other circumstances (e.g.
NETROM connects from the mailbox), but it does seem to be consistent
in the circumstances above. I never saw a case where a second possible
route was tried.

As I said before, this does not seem reasonable, but the more important
question is, is it correct?

[Readers who are still awake will see the similarity between this
and the parallel IP and AX.25 level routing which has been discussed
here previously....]

  Doc