Eric_Kolotyluk@cc.sfu.ca (09/26/88)
We have a number of KFPS-2 boxes and one KFPS-1 box. All boxes have been
upgraded with the capacitor fix and new ROMs. We're running etalkgw.srec
Version 1.0 (12-Nov-87).
The problem is the boxes routinely die. After restarting the box, by
cycling the power, the box will work fine for anywhere from a few hours
to a few days. Eventually the box seems to degrade, i.e. more and more
packets seem to be lost, until no more packets make it through. By this
time the the FastPath Manager can't even talk to the box and it's time
to recycle the power.
I have two supsicions but would like some comments from others who have
similar experience:
1) We have a very busy ethernet because it's *large* with many mac-level
bridges. The average load on the ethernet is 30 to 50 packets/second,
with peaks of 300 to 500 packets/second when we get ARP storms. Could
this be affecting the K-box adversely?
2) There are still bugs in the etalkgw.srec code???ntitley@zaphod.axion.bt.co.uk (nigel titley) (09/29/88)
From article <1212668@cc.sfu.ca>, by Eric_Kolotyluk@cc.sfu.ca: > We have a number of KFPS-2 boxes and one KFPS-1 box. All boxes have been > upgraded with the capacitor fix and new ROMs. We're running etalkgw.srec > Version 1.0 (12-Nov-87). > > The problem is the boxes routinely die. After restarting the box, by We have a very similar problem. Our network is also busy and we lose boxes every few days. It definitely seems related to network loading. We had a fault with a repeater a couple of weeks ago which resulted in a very high collision rate. The boxes were going out at much more frequent intervals. I think there are still some bugs in the firmware and/or etalkgw. Things were improved by the capacitor fix and new ROMS but they still go down, although less frequently. Our local distributor, JPY Associates, have been very sympathetic but not much help. The best they can do is get on to the suppliers in the States and they don't seem to get much joy. If you find what's causing it please let us know. Email: NTitley@axion.bt.co.uk Snail: British Telecom Research labs, Martlesham Heath, Ipswich, Suffolk, UK "Well, I'm disenchanted too. We're all disenchanted." (James Thurber)