timelord@eos (G. Murdock Helms) (06/28/88)
In January of this year we purchased an SGI 4D60Turbo. Around February or March of this year, this strange error began popping up on the console: enp0: bad recieve packet length of 0 Usually we noticed this on entering the lab in the morning, late evening, or on a weekend, usually when no one had been using the machine for a while. The number of messages started at about 6, went up to about 50, and currently hover around 12 to 25. We don't always find these messages: it seems to be a somewhat sporadic problem that crops up for a week or two and dies out again. We had originally thought it was just something wrong with our machine, until I went to use a 4D70 in another lab (not ours) and found the screen full of these errors. The Hotline is puzzled by these errors, and neither our 2500T or 3120 Irises show any indication of burping in this manner. Has anyone else experienced or found a solution to this little mystery? G. Murdock Helms
vjs@rhyolite.SGI.COM (Vernon Schryver) (06/29/88)
In article <967@eos.UUCP>, timelord@eos (G. Murdock Helms) writes: > In January of this year we purchased an SGI 4D60Turbo. Around February or > March of this year, this strange error began popping up on the console: > > enp0: bad recieve packet length of 0 This means that an invalidly small packet came in from the ether. Future releases of the driver will have a printf based on: "if_enp%d: packet too small (length = %d)\n" We infrequently get flurries of these in our network. They mean something is broken or not yet working somewhere on the wire. Vernon Schryver Silicon Graphics vjs@sgi.com or {pyramid,allegra,research,sun,pacvax}!sgi!vjs