lrr@princeton.UUCP (09/24/86)
At Princeton University, we are experiencing a problem that we cannot explain. We have 6 750s, some running 4.2, some running 4.3, and an 8600 running 4.3, nothing unusual in peripherals, all ethernetted together. Every so often - once every few minutes - the 4.3 750s experience momentary grand pauses when not even characters are echoed by those machines. The 8600 shows no problem. Monitoring the ethernet with the traffic program from SUN shows nothing unusual, nor do the 750s show calm followed by massive paging/swapping. The 750s are essentially unused and are running just the standard daemons from the distribution tape. Any leads would be much appreciated. Larry Rogers Princeton University Department of Computer Science Engineering Quadrangle Building, Room C334 Princeton, NJ 08544 UUCP: princeton!lrr CSNET: lrr%princeton@CSnet-relay PHONE: 609 452 6483
hartley@uvm-gen.UUCP (Stephen J. Hartley) (09/29/86)
> At Princeton University, we are experiencing a problem that we cannot > explain. Every so often the 4.3 750s experience momentary grand pauses > when not even characters are echoed by those machines. We have had similar problems with our 4.3 BSD VAX 750 and 780, although it seems to happen only once or twice a month. We have had to reboot to recover from the "grand pause". We have not a clue to the cause of the problem.
lrr@princeton.UUCP (09/29/86)
The problem turned out to be a lack of buffers in the DEUNA. By asserting dedebug in the driver, we found that it was unable to put the incoming information anywhere. We changed the line #define NRCV 7 to #define NRCV 31 and all is well. The problem did not show up on the 8600 because it appears to be able to get the info from the DEUNA fast enough to never (?) run out of buffer space. Larry Rogers Princeton University Department of Computer Science Engineering Quadrangle Building, Room C334 Princeton, NJ 08544 UUCP: princeton!lrr CSNET: lrr%princeton@CSnet-relay BITNET: lrr%princeton@pucc.bitnet PHONE: 609 452 6483