Hello, I'm really clueless about the following problem: I'm running a Sparc E250 server since december last year under potato. Two weeks ago the machine canceled network connection (no ping) and dosn't had any output on ttyS0 console. I switched it off :-( and on again some times but it doesn't helped.
When Sun service came the damn box worked again (doing several minutes e2fsck - hey are there any patches for XFS ready for production ???) and started working again. The only unusual thing logcheck reported was: Unusual System Events =-=-=-=-=-=-=-=-=-=-= Apr 4 08:38:10 bse kernel: power: Error, cannot register IRQ handler. Apr 4 08:38:10 bse kernel: request_irq: Old style IRQ registry attempt from 00000000005c7a68, irq 00000000. Looking at the date you see that it happened again. I found out that rebooting depends from the position of the keyswitch: If I switch the key to "Diagnostics" it doesn't start, but no failure LED is on at the Control Panel. If I switch on to "Power-On" position some seconds after switching on the "General fault" LED lights steadily. Kind of alarming but the service man thinks it is a Linux problem. If I would boot Solaris it would be off. I'm really clueles what to do now. Does anybody have any experiences which could give me a hint. I have to admit that I havn't observed the state of the "General fault" LED in the days before the first failure, but I can't imagine that it would be on. Moreover why isn't the box starting in the "Diagnostics"-mode? Could Linux be related to this LED / switching randomly off trouble. As I said, I'm using ~> uname -a Linux bse 2.2.18pre21 #1 SMP Tue Dec 19 22:53:19 CET 2000 sparc64 unknown on a pure potato machine with security and proposed-updates. Kind regards Andreas. -- We have joy, we have fun, we have Linux on our Sun.