Date: Sun, 11 Sep 2005 09:56:17 +0100 From: Robert Slade <bsd@bathnetworks.com> To: "freebsd-questions@freebsd.org" <freebsd-questions@freebsd.org> Subject: PL 5000 Random Reboots Message-ID: <1126428977.1803.48.camel@lmail.bathnetworks.co.uk>
next in thread | raw e-mail | index | archive | help
Hi, I have been having problems with random reboots when using the SMP Kernel on the above Quad processor machine. It occurs with bot 5.4 and 6.0B4. It does not happen with a single processor kernel, well not in the last 12 hours. Searching the mailing list and google has thrown up a number of possible causes: a. there was an issue with the SMP kernel under load in 5.3 but the errata says the fix was incorporated in 5.4; b. there appears to be some problems with with the ACPI with some machines. Unfortunately, the SMP kernel needs ACPI so turning it off is not an option. c. there appears to be an issue with some network cards and USB with interupt 'storms'. The machine has 2 network cards both identified in dmesg as Compaq Netelligent 10/100, athough only one is in use. I am getting tl0: tx underrun -- increasing tx threshold to 512 bytes when using vncserver to run xwindows with KDE which maybe related. It looks like the most likely cause is ACPI, SMP or both, but as ACPI relates to interups, the interup 'storm' issue could be impacting. However I do not know enough about the workings of ACPI etc to be able to judge this. It looks to me that there is a number of things I could do to isolate the problem: 1. try building a kernel without SMP but with ACPI to see what effect that has. Question is is this possible with 6.0B4? 2. if the above gives a problem try changing the network cards for another type. 3. fix the underrun issue but I am not sure how to do that. I would be grateful for any comments ideas on the way forward as the machine is quite slow without the SMP kernel. Thanks Rob
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1126428977.1803.48.camel>