[PlanetCCRMA] RT Kernel Crash

Donald Steven t6sn7gt at aim.com
Fri Jul 2 02:56:29 PDT 2010


Thanks Tracey.  I've now had the crash with the stock Fedora kernel, so 
it's clearly not a ccrma-rt issue.  I suspect a daemon and will continue 
to work it.  It's definitely a fedora issue, as there are no problems 
with centos, opensuse, etc.

Don

On 07/02/2010 03:40 AM, Tracey Hytry wrote:
> Thanks Donald, and you too, Jonathan.
>
> I can't really make much sense of the logs but they seem to vaguely like things I've seen on one of the kernel lists.
>
> Both of your logs indicated non-tainted kernels, which is something the rt-kernel devs can use.
>
> Most of the time when I was seeing very strange problems on this machine;  usually typing "dmesg" on a term was enough to get the kernel dump.  I'd scroll up from the bottom and copy over everything from just above the kernel dump.  Also, there was always "/var/log/messages" to look at.
>
> I'm not sure what Fernando thinks about it, but I'm all for posting bug reports for the ccrma-rt-kernels here on the list.  I'm sure that the stuff that matters will get forwarded to the right people.
>
> Just try to remember that the kernel folks don't look into problems with kernels that are running proprietary drivers, so if your dump says it's "tainted" instead of "non-tainted", you are on your own!
>
> BTW, on this machine I could never turn in a bug report because it was tainted with the "nvidia" driver.  I had to run the machine with the driver because it was an "sli" card and the nouveau and nv driver would not work with it.  I finally went and replaced it with a newer, but less power hungry :) nvidia card that works with the nouveau driver.
>
> When I was having crashes it took forever to find what was wrong, and this with a stock fedora 26.32.xx kernel.  Every crash would point to something else.  I eventually found the problem after turning off all of the daemons I could, and testing, and turning stuff back on until I found that the "cpuspeed" service cannot be used on this machine.  In my case I think it may be because I'm using an early dual core amd64.
>
> Tracey.
>
> _______________________________________________
> PlanetCCRMA mailing list
> PlanetCCRMA at ccrma.stanford.edu
> http://ccrma-mail.stanford.edu/mailman/listinfo/planetccrma
>    



More information about the PlanetCCRMA mailing list