Server with 3TB Crashing at boot

Michael Fuckner michael at fuckner.net
Mon Mar 16 10:17:50 UTC 2015


On 03/16/2015 10:17 AM, Konstantin Belousov wrote:
> On Sun, Mar 15, 2015 at 10:46:03PM +0100, Michael Fuckner wrote:
>>
>>> Send me two things:
>>> 1. output of acpidump -t
>>> 2. verbose dmesg of the boot with tunables set
>>> hw.dmar.enable=1
>>> hw.dmar.match_verbose=1
>>>
>>
>> http://dedi3.fuckner.net/~molli123/temp/head-dmar.log
>> http://dedi3.fuckner.net/~molli123/temp/acpidump.txt
>>
>> here are the logs you requested,
>
> Note that there are no fault logs from the DMAR blocking invalid DMA
> access.  Look at the release 10.1 logs, where there are lines like
>
> DMAR0: <unknown dev>:pci1:0:0 fault acc 0 adt 0x0 reason 0x5 addr 0
>
> and which I cannot find in your CURRENT kernel dmesg. I would say that
> the machine is fine, at least WRT the boot sequence.
>
> Then, I looked back at the previous logs from you, and both logs I
> have say 10.1-RELEASE-p6 in the herald. You overwrite the logs, so
> I cannot recheck the history, but it seems that you did not tested
> stable/10 ? There is currently no difference in DMAR code between head
> and stable/10.
Just generated, anything I should test in debugger?
Else I'll do the same with STABLE next.

http://dedi3.fuckner.net/~molli123/temp/head.txt

doing minimal network traffic causes the kernel to panic.

BTW: How to turn off empty lines in minicom logging?




More information about the freebsd-hackers mailing list