[DRBD-user] Kernel 2.6.20+ Drbd 8.0.0 (2738M): BUG: unable tohandle kernel NULL pointer

Jeronimo Zucco jczucco at ucs.br
Thu Feb 15 20:12:27 CET 2007

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Lars Ellenberg escreveu:
> / 2007-02-15 08:50:06 +0100
> \ Martin Bene:
>   
>> Hi Lars,
>>
>>     
>>>> I've seen the above bug several times, most recently after hard
>>>> resetting a node that was primary for one of the devices.
>>>>         
>>> as this happened in the dispatcher of the connector, I 
>>> suspect that the kernel configuration you built the drbd module
>>>       
>> against 
>>     
>>> does not match your running kernel.
>>>       
>> Ok, recompiling , installing & booting kernel/modules/drbd to be 100%
>> sure everything matches. 
>>
>>     
>>> or maybe there is also the CONNECTOR missing from your kernel,
>>> and for some reason our detection magic did not prevent to
>>> build the "built-in-backport", with non-matching netlink-ABI.
>>>       
>> Nope, I can definitely exclude that possibility:
>>
>> # zcat /proc/config.gz  |grep CONN
>> CONFIG_CONNECTOR=y
>>
>> Let's see if I can still get the same error or if it's gone away; 
>>
>> Nope, kernel/modules/drbd recompile + install didn't change a thing,
>> same error still. Next test: go back to 8.0.0 release (2713) instead of
>> currenct svn (2747M)...
>>
>> Nope, also no change, still crashes.
>>
>> Circumstances may be somewhat unusual:
>>
>> * start drbd on both nodes. Status connected, both devices on both nodes
>> are secondary.
>> * start heartbeat on just one node.
>>
>> Result: Heartbeat powers off node and switches drbd0 secondary =>
>> primary right next, resulting in kernel bug. BTW, doesn't happen
>> allways.
>>     
This issue happened to me once. I'm using vanilla linux 2.6.20 and drdb 
current from SVN, heartbeat 2.0.8.

>
> interessting.
> we may have an interessting race condition here :)
>
>
>   
>> Strange - this error happens pretty consistently when running via
>> heartbeat, I haven't been ablt to reproduce outside heartbeat though. 
>>
>> Any further hints what'S going on are greatly apreciated :-)
>>     
>
> does going back to kernel 2.6.19.3 or 2.6.16.something help?
>
>   


-- 
Jeronimo Zucco
LPIC-1 Linux Professional Institute Certified
Núcleo de Processamento de Dados
Universidade de Caxias do Sul

http://jczucco.blogspot.com




More information about the drbd-user mailing list