[DRBD-user] recovery from split-brain
jojy.varghese at gmail.com
Wed Oct 19 02:22:50 CEST 2011
Ok found out what the issue was. It took debugging the kernel module
to figure out what the issue was. I wish there was more user space
level error handling. The problem was that i was giving the command as
drbdadm -- -discard-my-data connect all
notice the "-" instead of "--". This was causing the kernel to
receive the data integrity configuration as some corrupted text.
On Tue, Oct 18, 2011 at 4:55 PM, Jojy Varghese <jojy.varghese at gmail.com> wrote:
> No when the error happened for the first time, the entry was not
> there. So i added one on both nodes.
> On Tue, Oct 18, 2011 at 4:45 PM, Pascal BERTON <pascal.berton3 at free.fr> wrote:
>> Ah! In that case, I'm not sure to understand what happened precisely : When you got the "'data-integrity-alg' not known" error message, did you have a 'data-integrity-alg' entry in your config file ? I understand you did not, but I would like to be sure!
>> Best regards,
>> -----Message d'origine-----
>> De : Jojy Varghese [mailto:jojy.varghese at gmail.com]
>> Envoyé : mercredi 19 octobre 2011 00:05
>> À : Pascal BERTON
>> Objet : Re: [DRBD-user] recovery from split-brain
>> I didnt have a data-integrity-alg to start with. Only when i saw this
>> error i added the config on one of the nodes and then on both :|
>> On Tue, Oct 18, 2011 at 2:58 PM, Pascal BERTON <pascal.berton3 at free.fr> wrote:
>>> I wouldn't swear, but I guess that if your kernel complains it doesn’t know
>>> this algorithm, it has good reasons to say so.
>>> Therefore, why not just suppressing the 'data-integrity-alg' entry within
>>> your config file, then issue a drbdadm adjust on both nodes to propagate
>>> this change and neutralize it, and finally issue your drbdadm command to
>>> make things run better.
>>> Just a guess...
>>> Best regards,
>>> -----Message d'origine-----
>>> De : drbd-user-bounces at lists.linbit.com
>>> [mailto:drbd-user-bounces at lists.linbit.com] De la part de Jojy Varghese
>>> Envoyé : mardi 18 octobre 2011 23:20
>>> À : drbd-user at lists.linbit.com
>>> Objet : [DRBD-user] recovery from split-brain
>>> How do you recover from a split-brain node with stale data? The
>>> status of the node is :
>>> StandAlone Primary/Unknown UpToDate/DUnknown
>>> I tried "drbdadm -- -discard-my-data connect all". But i get error:
>>> 1: Failure: (141) The 'data-integrity-alg' you specified is not known
>>> in the kernel. (Maybe you need to modprobe it, or modprobe hmac
>>> I verified that the hash i specified ( data-integrity-alg ) is
>>> supported by the kernel.
>>> drbd-user mailing list
>>> drbd-user at lists.linbit.com
More information about the drbd-user