[DRBD-user] primary/secondary problem

Gernot W. Schmied gernot.schmied at chello.at
Mon Jun 19 11:32:28 CEST 2006

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

nick wrote:
> Gernot W. Schmied wrote:
>> Lars Ellenberg wrote:
>>> / 2006-06-13 13:32:58 +0200
>>> \ Gernot W. Schmied:
>>>>>>>> setup (0.7.17 on both machines) on my primary machine suddenly
>>>>>>>> insists
>>>>>>>> to remain in "secondary" mode. Any idea why?
>>>>>>> well.
>>>>>>> what do you mean by "insist"?
>>>>>>> if you tell it to become primary, what exactly does happen?
>>>>>>> how do you try to do the "drbdadm primary all"?
>>>>>> Hmm, the primary node so far always remained primary after setting it
>>>>>> once regardless of the secondary being reachable and it maintained
>>>>>> that
>>>>>> state regardless of restarting drbd via /etc/init.d/drbd.
>>>>>> Now it always returns to secondary state when I stop/restart drbd
>>>>>> after
>>>>>> declaring it primary with "drbdadm primary all" before mounting my
>>>>>> file
>>>>>> systems. Maybe it is supposed to be that way but I'd like to
>>>>>> understand
>>>>>> why. Why does it return to secondary state when I stop drbd?
>>>>> I'd rather ask "what does switch drbd to primary".
>>>>> the init script (at least my init script) does not.
>>>> Mine does now. Should I conclude from your question that both
>>>> constituents of a drbd cluster start in secondary mode unless altered
>>>> via init script or heartbeat?
>>> if you insist on shooting yourself, just go ahead.
>> Very well, beautiful advice, witty, technically sound, one has to admire
>> the craftsmanship.
>> It strikes me as peculiar though that *some* folks here switch to
>> messiah-ego-mode as soon as somebody dares to say that he does not run
>> DRBD via heartbeat and sync-cluster. Of course I am running DRBD in
>> several heartbeat (crypto)clusters, makes the most sense anyway.
>> However, I was playing with an async configuration where I just power up
>> a remote secondary node (SOHO workstation) once a day as a replacement
>> for a daily tape backup, and tell you what, it works well and why
>> shouldn't it and it serves its purpose at the accepted risk of losing a
>> day's changes (same with the tape anyway).
>> So my question was related to an isolated primary node (secondary
>> powered down) which was declared primary once and why it insists to
>> return to secondary mode after restart of drbd.
>> Gernot
>> _______________________________________________
>> drbd-user mailing list
>> drbd-user at lists.linbit.com
>> http://lists.linbit.com/mailman/listinfo/drbd-user
> Just a simple question, why not run something like unison?
> http://www.cis.upenn.edu/~bcpierce/unison/
> I use it on my DRBD connected boxes for files in directories where the
> whole directory cannot be mirrored (due to configuration files present
> which need to remain unique)..
> You can run it over ssh, set it up as a cron job, and it works great for
>  remote sites cuz it will only send the changed data.
> For this type of problem I think using DRBD might be a bit of overkill.
> IMHO of course.
> Nick

DRBD is no overkill because it does not require a lot of attention and
it performes very well. I am using unison for async backup on notebooks
anyway, rsync for other purposes.


More information about the drbd-user mailing list