[DRBD-user] "syncer" crash when doing full resync

Philipp Reisner philipp.reisner at linbit.com
Thu Sep 23 15:30:09 CEST 2004

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


On Thursday 23 September 2004 15:19, you wrote:
> Hi Philipp,
>
> Thank you for your comments.  I will assume that DRBD-0.8 is at least 6
> months out and I am in need of a work around.  This means that I will
> probably try and implement your suggested fix in 0.6.12 and then send it to
> the development list for comments.
>
> Do you have any pointers for me (prototype) that may save me time as I look 
to tackle this issue.  My company will not allow me to use drbd in production 
unless I can demonstrate a stable subsystem.  I appreciate any help.  
Following I have attached the section from 0.8 roadmap that deals with this 
issue:
> > 5 It is possible that a secondary node crashes a primary by
> >   returning invalid block_ids in ACK packets. [This might be
> >   either caused by faulty hardware, or by a hostile modification
> >   of DRBD on the secondary node]
> >
> >   Proposed solution:
> >
> >  Extend the block_id field. (currently 64 bit) by at least
> >  32 bits (64?) . (=block_id_chk field). The primary node
> >  stores an encrypted (random key, changes every 15 minutes...)
> >  checksum (=signature) in the second field.
> >
> >  The secondary node can not fake (either intentionally or
> >  unintentionally) these signature.
> >
> >  The primary node will only dereference the block_id pointers
> >  if the signature is right.
>
> Best Regards,
> Moiz
>

*  Why do you want to use drbd-0.6.x ? Why do you not use drbd-0.7.x ?

*  You have to fix the RAM of the secondary machine.
   (In case you really have the issue you reference to)
   Maybe you should post that OOPS ...
   
   The item 5 of drbd-0.8 roadmap will have the effect that the
   primary node will write something like "Got invalid block-id"
   to its Syslog and disconnect. 
   --> You have to fix the RAM of the secondary...

*  Ever considered that it might be chaper to ask DRBD's authors
   to do that featur for drbd-0.7 (by getting in contact with 
   office at linbit.com) than understanding the internals of DRBD
   by yourself ?

   Just in case you need to protect yourself agains failing 
   RAMs in your machines :)

-Philipp

-- 
: Dipl-Ing Philipp Reisner                      Tel +43-1-8178292-50 :
: LINBIT Information Technologies GmbH          Fax +43-1-8178292-82 :
: Schönbrunnerstr 244, 1120 Vienna, Austria    http://www.linbit.com :



More information about the drbd-user mailing list