[DRBD-user] disaster recovery failover

Philipp Reisner philipp.reisner at linbit.com
Thu Aug 25 18:27:51 CEST 2005

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

Am Donnerstag, 25. August 2005 14:44 schrieb John S Little:
> Hi all,
> I saw the below post from "Subject: Re: [DRBD-user] 3 simple questions for
> a nifty setup" and it makes me wonder if I need to rethink part of our
> disaster recovery plan.
> We are building a disaster recovery center connected by fiber about 20
> miles from our location.  Where drbd is concerned I had planned on the
> primary here and the secondary at the dr site.  From what I read below
> this may not be such a hot idea.  Please let me know comments, insights
> and opinions..
> > buy it, and I'll tell you :->
> > well, basically, in case they'd need the "tertiary", what would happen
> is,
> > they phone us, and we do the tricky bit. *g*
> >
> > as I said, the "tertiary" is for desaster recovery, not for "high"
> > availability, so in case it is needed, there already has been a
> > desaster, and the additional latency of getting us involved is ok
> > compared to everything else.
> So we're looking at high availability with two machines locally and
> disaster recovery with a third at the dr site?
> What is the response time from the phone call?  Even though it is a file
> server server 90% of the documents are on it. We could stand some latency
> but as a hospital we would need to be up rather quickly.


The 3 node setup ... (DRBD+)
  Is a local DRBD pair at the main production site, with a third
  connection to the DR site. On the main production site you would
  usually run protocol C, but A to the DR site. 
  Also DRBD+ can use checksums to redurce the traffic generated
  by a resync on the link between the main site an the DR site.

  Btw, DRBD+ (read drbd plus) is a fork of the GPL'ed DRBD code.

The config is a bit more complex than the usual two node DRBD
setup. I guess it is a good idea if the config is done by LINBIT
for such a setup. I guess it is possible to do the knowhow transfer
about activating the DR site and switching back to the main site
in a workshop...     The mentioned customer decided, that they do
not want to have that knowhow, but to call us in that rare case, that
their two on-site data processing centers get destroyed....

: 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