Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Thu, Apr 27, 2006 at 11:54:10AM -0400, Charles Riley wrote: > > Greetings, > > In the "Data Redundancy By DRBD" paper, I see the following statement: > > "If you consider to stack DRBD on top of DRBD, think it over again. In a > fail-over case this will cause you more trouble than without it." I would guess they mean /dev/sda --> /dev/drbd0 --->/dev/drbd1 and the only motivation I can think of would be to try to replicate over more than two nodes. I understand they do a product for that for those who need it. the only consideration that comes to mind in the example you give is how many drbd volumes. see the list for the discussion of that. > I want to make sure I'm interpreting this correctly - > Does it mean that you can't/shouldn't consider a configuration like the > following? > > Physical device /dev/sda5 --> /dev/drbd0 mounted on /home/application > Physical device /dev/sdb5 --> /dev/drbd1 mounted on > /home/application/archive/das1 > Physical device /dev/sdc5 --> /dev/drbd2 mounted on > /home/application/archive/das2 > .... Etc. > > The application I'm attempting to run on top of drbd utilizes such a > setup. I'd rather not have to change the directory layout unless > absolutely necessary as it will require recoding the application, and > the archived data absolutely must exist on different physical devices > (there can be any number of das' attached, limited only by the number of > scsi controllers in the system) > > Comments/suggestions welcome - > > Many thanks, > > Charles > > > _______________________________________________ > drbd-user mailing list > drbd-user at lists.linbit.com > http://lists.linbit.com/mailman/listinfo/drbd-user -- Perl 6 will give you the big knob. -- Larry Wall