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, Feb 18, 2016 at 12:29:49PM +0100, Roland Kammerer wrote: > On Thu, Feb 18, 2016 at 10:01:02PM +1100, Adam Goryachev wrote: > > While this really is very simple, and it looks awesome, the issue is that > > most people will have existing data in existing drbd storage backed by > > physical block devices. How do you propose this existing data can be moved? > > > > Based on what you have said above, I really only see one option: > > > > node1-n:$ vgcreate drbdpool ... > > node1:$ drbdmanage init > > node1:$ drbdmange add-node node2 IP > > node1:$ drbdmange new-volume foo 10G --deploy 2 > > node1:$ dd if=/dev/drbdY/0 of=/dev/drbdX/0 > > Where /dev/drbdY is the old DRBD device that is not handled by drbdmanage and drbdX is the new one... > > > > Then adjust whatever utility is used to export the data to use the > > /dev/drbdX/0 instead of the old device. This just seems awkward, slow, and > > cumbersome..... > > > > Is there any better solution? > > No, there is no "import oldstuff" whatsoever feature. dd, restore a > backup, what ever. That is the state of the art. It looks easy, but as > always there are 10^23 corner cases and if we implement this "import > magic", people expect it to work for their - I can't even imagine what > crazy DRBD setups are out there - configurations. > > So again, it does not exist and I personally think if you jump to > DRBD9/drbdmanage land you have to deal with it, you are the admin that > knows your configuration best. Of course, eventually we need to provide a more or less convenient way to move 8.4 clusters to 9. And we will. But for now, if you have an existing, working, production DRBD 8.4 cluster, I see no reason at all to move that to DRBD 9, not in this year anyways. -- : Lars Ellenberg : http://www.LINBIT.com | Your Way to High Availability : DRBD, Linux-HA and Pacemaker support and consulting DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. __ please don't Cc me, but send to list -- I'm subscribed