[DRBD-user] MAX DRBD DEVICES

Lars Ellenberg lars.ellenberg at linbit.com
Fri Dec 19 12:58:26 CET 2008

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


On Fri, Dec 19, 2008 at 11:20:56AM +0100, Matthias Böttcher wrote:
> 2008/12/17 Lars Ellenberg <lars.ellenberg at linbit.com>:
> > On Wed, Dec 17, 2008 at 03:58:01PM +0100, Maciej Bogucki wrote:
> >>
> >>> With this setup you need hardware of at least 4 physical machines, but
> >>> because of the machine (host) independent storage (iSCSI targets on
> >>> the redundant storage server) you are free to run and to switch the
> >>> Xen-DomU's on the physical Xen-Servers.
> >>>
> >>>
> >> You need only two phicycal servers at the setup. iSCSI target and client
> >> could be on the same machine.
> >
> > don't do that. really.
> > has deadlock potential.
> 
> Where ist the deadlock potential?

resource starvation deadlock.
file system io
entering block io
entering network stack
re-entering file system or block io.

it can get stuck if someone in the stack needs to allocate some memory
or request object someone else has to free/release first. queue
unplugging (pdflush, kjournald, etc.) can get stuck.
just don't do that.
especially with iSCSI target in fileio mode, that is a No-No!

> And if I remember correctly, it was said not to run LVM on top of DRBD
> on top of LVM.
> Is that correct?

no basic technical problem with that, other than that you have to be
careful with your lvm filters and activation scripts, and can seriously
confuse your admins (and all sort of distribution magic, e.g. initramfs
or udev).

-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.
__
please don't Cc me, but send to list   --   I'm subscribed



More information about the drbd-user mailing list