Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Mon, Oct 20, 2008 at 01:15:41PM -0700, Jeffrey Froman wrote: > On Friday 17 October 2008 03:54:08 am Lars Ellenberg wrote: > > simple fix: just DO NOT explicitly set a size. > > it is only useful in very special cases if at all. > > I am under the impression that our case requires (or required) the > size parameter. We encountered overrun errors when initially syncing > our nodes, which contained a pre-existing filesystem, with hugely > different sized disks underneath. Adding the size parameter to > restrict DRBD to a size slightly smaller than the smaller disk > cleared up the problem. > > Now that both nodes are sync'd, maybe it's safe to remove this > parameter? Or maybe the overrun problems we had were specific to > drbd-7.x, which is what we were using at the time. We are currently > using version 8.2.5. most likely your "overrun problems" where due to the fact that you had a pre-existing file system on that disk, and truncated it by overwriting the last part with drbd meta data. reducing the drbd size won't even help. as long as you mkfs /dev/drbd (after both nodes have seen each other for the first time), all is well. > If we have to keep the size parameter, it sounds like the only thing > broken is "drbdadm adjust", which we can work around by performing a > full restart on each node when it is in the secondary role. you don't need to keep the size parameter. -- : 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