On 2010-09-09 19:17, Digimer wrote: > On 10-09-09 01:06 PM, Adam Gandelman wrote: >> On 09/09/2010 07:18 AM, Digimer wrote: >>> vgcreate -c y drbd_vg0 /dev/drbd0 >>> Found duplicate PV 4BbAHN2mFIP5pJsUE1ktdjSBwgFegmaG: using /dev/md3 >>> not /dev/drbd0 >>> Error locking on node <node name>: Command timed out >>> Can't get lock for orphan PVs >>> >> You're essentially not using DRBD anymore and now running CLVM commands >> on a non-shared (local) disk. Adjust your filters in lvm.conf to >> ignore those underlying devices and try again. >> >> Hope that helps, > > That was the trick. However, I've not been able to fathom a filter > argument that simply ignores '/dev/md3'. It's working now by only > accepting DRBD and rejecting the rest. > > Simply 'a|/dev/drbd*|' or 'r|/dev/md*|' doesn't seem to be enough. It > makes me a bit uncomfortable knowing that I'll have to remember to > explicitly allow all future devices, rather than specifically ignoring > the troubling one. Any suggestions there?
This is just a guess, but maybe it's that "md component detection" magic in LVM getting in the way here? Does the behavior of pvscan change in any way if you play with md_component_detection or some of the other md_* settings that lvm.conf supports? Cheers, Florian
signature.asc
Description: OpenPGP digital signature
_______________________________________________ drbd-user mailing list [email protected] http://lists.linbit.com/mailman/listinfo/drbd-user
