Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Il 25/03/2017 11:02, Den ha scritto: > Hi, Hi Den, > I'm testing a 3 node cluster, fresh installed with linbit repo for pve > with last versions Me too (even if mine is not fresh installed ;-) ) > I have configured LVM as backend storage and set redundancy to 1 for now. Do you mean LVL or LVM-thin ? > So when I create a VM on node1, for example, could happen that the resource > is allocated on node2 and i get a diskless resource on node1, but this > not seem to be a problem. It is, if it is not intended for pve1 to be a satellite node. Diskless is not a normal condition. > The problem is restoring adn creating VMs, I had a similar problem few days ago, try the workaround that Yannis provided me on this list: http://lists.linbit.com/pipermail/drbd-user/2017-March/023108.html > here some log: > 1) restoring a VM on pve215 It seems that restore is being attempted and fails because it is Diskless (no backing device) > 2) restoring a VM on pve215 What is the difference between 1) and 2) ? If you tried 2 after 1), it seems that storage is still locked because of former failed attempt. > 3) > If I try to create a new VM, often happens that drbdmanage miss something: > > root at pve214:~# drbdmanage list-assignments > +-----------------------------------------------------------------------------------------------------------+ > > | Node | Resource | Vol ID | | > State | > |------------------------------------------------------------------------------------------------------------| > > | pve216 | vm-102-disk-1 | * | > | pending actions: commission | > | pve216 | vm-102-disk-1 | 0 | | > pending actions: commission, attach | Same happens to me, see: http://lists.linbit.com/pipermail/drbd-user/2017-March/023117.html rob