Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Wed, Nov 26, 2008 at 11:15:00AM +0100, Federico Simoncelli wrote: > Hi all. I noticed that promoting to primary a "ds:Consistent/DUnknown" > device automatically triggers a "/sbin/drbdadm outdate-peer". > > 1) During normal operation the status is: > > 0: cs:Connected st:Primary/Primary ds:UpToDate/UpToDate C r--- > > 2) Unplugging the cable the node 1 fences the node 2 and its status becomes: > > 0: cs:WFConnection st:Primary/Unknown ds:UpToDate/Outdated C r--- > > 3) Booting the node 2 (keeping the cable disconnected) and running > "drbdadm adjust all" (for debug purpose) I can see that on this side > the status is: > > 0: cs:WFConnection st:Secondary/Unknown ds:Consistent/DUnknown C r--- > > 4) Running the command "drbdadm sh-b-pri all" on node 2 (as the drbd > init script would do after adjust and wait-connect) the "/sbin/drbdadm > outdate-peer" is triggered. > > The problem here is that the node 1 (the one holding the updated data) > get fenced without reason and it also risks to become later a > SyncTarget losing the updated data. > > I suppose that the problem is my outdate-peer which just fences the > node without actually outdate it. > Even if this is the reason I think that a "Consistent" device should > not be allowed to become primary (fencing the remote node) without any > confirmation. the "confirmation" is supposed to come from the outdate-peer handler. a current Primary will refuse to be Outdated, the outdate would fail, denying the promotion of the merely Consistent one. > What do you think about this? Am I missing something? > > My current configuration: > > become-primary-on both; > allow-two-primaries; > after-sb-0pri discard-zero-changes; > after-sb-1pri discard-secondary; > after-sb-2pri disconnect; > fencing resource-and-stonith; > outdate-peer "/usr/lib/drbd/obliterate-peer.sh"; > > The "/usr/lib/drbd/obliterate-peer.sh" script is available at: > > http://people.redhat.com/lhh/obliterate > > (It basically fences the other node using "fence_node" and it returns > the exit code 7 on success) > > -- > Federico. > _______________________________________________ > drbd-user mailing list > drbd-user at lists.linbit.com > http://lists.linbit.com/mailman/listinfo/drbd-user > -- : 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