Nobuhito-san, GFS2 freezes I/O when it loses connection to a cluster node until fencing completes, which is probably what's keeping you from unmounting.
And, you are running yourself into split brain from which you need to manually recover (assuming you haven't configured split brain autorecovery policies). Please take a look at Lon Hohberger's DRBD GFS cookbook http://sourceware.org/cluster/wiki/DRBD_Cookbook if the information in the DRBD User's Guide is not sufficient. Best regards, Florian On 2009-06-29 13:03, Nobuhito Ogawa wrote: > Hello everyone, > > I set up DRBD Primary/Primary connection by DRBD, CMAN, LVM2 and GFS2. > And tested about network cut off by pull out the network cable. > After three minute, reconnect the network cable but DRBD don't reconnect. > And the DRBD disk can't unmount by commend "umount". > It needs machine shutdown. > I tryed command "drbdadm connect r0" but it didn't work. > Please teach me how can I reconnect the DRBD if you know. > > Installed package is > OS Centos 5.2 > drbd drbd82.x86_64 8.2.6-1 > kmod-drbd82.x86_64 8.2.6-2 > cman.x86_64 2.0.84-2 > lvm2.x86_64 2.02.32-4 > lvm2-cluster.x86_64 2.02.32-4 > gfs2-utils.x86_64 0.1.44-1 > > Thanks, > Nobuhito > > > > __________ ESET NOD32 Antivirus からの情報, ウイルス定義データベースのバージョン 4195 (20090629) > __________ > > このメッセージは ESET NOD32 Antivirus によって検査済みです。 > > http://canon-its.jp > > > _______________________________________________ > drbd-user mailing list > [email protected] > http://lists.linbit.com/mailman/listinfo/drbd-user >
signature.asc
Description: OpenPGP digital signature
_______________________________________________ drbd-user mailing list [email protected] http://lists.linbit.com/mailman/listinfo/drbd-user
