Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
5 drbd devices:
drbd0 20GB ext2
drbd1 20GB ext2
drbd2 400GB lvm
drbd3 500GB lvm
drbd4 500GB lvm
Both nodes connected and synced
Reboot both nodes
before turning on drbd, I ran the following on both nodes:
lvextend -L500GB /dev/DISK/HOME
start drbd on both nodes
both nodes complain about drbd2 not having meta-data and telling me to
run 'drbdadm create-md'.
Both nodes running Centos 4.6
kernel 2.6.9-67.0.15.ELsmp
drbd82-8.2.6-1.el4.centos
kmod-drbd82-smp-8.2.6-1.2.6.9_67.0.15.EL
Logs:
Jun 30 13:30:31 msddisk1 kernel: drbd2: disk( Diskless -> Attaching )
Jun 30 13:30:31 msddisk1 kernel: drbd2: Starting worker thread (from
cqueue/0 [5052])
Jun 30 13:30:31 msddisk1 kernel: drbd2: Error while reading metadata,
magic not found.
Jun 30 13:30:31 msddisk1 kernel: drbd2: disk( Attaching -> Diskless )
Jun 30 13:30:31 msddisk1 kernel: drbd2: drbd_bm_resize called with
capacity == 0
Jun 30 13:30:31 msddisk1 kernel: drbd2: worker terminated
Jun 30 13:30:31 msddisk1 kernel: drbd2: Terminating worker thread
Jun 30 13:30:31 msddisk1 kernel: drbd2: conn( StandAlone -> Unconnected )
Jun 30 13:30:31 msddisk1 kernel: drbd2: Starting worker thread (from
cqueue/0 [5052])
Jun 30 13:30:31 msddisk1 kernel: drbd2: Starting receiver thread (from
drbd2_worker [5091])
Jun 30 13:30:31 msddisk1 kernel: drbd2: receiver (re)started
Jun 30 13:30:31 msddisk1 kernel: drbd2: conn( Unconnected -> WFConnection )
Jun 30 13:30:42 msddisk1 kernel: drbd2: Handshake successful: Agreed
network protocol version 88
Jun 30 13:30:42 msddisk1 kernel: drbd2: conn( WFConnection ->
WFReportParams )
Jun 30 13:30:42 msddisk1 kernel: drbd2: Starting asender thread (from
drbd2_receiver [5092])
Jun 30 13:30:42 msddisk1 kernel: drbd2: data-integrity-alg: <not-used>
Jun 30 13:30:42 msddisk1 kernel: drbd2: some backing storage is needed
Jun 30 13:30:42 msddisk1 kernel: drbd2: conn( WFReportParams ->
Disconnecting )
Jun 30 13:30:42 msddisk1 kernel: drbd2: error receiving ReportSizes, l: 32!
Jun 30 13:30:42 msddisk1 kernel: drbd2: meta connection shut down by peer.
Jun 30 13:30:42 msddisk1 kernel: drbd2: asender terminated
Jun 30 13:30:42 msddisk1 kernel: drbd2: Terminating asender thread
Jun 30 13:30:42 msddisk1 kernel: drbd2: tl_clear()
Jun 30 13:30:42 msddisk1 kernel: drbd2: Connection closed
Jun 30 13:30:42 msddisk1 kernel: drbd2: conn( Disconnecting -> StandAlone )
Jun 30 13:30:42 msddisk1 kernel: drbd2: ASSERT( mdev->receiver.t_state
== None ) in
/home/buildcentos/rpmbuild/BUILD/drbd-8.2.6/_kmod_build_smp/drbd/drbd_main.c:2412
Jun 30 13:30:42 msddisk1 kernel: drbd2: drbd_bm_resize called with
capacity == 0
Jun 30 13:30:42 msddisk1 kernel: drbd2: worker terminated
Jun 30 13:30:42 msddisk1 kernel: drbd2: Terminating worker thread
Jun 30 13:30:42 msddisk1 kernel: drbd2: receiver terminated
Jun 30 13:30:42 msddisk1 kernel: drbd2: Terminating receiver thread
Jun 30 13:30:59 msddisk1 kernel: drbd2: State change failed: Refusing to
be Primary without at least one UpToDate disk
Jun 30 13:30:59 msddisk1 kernel: drbd2: state = { cs:StandAlone
st:Secondary/Unknown ds:Diskless/DUnknown r--- }
...
Florian Haas wrote:
>> When resizing lower device (LVM) offline, drbd no longer sees the
>> meta-data when restarted. I'm using internal meta-data. How can I avoid
>> this?
>>
>
> Shouldn't happen, and normally doesn't. Can you give us the exact sequence
> of commands you issued, and the DRBD error message you are getting in your
> kernel log?
>
> Cheers,
> Florian
>
>
>
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user
>
--
Jeff Gamsby
Center for X-Ray Optics
Materials Sciences Division
Lawrence Berkeley Laboratory
510.486.7783
jfgamsby at lbl.gov