[DRBD-user] DRBD sync starts over again and again

Lars Ellenberg lars.ellenberg at linbit.com
Sat Dec 29 02:00:45 CET 2007

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


On Fri, Dec 28, 2007 at 10:56:06PM +0100, Harald Rinker wrote:
> Hello Lars,
> i use now 8.0.8 on both side´s but it´s anyway the same
> 
> my syslog say´s
> Dec 28 22:31:50 k641 kernel: drbd0: Writing meta data super block now.
> Dec 28 22:31:51 k641 kernel: drbd0: conn( WFBitMapT -> WFSyncUUID )
> Dec 28 22:31:51 k641 kernel: drbd0: conn( WFSyncUUID -> SyncTarget )
> Dec 28 22:31:51 k641 kernel: drbd0: Began resync as SyncTarget (will sync
> 814504672 KB [203626168 bits set]).
> Dec 28 22:31:51 k641 kernel: drbd0: Writing meta data super block now.
> Dec 28 22:31:59 k641 kernel: drbd0: magic?? on data m: 0xffffffff c: 65535 l:
> 65535

the packet that drbd received consists of all ones,
where drbd expected it to be the standard header
"drbd magic, command, payload length".
strange.

what exact kernel are you using?
does that happen to have CONFIG_LBD disabled?

if that is the case, sector_t would become
"unsigned long" instead of "u64".

this is just an unverified guess,
but that might cause some of the internal calculations to overflow,
where we mix u64, unsigned long long and sector_t.

> Dec 28 22:31:59 k641 kernel: drbd0: ASSERT( mdev->state.conn < Connected ) in /
> usr/src/drbd-8.0.8/drbd/drbd_receiver.c:2703
> Dec 28 22:31:59 k641 kernel: drbd0: ASSERT( mdev->state.conn < Connected ) in /
> usr/src/drbd-8.0.8/drbd/drbd_receiver.c:3436
> Dec 28 22:31:59 k641 kernel: drbd0: asender terminated
> Dec 28 22:31:59 k641 kernel: drbd0: tl_clear()
> Dec 28 22:31:59 k641 kernel: drbd0: Connection closed
> Dec 28 22:31:59 k641 kernel: drbd0: peer( Primary -> Unknown ) conn( SyncTarget
> -> Unconnected ) pdsk( UpToDate -> DUnknown )
> Dec 28 22:31:59 k641 kernel: drbd0: drbdd_init: (mdev->receiver.t_state !=
> Restarting) in /usr/src/drbd-8.0.8/drbd/drbd_receiver.c:3117

and there you trigger some assertions. "interessting" :)

please provide the kernel log from the "very beginning",
i.e. from drbd0 unconfigured to the first connection attempt.
there should be something about "disk attaching" and "capacity"
and so on ...

also the corresponding (time synchronized) log of the same event
(connection attempt) from the other node.

-- 
: Lars Ellenberg                            Tel +43-1-8178292-55 :
: LINBIT Information Technologies GmbH      Fax +43-1-8178292-82 :
: Vivenotgasse 48, A-1120 Vienna/Europe    http://www.linbit.com :



More information about the drbd-user mailing list