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

Harald Rinker harald.rinker at unitedprint.com
Wed Jan 16 00:10:57 CET 2008

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


Harald Rinker schrieb:
> Lars Ellenberg schrieb:
>> On Sat, Dec 29, 2007 at 03:18:32PM +0100, Lars Ellenberg wrote:
>>  
>>> On Sat, Dec 29, 2007 at 10:44:17AM +0100, Harald Rinker wrote:
>>>    
>>>> Hello Lars,
>>>>       
>>> please avoid html mail (at least on mailing lists).
>>> and top posting is not a good habbit either.
>>>
>>>    
>>>> I use Debian Etch Standard Kernel k713 2.6.18-4-amd64 and k641 
>>>> 2.6.18-5-686
>>>> and first i used drbd8 from backports now 8.0.8 from linbit website 
>>>> compiled as
>>>> module with the kernel headers
>>>>       
>>> in that case,  it could be a bug in drbd
>>>    somebody else should have noticed, there are lots of people using
>>>    that kernel, including us, with varying configurations.
>>>     
>>
>> hm.
>>
>> you DID hit a bug in drbd,
>> and no one (including us) did notice.
>>
>> that bug was only visible with protocol A,
>> and multipage bios.
>>
>> that means it would NOT show up with protocol B or C,
>> or with ext3, or when using any fs on top of lvm on top of drbd on 
>> top of anything,
>> or when using any fs on top of drbd on top of lvm or md raid0/5/10.
>>
>> it would show up e.g. when using xfs on top of drbd
>> on top of bare sata/scsi or md raid1.
>>
>> workaround: see above, for when it does NOT show up.
>>
>> the fix is there:
>>  http://git.drbd.org/?p=drbd-8.0.git;a=commitdiff;h=75a45802045fadaaaa926e58407726ace0317a0d 
>>
>>  http://git.drbd.org/?p=drbd-8.2.git;a=commitdiff;h=3a9e0f331315c487f79e32ff6e36b442d6d5369d 
>>
>>
>> and will be in the next releases.  yes, that means that we will likely
>> see yet an other release of drbd 8.2 within the next few days.
>>
>> embarrassing.
>>
>> cheers anyways,
>> and keep complaining!
>>
>>     ;->
>>
>>   
Hi list,

i have now new gb network cards use sync rate 5M  using protocol B set 
up node2
and make drbdadm adjust daten

/var/log/syslog say:
drbd0: Handshake successful: DRBD Network Protocol version 86
Jan 15 23:35:48 k713 kernel: drbd0: Peer authenticated using 32 bytes of 
'sha256' HMAC
Jan 15 23:35:48 k713 kernel: drbd0: conn( WFConnection -> WFReportParams )
Jan 15 23:35:48 k713 kernel: drbd0: Becoming sync target due to disk states.
Jan 15 23:35:48 k713 kernel: drbd0: Writing meta data super block now.
Jan 15 23:35:56 k713 kernel: drbd0: writing of bitmap took 1947 jiffies
Jan 15 23:35:56 k713 kernel: drbd0: 1499 GB (393203991 bits) marked 
out-of-sync by on disk bit-map.
Jan 15 23:35:56 k713 kernel: drbd0: Writing meta data super block now.
Jan 15 23:35:56 k713 kernel: drbd0: peer( Unknown -> Primary ) conn( 
WFReportParams -> WFBitMapT ) pdsk( DUnknown -> UpToDate )
Jan 15 23:35:56 k713 kernel: drbd0: Writing meta data super block now.

for about 5 Minutes it do nothing else and i must disconnect because the 
load of node1 the primary on goes over 20 and nfs gives no answer

drbdadm cstate daten tells WFBitMapT and no sync is running by cat 
/proc/drbd

dmesg on node2 (secondary) say:
drbd0: Handshake successful: DRBD Network Protocol version 86
drbd0: Peer authenticated using 32 bytes of 'sha256' HMAC
drbd0: conn( WFConnection -> WFReportParams )
drbd0: Becoming sync target due to disk states.
drbd0: Writing meta data super block now.
drbd0: writing of bitmap took 1947 jiffies
drbd0: 1499 GB (393203991 bits) marked out-of-sync by on disk bit-map.
drbd0: Writing meta data super block now.
drbd0: peer( Unknown -> Primary ) conn( WFReportParams -> WFBitMapT ) 
pdsk( DUnknown -> UpToDate )
drbd0: Writing meta data super block now.
eth2: no IPv6 routers present
drbd0: PingAck did not arrive in time.
drbd0: peer( Primary -> Unknown ) conn( WFBitMapT -> NetworkFailure ) 
pdsk( UpToDate -> DUnknown )
drbd0: asender terminated
drbd0: error receiving ReportBitMap, l: 4088!
drbd0: tl_clear()
drbd0: Connection closed
drbd0: Writing meta data super block now.
drbd0: conn( NetworkFailure -> Unconnected )  <- this one is because i 
take down eth2

I don´t understand. On my tests bevor drbd starts syncing immediately
How long did wrote of Metadata work

Cheers,
Harry








More information about the drbd-user mailing list