[DRBD-user] Starting DRBD on SLES9(64bit) with 2.5TB results in oops

Lars Ellenberg Lars.Ellenberg at linbit.com
Tue Aug 9 11:00:00 CEST 2005

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


/ 2005-08-08 17:12:24 -0400
\ Blake Atkins:
> Hello,
> 
> 	I am attempting to setup DRBD to mirror 2.5TB of direct attach storage.  I have no problems if I create a 500GB partiton on the 2.5TB array.  If I try to mirror the entire 2.5TB I get an oops as soon as I try to start DRBD:
> 
> server:~ # /etc/init.d/drbd start
> Starting DRBD resources:    [ d0 /etc/init.d/drbd: line 131:  6053 Killed                  $CMD
> 
> cmd /sbin/drbdsetup /dev/drbd0 disk /dev/sdc1 internal -1 --on-io-error=detach  failed!
> 
> Message from syslogd at server at Mon Aug  8 16:48:55 2005 ...
> server kernel: Oops: 0000 [1] SMP
> 
> Message from syslogd at server at Mon Aug  8 16:48:55 2005 ...
> server kernel: CR2: 0000000000000000
> 
> I am running SuSE SLES9 64-bit with all available updates on an AMD Opteron.  I have tried the drbd that came with SuSE as well as 0.7.11 with nearly identical results.  I have searched the list archives and it appears that I should be able to mirror up to 4TB with 4GB of RAM and a 64-bit architecture. 
> 
> Thanks very much,
> 
> Blake Atkins
> 
> Aug  8 16:48:54 cell4 kernel: drbd: module not supported by Novell, setting U taint flag.
> Aug  8 16:48:54 cell4 kernel: drbd: initialised. Version: 0.7.11 (api:77/proto:74)
> Aug  8 16:48:54 cell4 kernel: drbd: SVN Revision: 1807 build by root at cell4, 2005-08-08 16:25:00
> Aug  8 16:48:54 cell4 kernel: drbd: registered as block device major 147
> Aug  8 16:48:55 cell4 kernel: drbd0: Creating state block
> Aug  8 16:48:55 cell4 kernel: klogd 1.4.1, ---------- state change ----------
> Aug  8 16:48:55 cell4 kernel: drbd0: BM resizing failed. Leaving size unchanged at size = 2734838756 KB

       this is it: sort of "out of memory condition" that we do not
       handle gracefully (we used to BUG() right there.
       sorry.
       btw, it should have set its device size to zero,
       and current svn does it this way.  but we still cannot really
       handle it when allocation of the bitmap fails.
       btw, we need ~32MB of in kernel memory for each TB of storage.

> Aug  8 16:48:55 cell4 kernel: drbd0: Assuming that all blocks are out of sync (aka FullSync)
> Aug  8 16:48:55 cell4 kernel: drbd0: drbd_bm_set_all: (!(b && b->bm)) in /root/DRBD/drbd-0.7.11/drbd/drbd_bitmap.c:596

as a side node, at LINBITs currently largest deployment we effectively
have 15 TB on drbd (though still on several devices)...

-- 
: Lars Ellenberg                                  Tel +43-1-8178292-0  :
: LINBIT Information Technologies GmbH            Fax +43-1-8178292-82 :
: Schoenbrunner Str. 244, A-1120 Vienna/Europe   http://www.linbit.com :
__
please use the "List-Reply" function of your email client.



More information about the drbd-user mailing list