[DRBD-user] Kernel Oops, NULL pointer dereference during sync [was: Sync speed 0 K/sec]

Dan Didier dan at mapolce.com
Tue Aug 31 16:03:11 CEST 2004

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


 

> 
> / 2004-08-30 11:00:14 -0400
> \ Dan Didier:
> > This output is from primary system:
> > 
> > [root at linux1 root]# cat /proc/drbd
> > version: 0.7.2 (api:75/proto:74)
> > SVN Revision: 1483:1485 build by root at linux1.romestripsteel.local, 
> > 2004-08-30 09:48:47
> >  0: cs:SyncSource st:Primary/Secondary ld:Consistent
> >     ns:172 nr:0 dw:0 dr:1024 al:0 bm:0 lo:0 pe:2 ua:213 ap:0
> >         [>...................] sync'ed:  0.3% (1876852/1877020)K
> >         finish: 398:49:51 speed: 0 (0) K/sec
> >  1: cs:Unconfigured
> 
> though it probably is not relevant, I'd like to know the 
> kernel used, and hardware (smp? highmem?). Ah, I see, its 
> Redhats fabulous 2.4.20-31.9smp .
> still, is it highmem?

Lars,
Kernel is 2.4.20-31.9smp as you see.  It is SMP.  I'm not sure how to
check for highmem option.  Our system has a gig of memory and it sees it
all as it should, so maybe it is enabled.

> 
> can you reproduce this easily?

I can reproduce this very easily.  Every time I try to bring up the
second node we get this problem.
> if so, can you reproduce it still with a fresh svn checkout / 0.7.3 ?

I can't even compile it - here is what I get:

/usr/src/linux-2.4.20-31.9/include/linux/mm_inline.h:28: warning:
initialization makes integer from pointer without a cast
/usr/src/linux-2.4.20-31.9/include/linux/mm_inline.h:28: initializer
element is not computable at load time
make[3]: *** [drbd_worker.o] Error 1
make[3]: Leaving directory `/var/src/drbd-0.7.3/drbd'
make[2]: *** [_mod_/var/src/drbd-0.7.3/drbd] Error 2
make[2]: Leaving directory `/usr/src/linux-2.4.20-31.9'
make[1]: *** [kbuild] Error 2
make[1]: Leaving directory `/var/src/drbd-0.7.3/drbd'
make: *** [all] Error 2


> can you reproduce this with some 2.6. kernel?

Haven't tried this yet.  Would I be better off trying the default
redhat9 kernel?

Thanks,
Dan


> anyone else seen this with drbd 0.7.2 or later?
> 
> do we need to disable sendpage again, and fall back to sendmsg?
> 
> > EIP is at do_tcp_sendpages [kernel] 0x4e9 (2.4.20-31.9smp)
> > eax: 00000000   ebx: f76bbd80   ecx: 00000380   edx: 00000000
> > esi: 00000002   edi: f6cede3c   ebp: f6c99b80   esp: f6d1de84
> > ds: 0068   es: 0068   ss: 0068
> > Process drbd0_worker (pid: 2043, stackpage=f6d1d000)
> > Stack: f6d1def4 00000018 00000000 c0207bb8 f6c99b94 
> f6d1def4 00000018 
> > 00000000
> >        00000000 00000000 00001000 00000000 000005a8 
> 0000012c f6cedd00 
> > 00004000
> >        f6d1c000 00001000 c023105b f6cedd00 f6d1defc 
> fe2fd000 00001000 
> > 00004000
> > Call Trace:   [<c0207bb8>] sock_sendmsg [kernel] 0x78 (0xf6d1de90))
> > [<c023105b>] tcp_sendpage [kernel] 0xcb (0xf6d1decc)) [<f8a5fc57>] 
> > _drbd_send_page [drbd] 0x107 (0xf6d1def4)) [<f8a6026b>] 
> > drbd_send_block [drbd] 0x15b (0xf6d1df24)) [<f8a52a54>] 
> > w_e_end_rsdata_req [drbd] 0x64 (0xf6d1df74)) [<f8a53882>] 
> drbd_worker 
> > [drbd] 0x112 (0xf6d1df98)) [<f8a5eb88>] drbd_thread_setup 
> [drbd] 0x78 
> > (0xf6d1dfd0)) [<f8a5eb10>] drbd_thread_setup [drbd] 0x0 
> (0xf6d1dfe4)) 
> > [<c010759d>] kernel_thread_helper [kernel] 0x5 (0xf6d1dff0))
> > 
> > 
> > Code: f0 ff 42 14 8b 83 88 00 00 00 8b 6c 24 24 8d 04 f0 89 68 0c
> 
> 
> 	Lars Ellenberg
> 
> --
> please use the "List-Reply" function of your email client.
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user
> 



More information about the drbd-user mailing list