[DRBD-user] attempt to access beyond end of device

Alan Robertson alanr at unix.sh
Fri Sep 17 10:20:18 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.


Andrew wrote:
> Alan Robertson wrote:
> 
>> Andrew wrote:
>>
>>> Alan Robertson wrote:
>>>
>>>> Andrew wrote:
>>>>
>>>>> Alan Robertson wrote:
>>>>>
>>>>>> Andrew wrote:
>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> I'm using DRBD version 0.6.12. Having some troubles with EXT3 
>>>>>>> reporting "attempt to access beyond end of device".
>>>>>>>
>>>>>>> I first noticed a problem when i tried to mkdir. :
>>>>>>>
>>>>>>> [root at jack www]$ mkdir test
>>>>>>> mkdir: cannot create directory `test': Input/output error
>>>>>>>
>>>>>>> I'm getting EXT3 errors in /var/log/messages such as :
>>>>>>>
>>>>>>> Sep 17 11:15:20 jack kernel: attempt to access beyond end of device
>>>>>>> Sep 17 11:15:20 jack kernel: 2b:00: rw=0, want=100270088, 
>>>>>>> limit=100179416
>>>>>>> Sep 17 11:15:20 jack kernel: EXT3-fs error (device drbd(43,0)): 
>>>>>>> read_inode_bitmap: Cannot read inode bitmap - block_group = 765, 
>>>>>>> inode_bitmap = 25067521
>>>>>>> Sep 17 11:15:20 jack kernel: EXT3-fs error (device drbd(43,0)) in 
>>>>>>> ext3_new_inode: IO failure
>>>>>>>
>>>>>>> [root at jack nai2]# df -k /data/
>>>>>>> Filesystem           1K-blocks      Used Available Use% Mounted on
>>>>>>> /dev/nb0             100179416   5002344  90088204   6% /data
>>>>>>>
>>>>>>> [root at jack nai2]# /sbin/fdisk -s /dev/nb0
>>>>>>> 100179416
>>>>>>>
>>>>>>> in drbd.conf:
>>>>>>>    disk-size = 100179416k
>>>>>>>
>>>>>>> So i'm just wondering if anyone has any ideas on what might be 
>>>>>>> wrong ?
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> It might be corrupted.
>>>>>>
>>>>>> When did you last run fsck on it?
>>>>>>
>>>>>>
>>>>> would have been a fair while .... maybe 3-4 months.... should I 
>>>>> give fsck a go? any particular options I should use with fsck ?
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Not that I know of.
>>>>
>>>> If taking this system down is difficult, you could even do an fsck 
>>>> -n with it mounted.  It might produce errors, but if you do it 
>>>> several times, and each time it gives the same errors, then it's 
>>>> probably worth unmounting it and checking it out the right way.
>>>
>>>
>>>
>>>
>>>
>>> when i run fsck ... i get :
>>>
>>> [nai2 at jack nai2]$ sudo /sbin/fsck -n /dev/nb0
>>> fsck 1.34 (25-Jul-2003)
>>> e2fsck 1.34 (25-Jul-2003)
>>> Warning!  /dev/nb0 is mounted.
>>> Warning: skipping journal recovery because doing a read-only 
>>> filesystem check.
>>> The filesystem size (according to the superblock) is 25444352 blocks
>>> The physical size of the device is 25044854 blocks
>>> Either the superblock or the partition table is likely to be corrupt!
>>> Abort? no
>>>
>>> /dev/nb0 contains a file system with errors, check forced.
>>> Pass 1: Checking inodes, blocks, and sizes
>>> Deleted inode 1015829 has zero dtime.  Fix? no
>>
>>
>>
>> This is more or less what I expected...
>>
>> I suspect you made the filesystem (mkfs) on the lower layer (real 
>> disk) device.
> 
> 
> 
> Thanks for your help !
> 
> hmmmm ok .... I didn't setup the filesystem myself, so I'm not sure how 
> it was created. Basically your saying that when the filesystem was setup 
> it may have been done incorrectly by :
> 
> mkfs -b 4096 /dev/drbd0
> 
> instead of :
> 
> mkfs -b 4096 /dev/nb0

Right.  There are other ways to do this, but this is the most obvious 
error.  Another would be that someone came in and biffed the DRBD 
configuration to make it smaller -- when both sides were down.

I think it really unlikely that somehow the filesystem size on disk got 
stepped on without damaging much of anything else.

> Is that correct ?  If that were the case then, is this the kind of 
> errors you would expect to occur after a few months of seemingly 
> flawless operation ?
> Are there any other ways to work out whats going on here. I can't really 
> take the machines offline at the moment, nor can i umount the drbd 
> parititions till i know for sure what the problem is.

Your filesystem isn't very full yet.  It may just be the case that it's not 
had any occasion to try and allocate in this missing bit of disk before now.


-- 
     Alan Robertson <alanr at unix.sh>

"Openness is the foundation and preservative of friendship...  Let me claim 
from you at all times your undisguised opinions." - William Wilberforce



More information about the drbd-user mailing list