[DRBD-user] Question regarding read_inode_bitmap: cannot read inode bitmap -> IO failure

Lars Ellenberg lars.ellenberg at linbit.com
Mon Dec 6 11:59:49 CET 2004

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 03, 2004 at 12:52:00PM +0100, Roberto Nibali wrote:
> Hello,
> 
> Reading [1] on the subject, I learn that two things can cause that:
> 
> a. meta data is on top of en existing filesystem
> b. internal meta data is used but mkfs was done on lower level device
> 
> I'd like to share what I believe was our setup procedure for the DRBD 
> installation for a HA cluster:
> 
> 1. Configure two scsi disks of 74GB with 1 partition
> 2. Set label to "Linux Raid Autodectect"
> 3. Create raidtab
> 4. mkraid /dev/md0

[A]
|----------------- /dev/md0 --------------------|


> [... RAID is synching ...]

> 5. mke2fs -j /dev/md0
> 6. tunefs -c 0 -i 0 /dev/md0

[B]
|------ ext3 fs on /dev/md0 --------------------|


> 7. Create drbd.conf with device /dev/drbd0, disk /dev/md0 and 
> meta-disk-internal

[C]
|--- /dev/drbd0 --------------|-- meta data ----|
                                  of /dev/drbd0                
				  (do you see why we call it internal?)

> 8. Repeat 1-7 on second node
> 9. Invoke /etc/init.d/drbd start
>
> Do you see something in this shortened installation guideline that applies 
> to either a or b? If so could you kindly tell me where and how to address 
> this?
> 

comparing [B] with [C], ...
I guess you got the point.

> [1] http://marc.free.net.ph/message/20040919.215542.e4c54ac6.html



More information about the drbd-user mailing list