Growing a graid3?

[Eike Bernhardt]

On 11.05.2007, at 10:40, Pawel Jakub Dawidek wrote:

> It should be possible. First you swap one disk after another  
> waiting for
> synchronization. Once you have all new disks in place, you need to
> relabel your graid3 (it won't touch your data) and then growfs(8) it.
> You can try it with md(4) devices, eg:

*snip*

> 	# growfs /dev/raid3/test
> 	# mount -o ro /dev/raid3/test /mnt/test
> 	# dd if=/mnt/test/rand bs=1m count=12 | md5	# verify if file is ok  
> again
>
> Something like this.

Well, I just tried this with md devices, this works fine up to the  
growfs step:

# growfs /dev/raid3/test
growfs: we are not growing (8191->8191)

# graid3 list test
Geom name: test
State: COMPLETE
Components: 3
Flags: NONE
GenID: 0
SyncID: 1
ID: 2225454830
Zone64kFailed: 0
Zone64kRequested: 0
Zone16kFailed: 0
Zone16kRequested: 2
Zone4kFailed: 0
Zone4kRequested: 50
Providers:
1. Name: raid3/test
    Mediasize: 33553408 (32M)
    Sectorsize: 1024
    Mode: r0w0e0
*snip*

So something seems to be missing?

Thanks,
   Eike
-- 
Eike Bernhardt                                           ___ ___ _  _
   Institute for Science Networking Oldenburg GmbH       |_ _/ __| \| |
   email: eike.bernhardt at isn-oldenburg.de                 | |\__ \ .` |
   http://www.isn-oldenburg.de/                          |___|___/_|\_|