trying to understand setup-storage

[Michael Tautschnig]

[...]

> 
> After resolving a module related problem (see other thread), the
> installation almost completed. I made some errors in my configuration,
> though and wanted to re-run the installation, but without the two or
> three hours delay that are attributable to the creation and re-syncing
> of the software raid, if possible. Therefore, I changed the
> setup-storage configuration file to:
> 
> --------------------- cut
> disk_config sda bootable:1 preserve_reinstall:1,2
> primary         -                200    -       -
> primary         -                  0-   -       -
> 
> disk_config sdb bootable:1 preserve_reinstall:1,2
> primary         -                200    -       -
> primary         -                  0-   -       -
> 
> 
> disk_config raid preserve_reinstall:1,2
> raid1           /boot           sda1,sdb1  ext2  rw
> raid1           -               sda2,sdb2  -    -
> 
> 
> # disk_config lvm preserve_reinstall:1,2,3,4,5,6,7,8
> 
> disk_config lvm preserve_reinstall:1-8
> 
> vg uv0 md1
> 
> uv0-root        /               500    ext3    rw,errors=remount-ro
> uv0-swap        swap            4096    swap    sw
> uv0-usr         /usr           20480    ext3    rw
> uv0-var         /var            2048    ext3    rw
> uv0-qmail       /var/qmail        200    ext3    rw
> uv0-home       /home        60000-    ext3    rw
> uv0-var_log    /var/log        1024    ext3    rw
> uv0-qmail_queue    /var/qmail/queue  1024    ext3    rw
> --------------------- cut
> 
>  [ minor gripe: When I first tried to say this:
>    "preserve_reinstall:1,2,3,4,5,6,7,8", setup-storage errored out,
>    talking something about a syntax error. ]
> 

Actually, both variants are wrong for logical volumes: you must put the logical
volume names in there. Like this:

disk_config lvm preserve_reinstall:uv0-root,uv0-swap,uv0-usr,uv0-var,uv0-qmail,uv0-home,uv0-var_log,uv0-qmail_queue

Your attempt using 1-8 parses, but means "a volume group named 1 with a logical
volume named 8"; you don't have such things here.

> With this configuration, I now get another problem that I'm unsure
> about how to solve:
> 
> 
> Starting setup-storage 1.0.5
> Can't preserve /dev/md1 because it is not defined in the current config
> 
> 
> This I don't understand because I only added "preserve_reinstall
> attributes" to all partitions, but didn't really change anything.
> Running "mdadm-startall" gives me the two raid partitions back,
> although with errors:
> 
> 
> installclient# mdadm-startall 
> W: mdadm: using ARRAYs defined in existing mdadm.conf.
> Assembling MD array md0...failed (no devices found).
> Assembling MD array md1...failed (no devices found).
> Assembling MD array md2...failed (no devices found).
> Assembling MD array md3...failed (no devices found).
> Assembling MD array md1...done (started [2/2]).
> Assembling MD array md0...done (started [2/2]).
> Generating udev events for MD arrays...done.
> 
> 
> installclient# cat /proc/mdstat 
> Personalities : [raid1] 
> md1 : active raid1 sdb2[0] sda2[1]
>       488183104 blocks [2/2] [UU]
>       
> md0 : active raid1 sda1[0] sdb1[1]
>       200704 blocks [2/2] [UU]
>       
> unused devices: <none>
> 

The problem is that this array has not been auto-started. You could add the
mdadm-startall somewhere before setup-storage is called. Maybe we should even do
that within setup-storage. Most probably nobody ever tried preserve in the
context of SW raid before (including myself :-( ).

Best,
Michael

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
Url : http://lists.uni-koeln.de/pipermail/linux-fai/attachments/20091016/cada5c1d/attachment.bin