1 |
* Rich Freeman <rich0@g.o> [150203 08:36]: |
2 |
> On Tue, Feb 3, 2015 at 8:14 AM, Alan Mackenzie <acm@×××.de> wrote: |
3 |
> > So, what was it that chewed up my RAID configuration so badly that |
4 |
> > /dev/md6 got renamed to /dev/md127? Can I change it back to /dev/md6, |
5 |
> > somehow? Do I need to bother? |
6 |
> |
7 |
> I ran into similar issues a while back. In my case some of my arrays |
8 |
> were using older metadata (which was required at the time to boot |
9 |
> without an initramfs). I suspect that either this metadata lacked the |
10 |
> info needed for a boot CD to assign the same ID, or perhaps the ID I |
11 |
> was using was already allocated somehow and the boot CD chose another |
12 |
> one and wrote that ID to the metadata so that it stuck. |
13 |
> |
14 |
> My solution was to move to using UUIDs or labels for everything and |
15 |
> not relying on array numbering. This of course requires an initramfs |
16 |
> - personally I've found Dracut to be the best one out there. It is |
17 |
> just far less prone to breakage when some update causes stuff to move |
18 |
> around like this. |
19 |
|
20 |
I also had the same problem a while ago and like Rich I started using |
21 |
UUIDs (actually I had started on another system where it mounted my |
22 |
/home partition as /tmp and rm -rf'd it during startup because of the |
23 |
/dev/md devices being scrambled around, but at that point I switched all |
24 |
my systems to UUIDs.) |
25 |
|
26 |
I also use dracut and aside from some problems with it starting up my |
27 |
raid arrays, it works well and I don't think much about it. |
28 |
|
29 |
Todd |