On Thu, 17 Apr 2008 15:40:03 +0100
Hugo Mills <hugo@???> wrote:
> > My understanding of Raid1 from
> > http://tldp.org/HOWTO/html_single/Software-RAID-HOWTO/
> >
> > is that each disk is an exact mirror of the information on one
> > disk on the other disk(s). So that if I can get at any one disk I
> > can copy off the geneweb data in /var/lib/geneweb and save it to
> > the zip drive which I have already checked can be mounted.
>
> OK. Close enough -- but in this case, not quite true (see below).
Yup I can see that now but I thought I'd put all 3 disks into the
array without having a spare.
<snipped>
> > Update Time : Wed Apr 16 2008
> > State : Clean
> > Active Devices : 2
> > Working Devices : 3
> > Failed Devices : 0
> > Spare Devices : 1
>
> You have two devices in the arracy actually storing data, and one
> "hot spare" -- spun up, but not participating in the RAID array at
> the moment.
>
> > then further on active sync /dev/sda1
> > active sync /dev/sdb1
> > spare /dev/sdc1
>
> /dev/sdc1 is the hot spare -- it won't contain any
> data. /dev/sda1 and /dev/sdb1 have the data on.
>
> > using fsck -l each disk is reported as
> > Device Boot /dev/sd*1 start 1 End 1111 Id fd System Linux Raid
> > autodetect
>
> You didn't answer my second question, regarding the content of
> /dev/md0 -- is this a single filesystem on its own, or is there a
> partition table on it, or LVM?
that is cos I didn't read the question carefully enough :-(
Well it won't have LVM for reasons mentioned in earlier email. So far
as I can recall I did an <mdraid --create /dev/md0> or similar and
then added drives to it.
I would have followed instructions similar to the Howto I mentioned
earlier and I don't think that talks about creating file systems on
the /dev/md0 device, I'd have to have a careful read of it to be sure.
Does this answer the question?
mdadm --examine /dev/md0
mdadm: No md superblock detected on /dev/md0
--
John Lewis
Debian (Sid) & the GeneWeb genealogical data server
