Re: RAID 0+1 (Problem)


ECJ wrote:

Hello All,

I followed the steps in the link below but my RAID 10 array had a problem
after a reboot.
What I did was:
- constructed 2 RAID0s (3x73GB each)
- combined these 2 into a RAID1
- mkfs.ext3 /dev/md45 (the name of my RAID10 array)
- mounted it on /data

All seemed ok. It shows on /proc/mdstat. Can write to the partition where
I mounted it. But after a reboot, I got errors re this new RAID device
and
was prompted for the root password to enter repair mode.
Is there something I am missing here?



Did some tests, and was able to replicate your problem so hopefully this
will work on your system.

First, make sure the OS can start the RAID-0's. If it can, see if you
can start the RAID10 with:
mdadm -As /dev/md45

If you get something like:
mdadm: no devices found for /dev/md45

Then the problem is your mdadm.conf file. In that case, run:
mdadm --examine --scan /dev/md4 /dev/md5 >> /etc/mdadm.conf

And for some reason, it didn't like the devices= field on my system, so
get rid of those and replace it with a "DEVICE partitions" line at the
very top of the file.

If you can then do the mdadm -As command, you should be good to go.

And definitely talk to your boss about WHY he wants mirrored stripes.
If its for 'performance reasons' offer to bench the two and that should
show him that any performance difference would be statistically
insignificant and not worth the loss of redundancy. Having numbers to
back up your case may help.

Hope this helps,

Daniel

I scrapped the mirrored stripe and switched to striped mirror. What the
heck. So I then created 3 mirrors and joined them together. Everything
went ok.
- 3 RAID1s - /dev/md4 /dev/md5 /dev/md6
- combined them to RAID 0 - /dev/md10
- Did mkfs.ext3 /dev/md10
- Added md10 to /etc/mdadm.conf. Doing 'mdadm --examine --scan
--config=/etc/mdadm.conf' shows all my RAID arrays correctly.

Now I tried stopping /dev/md10 then restarted it with 'mdadm -As
/dev/md10' but then it gives me an error that 'no devices found for
/dev/md10.' If I do 'mdadm --examine --scan --config=/etc/mdadm.conf,'
/dev/md10 shows up. Any more suggestions? I've given myself until today to
make this work, or else I'm gonna go for RAID5.

Thanks to all.
Edward





--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list

Relevant Pages

  • Re: RAID 0+1 (Problem)
    ... I followed the steps in the link below but my RAID 10 array had a problem ... But after a reboot, I got errors re this new RAID device and ... mdadm: ... If its for 'performance reasons' offer to bench the two and that should show him that any performance difference would be statistically insignificant and not worth the loss of redundancy. ...
    (Fedora)
  • Re: RAID 0+1 (Problem)
    ... I followed the steps in the link below but my RAID 10 array had a problem ... But after a reboot, I got errors re this new RAID device ... mdadm: ... If its for 'performance reasons' offer to bench the two and that should ...
    (Fedora)
  • Re: [opensuse] Software RAID with existing drive
    ... If you are talking about software raid 1, ... done it without a reboot. ... I'm running RAID 5 and if I use mdadm to remove the drive, ...
    (SuSE)
  • Re: [opensuse] 10.2 no RAID to 11.0 RAID 1
    ... 10.2 no RAID to 11.0 RAID 1 ... you can force rebuilds in mdadm in situations where no firmware raid will ever let you. ... Send me your problem disks that you think are impossible to assemble and I bet in a little while I can tell you how to assemble the array as long as there actually is enough there to use. ... One thing I've seen, which I don't think is your problem but shows the kind of thing that happens, a disk will drop out of the array and reappear instantly as a higher drive letter than the system really has. ...
    (SuSE)
  • HELP Recover software RAID5...
    ... I am hoping some brilliant sysadm can help me recover my software raid 5. ... Over Time as the media drive was filled up, I add additional drives to the ... mdadm: /dev/sdh1 requires wrong number of drives. ... mdadm: /dev/sdh has wrong uuid. ...
    (comp.os.linux.misc)