Expert Help needed (raid1+lvm) Did I destroy 2TB of backups?
Dean S. Messing
deanm at sharplabs.com
Thu Nov 11 23:09:36 UTC 2010
My System: up-to-date Fedora 13
The Situation:
I have two 2TB drives (/dev/sde1, /dev/sdf1) bound into a RAID-1 on /dev/md1.
This device (md1) is an lvm PV on which there is a single VG
(vg_medulla_bkup) In the VG is a single LV (vg_bkup). All has been
working well since I built it two weeks ago.
Today I did something VERY stupid: I accidently stopped the array by
typing md1 instead md2. I was able to start it again by assembling it,
and I can query the RAID just fine, but somehow the LV device has become
"disconnected" from md1.
Wheneven I query the LV I get lines like.
/dev/vg_medulla_bkup/lv_bkup: read failed after 0 of 4096 at 2000393535488: Input/output error
/dev/vg_medulla_bkup/lv_bkup: read failed after 0 of 4096 at 0: Input/output error
/dev/vg_medulla_bkup/lv_bkup: read failed after 0 of 4096 at 4096: Input/output error
I suppose what I did was the equivalent of unplugging the phsyical
device out from underneath an ordinary (non-RAID) Volume Group sitting
on it.
I could try to reboot and hope for the best. But this RAID contains a
year's worth of backups and I really don't want to destroy it.
So I thought I'd try the Font of Knowledge first. Are there lvm or
other commands that will "reconnect" lvm? Maybe `vgmknodes' ?
Thanks for your help.
More information about the users
mailing list