Ok, I have a nice 700GB LVM sitting on top of a couple of 350GB software
RAID5 volumes (on 4 250GB SATA disks). I want to add another 330GB of
software RAID 5 (on 4 120GB PATA disks), so here's what I did:
mdadm --create /dev/md2 --chunk 256 --level 5 -n 4 /dev/hdc1 /dev/hdd1
/dev/hdg1 /dev/hdh1
This went fine, and /dev/md2 is working as well as can be expected
considering the master/slave drives used to build it:
md2 : active raid5 hdh1[3] hdg1[2] hdd1[1] hdc1[0]
351661824 blocks level 5, 256k chunk, algorithm 2 [4/4] [UUUU]
[root@backup root]# lvm pvs
PV VG Fmt Attr PSize PFree
/dev/md0 backups lvm2 a- 349.33G 0
/dev/md1 backups lvm2 a- 349.31G 8.64G
Here is where it gets weird. So I want to add /dev/md2:
[root@backup root]# lvm pvcreate /dev/md2
No physical volume label read from /dev/md2
Physical volume "/dev/md2" successfully created
[root@backup root]# lvm pvs
PV VG Fmt Attr PSize PFree
/dev/hdc1 lvm2 -- 335.37G 335.37G
/dev/md0 backups lvm2 a- 349.33G 0
/dev/md1 backups lvm2 a- 349.31G 8.64G
Wha?! Thats not right! It sees disk 0 of the RAID5 set as having the pv.
Why is this? I created and added md0 and md1 with the same methods
almost exactly, albeit on a 3Ware card, so they show up as sda - sdd. I
can see why this might happen, but it still is a bit disconcerting.
A little more weirdness:
[root@backup root]# lvm pvremove /dev/hdc1
Labels on physical volume "/dev/hdc1" successfully wiped
[root@backup root]# lvm pvs
PV VG Fmt Attr PSize PFree
/dev/md0 backups lvm2 a- 349.33G 0
/dev/md1 backups lvm2 a- 349.31G 8.64G
/dev/md2 lvm2 -- 335.37G 335.37G
This looks more like what I want.. .but still.. I'm afraid to use this
PV. I can use vgextend and add it into my backups vg, and all seems
fine:
[root@backup root]# lvm vgextend -v backups /dev/md2
Checking for volume group "backups"
Archiving volume group "backups" metadata.
Adding physical volume '/dev/md2' to volume group 'backups'
Volume group "backups" will be extended by 1 new physical volumes
Creating volume group backup "/etc/lvm/backup/backups"
Volume group "backups" successfully extended
[root@backup root]# lvm vgs
VG #PV #LV #SN Attr VSize VFree
backups 3 1 0 wz-- 1.01T 344.01G
But, whats this? Weirdness returns!:
[root@backup root]# lvm pvs
PV VG Fmt Attr PSize PFree
/dev/hdc1 backups lvm2 a- 335.37G 335.37G
/dev/md0 backups lvm2 a- 349.33G 0
/dev/md1 backups lvm2 a- 349.31G 8.64G
If I remove hdc1 from /etc/lvm/.cache, I get this:
[root@backup root]# lvm pvs
PV VG Fmt Attr PSize PFree
/dev/md0 backups lvm2 a- 349.33G 0
/dev/md1 backups lvm2 a- 349.31G 8.64G
/dev/md2 backups lvm2 a- 335.37G 335.37G
So I must assume that this might just be a display problem. However,
this definitely looks like a bug. Either way I'm afraid to rely on
/dev/md2 until I figure out why this is happening. Anybody have any
ideas?
More info:
[root@backup root]# lvm version
LVM version: 2.00.08 (2003-11-14)
Library version: 1.00.07-ioctl (2003-11-21)
Driver version: 4.1.0
OS: Fedora Core 1
[root@backup root]# uname -a
Linux backup 2.6.5 #1 SMP Tue Apr 6 14:42:41 PDT 2004 i686 athlon i386 GNU/Linux
Hardware: Dual Athlon MP 1900+, Tyan S2466-N4M Motherboard, 512MB DDR
REG ECC RAM. 3ware Escalade 8506-8 SATA card. Highpoint HPT302 ATA/133
Controller. 4x WDC 250GB SATA Drives. 4xSeagate 120GB PATA drives.
Thanks!
--
Clint Byrum
Systems Administrator
CareerCast, Inc.
_______________________________________________
linux-lvm mailing list
linux-lvm@redhat.com
https://www.redhat.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
