trouble adding a disk to exisiting lvm
Hi,
Probably this is a known problem but anyway as I'm new to the subject of LVM I give it a go.
We set up on a server machine an initial LVM consisting of two 80 Gb drives dangling on a Promise RAID controller (just using them as /dev/hde and /dev/hdg, aka two separate IDE devices, no raid magic involved)
This worked as a charm and the resulting LVM was setup with ext3 filesystem without a hitch. The LVM was filled with data from other IDE devices (source) and when these were copied (they were FAT32 volumes) these source drives were emptied and with fdisk the partition table was emptied too.
Pvcreate was done on the first drive to be added and after vgscan and vgextend an fsck -f was done on the lvm volume before resize2fs was done. All this still went without problems, remounting the extended volume gave no errors.
The volume was unmounted and the next drive was to be added, same procedure (partition table clear with fdisk, pvcreate, vgscan, vgextend) but then as fsck was run a huge pile of numbers came up with at the end the question fix<y>? And I aborted this by Ctrl-C to keep the filesystem as it were.
I guess the adding of the third drive issued this problem, did resize2fs
foul up ? Is the whole volume at risk ?
Regards,
Gert
It would be very helpful if you were to include useful information such as the output that fsck spewed at you, some general data about your LVs and this troublesome drive, or even a log of the entire event that you tried if you have it.
Otherwise, you can scan your palm and we can try to read that for you. =)
I like to run something like badblocks, Bonnie++, and smartmontools, before I trust a new drive. Write a little script, let it run, and see if it fouls up over a 24-48 hour period. Then you can blame it on software later if something goes wrong.
Quoted 28 lines · Gert van der Knokke
Hi,
Probably this is a known problem but anyway as I'm new to the subject of LVM I give it a go.
We set up on a server machine an initial LVM consisting of two 80 Gb drives dangling on a Promise RAID controller (just using them as
/dev/hdeand/dev/hdg, aka two separate IDE devices, no raid magic involved)
This worked as a charm and the resulting LVM was setup with ext3 filesystem without a hitch. The LVM was filled with data from other IDE devices (source) and when these were copied (they were FAT32 volumes) these source drives were emptied and with fdisk the partition table was emptied too.Pvcreate was done on the first drive to be added and after vgscan and vgextend an fsck -f was done on the lvm volume before resize2fs was done. All this still went without problems, remounting the extended volume gave no errors.
The volume was unmounted and the next drive was to be added, same procedure (partition table clear with fdisk, pvcreate, vgscan, vgextend) but then as fsck was run a huge pile of numbers came up with at the end the question fix<y>? And I aborted this by Ctrl-C to keep the filesystem as it were.
I guess the adding of the third drive issued this problem, did resize2fs
foul up ? Is the whole volume at risk ?Regards,
Gert_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
Signature · 5 lines
# Jesse Molina # Mail = [email protected] # Page = [email protected] # Cell = 1.407.970.0280 # Web = http://www.opendreams.net/jesse/
Quoted 2 lines · Jesse Molina
It would be very helpful if you were to include useful information such as the output that fsck spewed at you, some general data about your LVs and this troublesome drive, or even a log of the entire event that you tried if you have it.
Ok, I will try to capture this output. It's more than Shift-PgUp can handle anyway so I could not see what was at the start of the output. And since the server is a AMD 1800 (Ok it's overkill...) the messages swoosh faster than an eye can blink :-) The machine is not here with me now.
Quoted 2 lines
Otherwise, you can scan your palm and we can try to read that for you. =)
Ah, this is the filosofical approach I guess :-)
Quoted 2 lines
I like to run something like badblocks, Bonnie++, and smartmontools, before I trust a new drive. Write a little script, let it run, and see if it fouls up over a 24-48 hour period. Then you can blame it on software later if something goes wrong.
Hmmm.. As I wrote, the original setup of the lvm consisiting of two 80 Gb went without problems, and adding the third drive later 'seemed' to work until the fourth drive was added an the fsck was run on the lvm (then consisting of the three drives). Note I have not put any data on the lvm since I added the third drive (so there is no more than 160 Gb used)
Is it principally safe to extend an ext3 filesystem with resize2fs ? I'd better had put reiserfs on it I guess instead of ext3 but now I'm stuck with 160 Gb of data on the lvm which I can't park easily somewhere else so I'm reluctant to open the bag of tricks.
Maybe I should reduce the volume back to 160 Gb (2x80) and move the data back on the source drives and start all over ? This time using reiserfs ?
Any help will do.
Gert