[CentOS] mkswap, lvm and bootbits sectors

I'd say it's safe but why do it?

What likely happened is that the lv you ran mkswap on contained a valid
partition table (or at least boot signature). Out of paranoia mkswap left the
first block alone. So what you could gain is maybe another 512 byte in
swapspace by adding -f.

You could write a meg of zeroes to the device first and see if the mkswap
warning goes away.

/Peter

<snip>
Are you making swap from a logical partition, or a swapfile? If the
former, perhaps recreating the partition might help.

mark

...
Don't confuse the poor guy. 1) yes he's doing mkswap on an lv which was
obvious had you read the post 2) an lv does not have a partition table so your
statement about recreating it makes no sense.

/Peter

Thanks all for your answers.

Actually, my problem concern lvm, not mkswap. I don't know yet if I
have completely messed with vgcreate/lvcreate or if what I get is due
to some difference between CentOS-5 and CentOS-6, but all my logical
volumes seem to appear as distinct disks.

On a CentOS-6 machine :

$ lvcreate -L 10M --name try_lvcreate --zero=y VolGroup
Rounding up size to full physical extent 12.00 MiB
Logical volume "try_lvcreate" created

$ fdisk -l
<snip the "normal" partitions table for /dev/sda>

Disk /dev/mapper/VolGroup-try_lvcreate: 12 MB, 12582912 bytes
255 heads, 63 sectors/track, 1 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

Disk /dev/mapper/VolGroup-try_lvcreate doesn't contain a valid partition table

$ mkswap /dev/mapper/VolGroup-try_lvcreate
mkswap: /dev/mapper/VolGroup-try_lvcreate: warning: don't erase bootbits sectors
on whole disk. Use -f to force.
Setting up swapspace version 1, size = 12284 KiB
no label, UUID=be229ca5-bcb9-4116-87fc-8878abb44742

The same commands on a CentOS-5 machine give me a completely
different output. Can you guess where I have messed, or is this
behavior correct on CentOS-6 ?

Thanks,

No, afaict there's nothing wrong with your lvm. The only problem I see is that
mkswap prints our an incorrect or at least pointless warning.

I can confirm that mkswap on an lv on one of my test machines also says
"warning: don't erase bootbits...". It also says this when running against a
file full of zeroes...

Regarding fdisks complaints about "doesn't contain a valid partition
table...", this is normal (and may be considered a bug in the "-l" option, it
could have excluded lvm-devices from its list...).

If you want to get to the bottom of this I suspect the easiest way is to look
at the mkswap source code (under which circumstances it prints that message).

/Peter

Ahh, perfect, thanks a lot. These bugs^W features of fdisk and mkswap
were absent of CentOS-5 : I was lost, wondering what can be my
mistakes.

Thanks again,