On Tue, Mar 13, 2012 at 10:52 AM, Bonno Bloksma <b.blok...@tio.nl> wrote:
> I have been using Linux for several years but now I have another learning > moment. ;-) I have read several LVM howto sites but still run into things I > do not understand. > Can someone please answer the 4 embedded questions and tell me what > I(?) did wrong. > > Using the 6.0.4 amd netinst CD I created a small 10GB virtual machine > (VM). I then realized I needed it to be a bit bigger so I wanted to extend > the LVM environment and add that space to the /var logical volume. > Of course as this is a new VM I just could have started from scratch > but I am trying to learn something as well. ;-) > > First I had a look at the current disk layout using fdisk -l > > root@wwwgw:~# fdisk -l > > Disk /dev/sda: 10.7 GB, 10737418240 bytes > 255 heads, 63 sectors/track, 1305 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: 0x000d6e97 > > Device Boot Start End Blocks Id System > /dev/sda1 1 37 291840 83 Linux > Partition 1 does not end on cylinder boundary. > /dev/sda2 37 1306 10190849 5 Extended > /dev/sda5 37 1306 10190848 8e Linux LVM > > 1) > Why the warning about the 300MB /boot partition not ending on a cylinder > boundary? I used the manual setup in the Debian 6.0.4 installation and > told it to create a 300MB partition at the beginning of the disk. Did > the installation software do something wrong or should I have know > something I do not know yet? I've never liked this message and I therefore partition manually. Someone else on the list might know whether you can incur a performance hit on a "regular" disk; you certainly do with an SSD or an "advanced format" disk. The wheezy installer corrects this and starts at 2048 (units=sectors). > 2) > After that came a lot of warnings about /dev/dm-0, /dev/dm-1, /dev/dm-2, > /dev/dm-3 and /dev/dm-4. I know those are LVM2 devices but... > Why is fdisk (still) seeing them as disks/partitions it has to show during > a listing, and then complain they are not valid? I assume that "fdisk -l" goes through "/proc/partitions" and displays the information that it can about its contents. It doesn't recognize the device mapper devices because they are virtual block devices created by dmsetup/lvm within a "real"/physical block device. > At the VMware level I increased the disk from 10GB to 12GB. Using > cfdisk, which in my opinion gives less cause for a user error, I created > a new logical sda6 partition in the free space. The end result is: > root@wwwgw:~# fdisk -l > > Disk /dev/sda: 12.9 GB, 12884901888 bytes > 255 heads, 63 sectors/track, 1566 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: 0x000d6e97 > > Device Boot Start End Blocks Id System > /dev/sda1 1 37 291840 83 Linux > Partition 1 does not end on cylinder boundary. > /dev/sda2 37 1566 12285008 5 Extended > /dev/sda5 37 1306 10190848 8e Linux LVM > /dev/sda6 1306 1566 2094127+ 8e Linux I've always assumed that it meant that the partition's size is between 2094127 and 2094128 blocks but assumption can be the mother of all <deleted>! > 3) > Why is there a + at the end of the number of blocks? > > I then want to make the sda6 partition a LVM physical volume using > root@wwwgw:~# pvcreate /dev/sda6 > Device /dev/sda6 not found (or ignored by filtering). > root@wwwgw:~# > indeed, there is no /dev/sda6 yet. > > 4) > Why is /dev/sda6 not there yet? What step am I missing? I've never used cfdisk; don't you get a message that the kernel's still using the old partition table? > Ok, after a reboot (it is not a production server yet) the /dev/sda6 is there. > Form here on it was (almost) straight sailing. ;-) No need to reboot; partprobe or kpartx can load the new partition table for you. > root@wwwgw:~# pvcreate /dev/sda6 > Physical volume "/dev/sda6" successfully created > root@wwwgw:~# vgextend vgroup1 /dev/sda6 > Volume group "vgroup1" successfully extended > > Then to runlevel 1 to make sure (almost) nothing is using the /var directory > tree > and > root@wwwgw:~# lvextend -l+2G /dev/vgroup1/lvvar > Extending logical volume lvvar to 3.86GiB > Logical volume lvvar successfully resized > > root@wwwgw:~# umount /var > Then first a filesystem check as that seems to be needed before resizing. Not > doing so will give me a warning although this was not mentioned in the HOWTOs > I have read. > root@wwwgw:~# fsck -f /dev/vgroup1/lvvar > root@wwwgw:~# resize2fs /dev/vgroup1/lvvar > root@wwwgw:~# mount /var > > and back to runlevel 2 I'm surprised that you were able to unmount "/var", even in runlevel 1, but I assume that it depends on your setup. Running fsck is a good idea. You can also do so after "resize2fs ...". I'd have booted via the network, a rescue CD, or a live CD to do this. You could probably even do this via the initramfs (untested!). -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOdo=sw34o7hmypcybkzhpntnfewomjywq96saaurtwfxbn...@mail.gmail.com
