PVE

Da Helpedia.

VZDump

Configurazione di vzdump si effettua su /etc/vzdump.conf, con i seguenti parametri

tmpdir: DIR
dumpdir: DIR
storage: STORAGE_ID
mode: snapshot|suspend|stop
bwlimit: KBPS
lockwait: MINUTES
stopwait: MINUTES
size: MB
maxfiles: N
script: FILENAME

E' possibile indicare solo i parametri che interessano

esempio:

tmpdir: /var/lib/vz/tmp
maxfiles: 3

RAID

The Problem

Add Linux software RAID1 to a ProxMoxVE install. This assumes you got 2 similar HDDs already setup in a server and that you know how to install ProxMoxVE (just plop the CD in the tray and blindly follow the prompts).

The Solution

Install Proxmox as usual onto 1 drive(/dev/sda)

Get extra packages

mdadm will ask which MD arrays are needed for the root fs. Go with the default (all)

apt-get install mdadm initramfs-tools vim

Add raid1 to /etc/modules

echo "raid1" >> /etc/modules

Regenerate initrd.img file

mkinitramfs -o /boot/test -r /dev/mapper/pve-root

Rename old img file (replace .x for whatever kernel version you are using)

mv /boot/initrd.img-2.6.x-pve /boot/initrd.img-2.6.x-pve.original

Rename new img file

mv /boot/test /boot/initrd.img-2.6.x-pve

Make sure grub is setup on both hdds

grub-install --no-floppy /dev/sda
grub-install --no-floppy /dev/sdb

Change UUID for (hd0,0) on /boot/grub/menu.1st if your file has the UUID in it.

This most likely involves replacing the root UUID=XXXXX line with root hd(0,0)

vim /boot/grub/menu.1st

Change UUIDs for proper md0 devices on /etc/fstab if your file has the UUID in it.

Now do the same thing for the /etc/fstab file.

Replace the UUID=XXXXXXXXX /boot ext3 defaults 0 1 line with /dev/md0 /boot ext3 defaults 0 1 .

vim /etc/fstab

Clone the partition table from 1st drive to 2nd

sfdisk -d /dev/sda | sfdisk --force /dev/sdb

Create md devices with second drive only

mdadm --create /dev/md0 --level=1 --raid-devices=2 missing /dev/sdb1
mdadm --create /dev/md1 --level=1 --raid-devices=2 missing /dev/sdb2

Save new mdconf file

mdadm --detail --scan >> /etc/mdadm/mdadm.conf

Get boot device setup

mkfs.ext3 /dev/md0
mkdir /mnt/md0
mount /dev/md0 /mnt/md0
cp -ax /boot/* /mnt/md0
umount /mnt/md0
umount /boot; mount /dev/md0 /boot 
sfdisk --change-id /dev/sda 1 fd
mdadm --add /dev/md0 /dev/sda1

Setup data device

pvcreate /dev/md1
vgextend pve /dev/md1

this step takes a LOOOOONG time

pvmove /dev/sda2 /dev/md1
vgreduce pve /dev/sda2
sfdisk --change-id /dev/sda 2 fd
mdadm --add /dev/md1 /dev/sda2

That is it!

You can monitor the array rebuild progress with the following command:

watch -n 1 cat /proc/mdstat

If for some reason mdstat reports that the raid will take days to complete, you can adjust the minimal raid rebuild rate with the following command:

echo 60000 >/proc/sys/dev/raid/speed_limit_min

The default setting of 1000 may be too slow on most situations. Again, upping this too much on a system that is in use is not advisable. Also make sure you don't put the value higher than /proc/sys/dev/raid/speed_limit_max.

After this process is completed you can reboot the system to make sure everything starts up nicely.

Easy Peasy!

-PCP

from [1]

Strumenti personali
Namespace
Varianti
Azioni
Navigazione
Strumenti