What is the best practise for using Proxmox 4.1 and 2x SATA HDD in mirror (without raid hardware)?
Best practise for softraid 2x SATA HDD and Proxmox 4.1?
- Thread starter math jansen
- Start date
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
This is easy to install because it is just point and click in the ISO install, but I have a little reservation about it. There is a known bug with the partitioner placing SWAP on a so-called ZVOL (a virtual block device on zfs). As soon as swap gets used a little, your machine will be prone to sudden reboots.
https://forum.proxmox.com/threads/zfs-swap-crashes-system.25208/#post-126215
Also there have been some incidents with quiet data corruption, seemingly connected to hardware configurations. Therefore I would recommend doing some basic "burn-in" to establish the correct behaviour of the system.
You can make a large file with random data of your own choosing and afterwards do a zpool scrub to check if the raid array behaves correctly.
Zpool scrubbing is the practice of checking all stored blocks in the background against their checksums. The default zpool is called rpool with Proxmox.
http://docs.oracle.com/cd/E23823_01/html/819-5461/gbbwa.html
You can quickly create a file with random data with this command (replace nn with the number of 32M blocks you want to burn, 32=1G, replace path/to with your own location):
I would recommend doing a burn-in with at least 200GB. This will also give you a good benchmark of the disk throughput.
https://forum.proxmox.com/threads/zfs-swap-crashes-system.25208/#post-126215
Also there have been some incidents with quiet data corruption, seemingly connected to hardware configurations. Therefore I would recommend doing some basic "burn-in" to establish the correct behaviour of the system.
You can make a large file with random data of your own choosing and afterwards do a zpool scrub to check if the raid array behaves correctly.
Zpool scrubbing is the practice of checking all stored blocks in the background against their checksums. The default zpool is called rpool with Proxmox.
http://docs.oracle.com/cd/E23823_01/html/819-5461/gbbwa.html
You can quickly create a file with random data with this command (replace nn with the number of 32M blocks you want to burn, 32=1G, replace path/to with your own location):
Code:
openssl enc -aes-256-ctr -pass:dummy -nosalt < /dev/zero | dd of=path/to/randomfile bs=32M count=nn iflag=fullblock oflag=syncI would recommend doing a burn-in with at least 200GB. This will also give you a good benchmark of the disk throughput.