meta data for this page
  •  

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
vm:proxmox:kvm:linux [2025/01/06 21:27] – niziakvm:proxmox:kvm:linux [2025/03/04 15:20] (current) – niziak
Line 1: Line 1:
 ====== Linux Guests ====== ====== Linux Guests ======
 +
 +===== Filesystem =====
 +
 +If ZFS is used on host (CoW filesystem) using CoW filesystem on guest generates overhead. Better is to use simple XFS or Ext4 if BTRFS snapshot and subvolumes functionality is not needed in guest.
  
 Linux Guest on BTRFS: Linux Guest on BTRFS:
 +
   * BTRFS with single copy of meta, system and data   * BTRFS with single copy of meta, system and data
   * I/O is better on modern guest Kernel 6.1 (vs 5.10)   * I/O is better on modern guest Kernel 6.1 (vs 5.10)
   * I/O is better whem VM storage is "no cache"   * I/O is better whem VM storage is "no cache"
-    * sync writes goes to ZFS ZIL +      * sync writes goes to ZFS ZIL 
-    * With "Writeback" cache - ZFS ZIL was not utilized so high like with "no cache" +      * With "Writeback" cache - ZFS ZIL was not utilized so high like with "no cache"
  
 Block sizes: Block sizes:
  
-  * Default ''zvol'' ''volblocksize'' was 8k +  * Default ''zvol''''volblocksize''  was 8k 
-  * Latest Proxmox creates ''zvols'' with 16k ''volblocksize''+  * Latest Proxmox creates ''zvols''  with 16k ''volblocksize''
  
 So it is best to fit guest block sizes into ZFS block (16k) sizes: So it is best to fit guest block sizes into ZFS block (16k) sizes:
-  * mkfs.btfs options: --nodesize 16k+ 
 +  * mkfs.btfs options: –nodesize 16k
   * btrfs mount options: -o noatime   * btrfs mount options: -o noatime
   * mkfs.ext4 options: -b 4096 -E stride=4,stripe-width=8   * mkfs.ext4 options: -b 4096 -E stride=4,stripe-width=8
   * ext4 mount options: -o noatime,user_xattr   * ext4 mount options: -o noatime,user_xattr
 +