Proxmox 7 Very Slow restoring

[Jose A]

Hi guys!
I have a server with an SSD disk, only 1 disk for probe performance, i know, but the problem is, when i try to restore a VM, i just wait 13 hours for 95%!! the normal time is about 20 minutes, i think the problem is the SSD, it is possible?

I'm testing a Seagate Firecuda 120 SSD disk (1 TB), a year ago i had bad experiences using a desktop ssd in proxmox, i choose this Firecuda because have more TBW than other desktop disk.

 

[spirit]

It's a consumer ssd, It's quite possible that it's pretty slow with small block write.

what is the result of "#pveperf" ?

 

[Jose A]

# pveperf
BOGOMIPS CPU: 24740.60
REGEX / SEGUNDO: 3330163
TAMAÑO HD: 93,93 GB (/ dev / mapper / pve-root)
LECTURAS EN BÚFER: 252,35 MB / seg
TIEMPO DE BÚSQUEDA PROMEDIO: 0,12 ms
FSYNCS / SEGUNDO: 170.61
EXT DNS: 242,88 ms
INT DNS: 249,27 ms (lan)

 

[Jose A]

Hello Again, I make a new ProxMox instalation, only 1 disk in ext4 (Seagate Firecuda 120 SSD, 1 TB), i put the backup in the /var/lib/vz/dump and just make a restoring, result:


restore vma archive: zstd -q -d -c /var/lib/vz/dump/vzdump-qemu-100-2021_06_07-12_23_29.vma.zst | vma extract -v -r /var/tmp/vzdumptmp2385.fifo - /var/tmp/vzdumptmp2385
CFG: size: 427 name: qemu-server.conf
DEV: dev_id=1 size: 214748364800 devname: drive-scsi0
CTIME: Mon Jun 7 12:23:31 2021
Logical volume "vm-100-disk-0" created.
new volume ID is 'local-lvm:vm-100-disk-0'
map 'drive-scsi0' to '/dev/pve/vm-100-disk-0' (write zeros = 0)
progress 1% (read 2147483648 bytes, duration 9 sec)
progress 2% (read 4294967296 bytes, duration 23 sec)
progress 3% (read 6442450944 bytes, duration 299 sec)
progress 4% (read 8589934592 bytes, duration 1364 sec)

look at the time!!! in each percentage it grows a lot!!! i wait for the 5% more than mid hour!
This backup was 40 Gb, i think in this conditions it takes more than 15 hours te restoring process.

 

[Jose A]

In the same server, i installed promox 6, same result:

restore vma archive: zstd -q -d -c /var/lib/vz/dump/vzdump-qemu-100-2021_06_07-12_23_29.vma.zst | vma extract -v -r /var/tmp/vzdumptmp3955.fifo - /var/tmp/vzdumptmp3955
CFG: size: 427 name: qemu-server.conf
DEV: dev_id=1 size: 214748364800 devname: drive-scsi0
CTIME: Mon Jun 7 12:23:31 2021
Logical volume "vm-100-disk-0" created.
new volume ID is 'local-lvm:vm-100-disk-0'
map 'drive-scsi0' to '/dev/pve/vm-100-disk-0' (write zeros = 0)
progress 1% (read 2147483648 bytes, duration 19 sec)
progress 2% (read 4294967296 bytes, duration 39 sec)
progress 3% (read 6442450944 bytes, duration 272 sec)
progress 4% (read 8589934592 bytes, duration 1351 sec)

I think is the SSD disk, i was install another server with SSD with no problems like this. I don't understand !

 

[Markoto]

Hello José,

I have the same problem as you. We noticed that everything that is installed with Kernel version 5 or UP have this issue. If you use a Kernel version of 4 you should not have a problem. Does someone have a work around this ?