Very slow cloning VM (50 minutes for 32 Go)

[cdarsac]

Hi,

I have created an Ubuntu VM 32 Go with Proxmox.

The duration of cloning is about 50mn in my file-system (RAID0 with 7 disk SAS 2,5'' 10k rpm)


The transfert rapidely goes to 100% during the first minute, then nothing seems to happend until the cloning end.


During cloning, the graphs stop displaying the curves, although the digital counters (CPU Usage, Load average, IO Delay, RAM...) continue to be updated



If I clone the same VM from a HDD-Usb disk to the file-system (not same location) , the duration is the same (about 48 mn).

But if I reversely clone from the file-system to the HDD-Usb disk, the duration is only 43 seconds !!!


Then I tested transfert speed with hdd command ( (dd if=/dev/zero of=/tmp/test.data bs=10M count=1024 conv=fdatasync):

to FS	338 Mo/s​	​	​	​
to HDD-USB	36 Mo/s​	​	​	​
	​	​	​	​

So the file-system seems to work properly...


Specifications:
- RAID0 with 7 disques SAS 2.5'' 10krpm, 660 Go (NETAPP X422A-R5)
- controleur RAID: P420i, with no "RAM module"
- Server HP Proliant DL360P G8 ( https://www.serveur-occasion.com/fr...-x-xeon-eight-core-e5-2650-64-go-rack-1u.html )

CPU(s): 32 x Intel(R) Xeon(R) CPU E5-2650 0 @ 2.00GHz (2 Sockets)

Kernel Version: Linux 5.4.106-1-pve #1 SMP PVE 5.4.106-1 (Fri, 19 Mar 2021 11:08:47 +0100)

PVE Manager Version: pve-manager/6.4-4/337d6701

Do you know where the very slow cloning problem may come from ?

Best regards,
Christophe.

 

[bedo]

I am having almost the same problem as well. Did you find a solution?

 

[cleanden]

Same problem here with PVE 8.0.3
My storage is 8 x 300GB 15K 6Gb/s SFF SAS drives in RAID5 configuration on an HPE ProLiant DL360 Gen9, connected via a Smart HBA H240a controller.

 

[_gabriel]

h240a missing onboard cache acceleration + i think any raid mode disable disks writecache (as smart array do), so write performance is bad.
You can confirm with proper fio benchmark like described in documentation + a 5 minutes test duration.