Slow cloning and backup/restore

JesperAP

New Member
Jun 18, 2024
24
0
1
Hi All,

We have a proxmox cluster of 9 Dell R640's with 2x Xeon Gold 6154 and 1.5TB RAM.

We use NVMe/TCP with a NetApp AFF-A250 as our storage for the VM's.

The cloning of VM's and backup/restoring is very slow and has always been slow for us.

Is anyone else experiencing the same? Is this normal for proxmox? We came from ESXi and it was much faster there.
 
I am doing a disk move from NAS storage (synology) to the NetApp (NVMe/TCP) because i chose the wrong storage when installing.

The moving is at 100% for about 15 minutes now, anyone else also experiencing this?
 
could you describe your storage setup in more detail (e.g., storage.cfg and sample VM configs?) and post a task log showing the problematic behaviour?

thanks!
 
could you describe your storage setup in more detail (e.g., storage.cfg and sample VM configs?) and post a task log showing the problematic behaviour?

thanks!
Hi Fabian,

Sorry for the extreme late reply. We are still experiencing this problem.

This is the storage.cfg

Code:
dir: local
        disable
        path /var/lib/vz
        content iso
        prune-backups keep-all=1
        shared 0

lvm: NVMe
        vgname nvme_vg
        content images,rootdir
        saferemove 0
        shared 1

And this is one of our VM configs:

Code:
agent: 1
bios: ovmf
boot: order=scsi0;sata0
cores: 4
cpu: host
efidisk0: NVMe:vm-105-disk-3,size=128K
machine: pc-i440fx-8.1
memory: 16384
meta: creation-qemu=8.1.5,ctime=1715087706
name: *******************
net0: vmxnet3=00:50:56:8c:e8:f7,bridge=vmbr0,tag=16
numa: 1
ostype: win10
sata0: none,media=cdrom
scsi0: NVMe:vm-105-disk-0,discard=on,iothread=1,size=60G,ssd=1
scsi1: NVMe:vm-105-disk-1,discard=on,iothread=1,size=100G,ssd=1
scsi2: NVMe:vm-105-disk-2,discard=on,iothread=1,size=100G,ssd=1
scsihw: virtio-scsi-single
smbios1: uuid=564d778a-1fee-a555-2adf-f2e859e7f5ba
sockets: 1
tags: dvm
vmgenid: 348a4bef-922d-4dae-9ab1-3585fc32ccc7


And this is a clone log:

Code:
create full clone of drive efidisk0 (NVMe:vm-234-disk-0)
  Rounding up size to full physical extent 4.00 MiB
  Logical volume "vm-242-disk-0" created.
create full clone of drive scsi0 (NVMe:vm-234-disk-1)
  Wiping PMBR signature on /dev/nvme_vg/vm-242-disk-1.
  Logical volume "vm-242-disk-1" created.
transferred 0.0 B of 250.0 GiB (0.00%)
transferred 2.5 GiB of 250.0 GiB (1.00%)
transferred 5.0 GiB of 250.0 GiB (2.00%)
transferred 7.5 GiB of 250.0 GiB (3.00%)
transferred 10.0 GiB of 250.0 GiB (4.00%)
transferred 12.5 GiB of 250.0 GiB (5.00%)
transferred 15.0 GiB of 250.0 GiB (6.00%)
transferred 17.5 GiB of 250.0 GiB (7.01%)
transferred 20.0 GiB of 250.0 GiB (8.01%)
transferred 22.5 GiB of 250.0 GiB (9.01%)
transferred 25.0 GiB of 250.0 GiB (10.01%)
transferred 27.5 GiB of 250.0 GiB (11.01%)
transferred 30.0 GiB of 250.0 GiB (12.01%)
transferred 32.5 GiB of 250.0 GiB (13.01%)
transferred 35.0 GiB of 250.0 GiB (14.01%)
transferred 37.5 GiB of 250.0 GiB (15.01%)
transferred 40.0 GiB of 250.0 GiB (16.01%)
transferred 42.5 GiB of 250.0 GiB (17.01%)
transferred 45.0 GiB of 250.0 GiB (18.01%)
transferred 47.5 GiB of 250.0 GiB (19.01%)
transferred 50.0 GiB of 250.0 GiB (20.02%)
transferred 52.5 GiB of 250.0 GiB (21.02%)
transferred 55.0 GiB of 250.0 GiB (22.02%)
transferred 57.5 GiB of 250.0 GiB (23.02%)
transferred 60.0 GiB of 250.0 GiB (24.02%)
transferred 62.5 GiB of 250.0 GiB (25.02%)
transferred 65.0 GiB of 250.0 GiB (26.02%)
transferred 67.5 GiB of 250.0 GiB (27.02%)
transferred 70.0 GiB of 250.0 GiB (28.02%)
transferred 72.5 GiB of 250.0 GiB (29.02%)
transferred 75.0 GiB of 250.0 GiB (30.02%)
transferred 77.5 GiB of 250.0 GiB (31.02%)
transferred 80.0 GiB of 250.0 GiB (32.02%)
transferred 82.6 GiB of 250.0 GiB (33.03%)
transferred 85.1 GiB of 250.0 GiB (34.03%)
transferred 87.6 GiB of 250.0 GiB (35.03%)
transferred 90.1 GiB of 250.0 GiB (36.03%)
transferred 92.6 GiB of 250.0 GiB (37.03%)
transferred 95.1 GiB of 250.0 GiB (38.03%)
transferred 97.6 GiB of 250.0 GiB (39.03%)
transferred 100.1 GiB of 250.0 GiB (40.03%)
transferred 102.6 GiB of 250.0 GiB (41.03%)
transferred 105.1 GiB of 250.0 GiB (42.03%)
transferred 107.6 GiB of 250.0 GiB (43.03%)
transferred 110.1 GiB of 250.0 GiB (44.03%)
transferred 112.6 GiB of 250.0 GiB (45.04%)
transferred 115.1 GiB of 250.0 GiB (46.04%)
transferred 117.6 GiB of 250.0 GiB (47.04%)
transferred 120.1 GiB of 250.0 GiB (48.04%)
transferred 122.6 GiB of 250.0 GiB (49.04%)
transferred 125.1 GiB of 250.0 GiB (50.04%)
transferred 127.6 GiB of 250.0 GiB (51.04%)
transferred 130.1 GiB of 250.0 GiB (52.04%)
transferred 132.6 GiB of 250.0 GiB (53.04%)
transferred 135.1 GiB of 250.0 GiB (54.04%)
transferred 137.6 GiB of 250.0 GiB (55.04%)
transferred 140.1 GiB of 250.0 GiB (56.04%)
transferred 142.6 GiB of 250.0 GiB (57.04%)
transferred 145.1 GiB of 250.0 GiB (58.05%)
transferred 147.6 GiB of 250.0 GiB (59.05%)
transferred 150.1 GiB of 250.0 GiB (60.05%)
transferred 152.6 GiB of 250.0 GiB (61.05%)
transferred 155.1 GiB of 250.0 GiB (62.05%)
transferred 157.6 GiB of 250.0 GiB (63.05%)
transferred 160.1 GiB of 250.0 GiB (64.05%)
transferred 162.6 GiB of 250.0 GiB (65.05%)
transferred 165.1 GiB of 250.0 GiB (66.05%)
transferred 167.6 GiB of 250.0 GiB (67.05%)
transferred 170.1 GiB of 250.0 GiB (68.05%)
transferred 172.6 GiB of 250.0 GiB (69.05%)
transferred 175.1 GiB of 250.0 GiB (70.05%)
transferred 177.6 GiB of 250.0 GiB (71.06%)
transferred 180.1 GiB of 250.0 GiB (72.06%)
transferred 182.6 GiB of 250.0 GiB (73.06%)
transferred 185.1 GiB of 250.0 GiB (74.06%)
transferred 187.6 GiB of 250.0 GiB (75.06%)
transferred 190.1 GiB of 250.0 GiB (76.06%)
transferred 192.6 GiB of 250.0 GiB (77.06%)
transferred 195.1 GiB of 250.0 GiB (78.06%)
transferred 197.6 GiB of 250.0 GiB (79.06%)
transferred 200.1 GiB of 250.0 GiB (80.06%)
transferred 202.6 GiB of 250.0 GiB (81.06%)
transferred 205.1 GiB of 250.0 GiB (82.06%)
transferred 207.6 GiB of 250.0 GiB (83.06%)
transferred 210.2 GiB of 250.0 GiB (84.07%)
transferred 212.7 GiB of 250.0 GiB (85.07%)
transferred 215.2 GiB of 250.0 GiB (86.07%)
transferred 217.7 GiB of 250.0 GiB (87.07%)
transferred 220.2 GiB of 250.0 GiB (88.07%)
transferred 222.7 GiB of 250.0 GiB (89.07%)
transferred 225.2 GiB of 250.0 GiB (90.07%)
transferred 227.7 GiB of 250.0 GiB (91.07%)
transferred 230.2 GiB of 250.0 GiB (92.07%)
transferred 232.7 GiB of 250.0 GiB (93.07%)
transferred 235.2 GiB of 250.0 GiB (94.07%)
transferred 237.7 GiB of 250.0 GiB (95.07%)
transferred 240.2 GiB of 250.0 GiB (96.07%)
transferred 242.7 GiB of 250.0 GiB (97.08%)
transferred 245.2 GiB of 250.0 GiB (98.08%)
transferred 247.7 GiB of 250.0 GiB (99.08%)
transferred 250.0 GiB of 250.0 GiB (100.00%)
transferred 250.0 GiB of 250.0 GiB (100.00%)
create full clone of drive scsi1 (NVMe:vm-234-disk-2)
  Logical volume "vm-242-disk-2" created.
transferred 0.0 B of 60.0 GiB (0.00%)
transferred 614.4 MiB of 60.0 GiB (1.00%)
transferred 1.2 GiB of 60.0 GiB (2.01%)
transferred 1.8 GiB of 60.0 GiB (3.01%)
transferred 2.4 GiB of 60.0 GiB (4.01%)
transferred 3.0 GiB of 60.0 GiB (5.01%)
transferred 3.6 GiB of 60.0 GiB (6.02%)
transferred 4.2 GiB of 60.0 GiB (7.02%)
transferred 4.8 GiB of 60.0 GiB (8.02%)
transferred 5.4 GiB of 60.0 GiB (9.02%)
transferred 6.0 GiB of 60.0 GiB (10.03%)
transferred 6.6 GiB of 60.0 GiB (11.03%)
transferred 7.2 GiB of 60.0 GiB (12.03%)
transferred 7.8 GiB of 60.0 GiB (13.03%)
transferred 8.4 GiB of 60.0 GiB (14.04%)
transferred 9.0 GiB of 60.0 GiB (15.04%)
transferred 9.6 GiB of 60.0 GiB (16.04%)
transferred 10.2 GiB of 60.0 GiB (17.04%)
transferred 10.8 GiB of 60.0 GiB (18.05%)
transferred 11.4 GiB of 60.0 GiB (19.05%)
transferred 12.0 GiB of 60.0 GiB (20.05%)
transferred 12.6 GiB of 60.0 GiB (21.05%)
transferred 13.2 GiB of 60.0 GiB (22.06%)
transferred 13.8 GiB of 60.0 GiB (23.06%)
transferred 14.4 GiB of 60.0 GiB (24.06%)
transferred 15.0 GiB of 60.0 GiB (25.07%)
transferred 15.6 GiB of 60.0 GiB (26.07%)
transferred 16.2 GiB of 60.0 GiB (27.07%)
transferred 16.8 GiB of 60.0 GiB (28.07%)
transferred 17.4 GiB of 60.0 GiB (29.08%)
transferred 18.0 GiB of 60.0 GiB (30.08%)
transferred 18.6 GiB of 60.0 GiB (31.08%)
transferred 19.2 GiB of 60.0 GiB (32.08%)
transferred 19.9 GiB of 60.0 GiB (33.09%)
transferred 20.5 GiB of 60.0 GiB (34.09%)
transferred 21.1 GiB of 60.0 GiB (35.09%)
transferred 21.7 GiB of 60.0 GiB (36.09%)
transferred 22.3 GiB of 60.0 GiB (37.10%)
transferred 22.9 GiB of 60.0 GiB (38.10%)
transferred 23.5 GiB of 60.0 GiB (39.10%)
transferred 24.1 GiB of 60.0 GiB (40.10%)
transferred 24.7 GiB of 60.0 GiB (41.11%)
transferred 25.3 GiB of 60.0 GiB (42.11%)
transferred 25.9 GiB of 60.0 GiB (43.11%)
transferred 26.5 GiB of 60.0 GiB (44.11%)
transferred 27.1 GiB of 60.0 GiB (45.12%)
transferred 27.7 GiB of 60.0 GiB (46.12%)
transferred 28.3 GiB of 60.0 GiB (47.12%)
transferred 28.9 GiB of 60.0 GiB (48.12%)
transferred 29.5 GiB of 60.0 GiB (49.13%)
transferred 30.1 GiB of 60.0 GiB (50.13%)
transferred 30.7 GiB of 60.0 GiB (51.13%)
transferred 31.3 GiB of 60.0 GiB (52.14%)
transferred 31.9 GiB of 60.0 GiB (53.14%)
transferred 32.5 GiB of 60.0 GiB (54.14%)
transferred 33.1 GiB of 60.0 GiB (55.14%)
transferred 33.7 GiB of 60.0 GiB (56.15%)
transferred 34.3 GiB of 60.0 GiB (57.15%)
transferred 34.9 GiB of 60.0 GiB (58.15%)
transferred 35.5 GiB of 60.0 GiB (59.15%)
transferred 36.1 GiB of 60.0 GiB (60.16%)
transferred 36.7 GiB of 60.0 GiB (61.16%)
transferred 37.3 GiB of 60.0 GiB (62.16%)
transferred 37.9 GiB of 60.0 GiB (63.16%)
transferred 38.5 GiB of 60.0 GiB (64.17%)
transferred 39.1 GiB of 60.0 GiB (65.17%)
transferred 39.7 GiB of 60.0 GiB (66.17%)
transferred 40.3 GiB of 60.0 GiB (67.17%)
transferred 40.9 GiB of 60.0 GiB (68.18%)
transferred 41.5 GiB of 60.0 GiB (69.18%)
transferred 42.1 GiB of 60.0 GiB (70.18%)
transferred 42.7 GiB of 60.0 GiB (71.18%)
transferred 43.3 GiB of 60.0 GiB (72.19%)
transferred 43.9 GiB of 60.0 GiB (73.19%)
transferred 44.5 GiB of 60.0 GiB (74.19%)
transferred 45.1 GiB of 60.0 GiB (75.20%)
transferred 45.7 GiB of 60.0 GiB (76.20%)
transferred 46.3 GiB of 60.0 GiB (77.20%)
transferred 46.9 GiB of 60.0 GiB (78.20%)
transferred 47.5 GiB of 60.0 GiB (79.21%)
transferred 48.1 GiB of 60.0 GiB (80.21%)
transferred 48.7 GiB of 60.0 GiB (81.21%)
transferred 49.3 GiB of 60.0 GiB (82.21%)
transferred 49.9 GiB of 60.0 GiB (83.22%)
transferred 50.5 GiB of 60.0 GiB (84.22%)
transferred 51.1 GiB of 60.0 GiB (85.22%)
transferred 51.7 GiB of 60.0 GiB (86.22%)
transferred 52.3 GiB of 60.0 GiB (87.23%)
transferred 52.9 GiB of 60.0 GiB (88.23%)
transferred 53.5 GiB of 60.0 GiB (89.23%)
transferred 54.1 GiB of 60.0 GiB (90.23%)
transferred 54.7 GiB of 60.0 GiB (91.24%)
transferred 55.3 GiB of 60.0 GiB (92.24%)
transferred 55.9 GiB of 60.0 GiB (93.24%)
transferred 56.5 GiB of 60.0 GiB (94.24%)
transferred 57.1 GiB of 60.0 GiB (95.25%)
transferred 57.8 GiB of 60.0 GiB (96.25%)
transferred 58.3 GiB of 60.0 GiB (97.25%)
transferred 59.0 GiB of 60.0 GiB (98.26%)
transferred 59.6 GiB of 60.0 GiB (99.26%)
transferred 60.0 GiB of 60.0 GiB (100.00%)
transferred 60.0 GiB of 60.0 GiB (100.00%)
create full clone of drive scsi2 (NVMe:vm-234-disk-3)
  Wiping dos signature on /dev/nvme_vg/vm-242-disk-3.
  Logical volume "vm-242-disk-3" created.
transferred 0.0 B of 60.0 GiB (0.00%)
transferred 614.4 MiB of 60.0 GiB (1.00%)
transferred 1.2 GiB of 60.0 GiB (2.01%)
transferred 1.8 GiB of 60.0 GiB (3.01%)
transferred 2.4 GiB of 60.0 GiB (4.01%)
transferred 3.0 GiB of 60.0 GiB (5.01%)
transferred 3.6 GiB of 60.0 GiB (6.02%)
transferred 4.2 GiB of 60.0 GiB (7.02%)
transferred 4.8 GiB of 60.0 GiB (8.02%)
transferred 5.4 GiB of 60.0 GiB (9.02%)
transferred 6.0 GiB of 60.0 GiB (10.03%)
transferred 6.6 GiB of 60.0 GiB (11.03%)
transferred 7.2 GiB of 60.0 GiB (12.03%)
transferred 7.8 GiB of 60.0 GiB (13.03%)
transferred 8.4 GiB of 60.0 GiB (14.04%)
transferred 9.0 GiB of 60.0 GiB (15.04%)
transferred 9.6 GiB of 60.0 GiB (16.04%)
transferred 10.2 GiB of 60.0 GiB (17.04%)
transferred 10.8 GiB of 60.0 GiB (18.05%)
transferred 11.4 GiB of 60.0 GiB (19.05%)
transferred 12.0 GiB of 60.0 GiB (20.05%)
transferred 12.6 GiB of 60.0 GiB (21.05%)
transferred 13.2 GiB of 60.0 GiB (22.06%)
transferred 13.8 GiB of 60.0 GiB (23.06%)
transferred 14.4 GiB of 60.0 GiB (24.06%)
transferred 15.0 GiB of 60.0 GiB (25.07%)
transferred 15.6 GiB of 60.0 GiB (26.07%)
transferred 16.2 GiB of 60.0 GiB (27.07%)
transferred 16.8 GiB of 60.0 GiB (28.07%)
transferred 17.4 GiB of 60.0 GiB (29.08%)
transferred 18.0 GiB of 60.0 GiB (30.08%)
transferred 18.6 GiB of 60.0 GiB (31.08%)
transferred 19.2 GiB of 60.0 GiB (32.08%)
transferred 19.9 GiB of 60.0 GiB (33.09%)
transferred 20.5 GiB of 60.0 GiB (34.09%)
transferred 21.1 GiB of 60.0 GiB (35.09%)
transferred 21.7 GiB of 60.0 GiB (36.09%)
transferred 22.3 GiB of 60.0 GiB (37.10%)
transferred 22.9 GiB of 60.0 GiB (38.10%)
transferred 23.5 GiB of 60.0 GiB (39.10%)
transferred 24.1 GiB of 60.0 GiB (40.10%)
transferred 24.7 GiB of 60.0 GiB (41.11%)
transferred 25.3 GiB of 60.0 GiB (42.11%)
transferred 25.9 GiB of 60.0 GiB (43.11%)
transferred 26.5 GiB of 60.0 GiB (44.11%)
transferred 27.1 GiB of 60.0 GiB (45.12%)
transferred 27.7 GiB of 60.0 GiB (46.12%)
transferred 28.3 GiB of 60.0 GiB (47.12%)
transferred 28.9 GiB of 60.0 GiB (48.12%)
transferred 29.5 GiB of 60.0 GiB (49.13%)
transferred 30.1 GiB of 60.0 GiB (50.13%)
transferred 30.7 GiB of 60.0 GiB (51.13%)
transferred 31.3 GiB of 60.0 GiB (52.14%)
transferred 31.9 GiB of 60.0 GiB (53.14%)
transferred 32.5 GiB of 60.0 GiB (54.14%)
transferred 33.1 GiB of 60.0 GiB (55.14%)
transferred 33.7 GiB of 60.0 GiB (56.15%)
transferred 34.3 GiB of 60.0 GiB (57.15%)
transferred 34.9 GiB of 60.0 GiB (58.15%)
transferred 35.5 GiB of 60.0 GiB (59.15%)
transferred 36.1 GiB of 60.0 GiB (60.16%)
transferred 36.7 GiB of 60.0 GiB (61.16%)
transferred 37.3 GiB of 60.0 GiB (62.16%)
transferred 37.9 GiB of 60.0 GiB (63.16%)
transferred 38.5 GiB of 60.0 GiB (64.17%)
transferred 39.1 GiB of 60.0 GiB (65.17%)
transferred 39.7 GiB of 60.0 GiB (66.17%)
transferred 40.3 GiB of 60.0 GiB (67.17%)
transferred 40.9 GiB of 60.0 GiB (68.18%)
transferred 41.5 GiB of 60.0 GiB (69.18%)
transferred 42.1 GiB of 60.0 GiB (70.18%)
transferred 42.7 GiB of 60.0 GiB (71.18%)
transferred 43.3 GiB of 60.0 GiB (72.19%)
transferred 43.9 GiB of 60.0 GiB (73.19%)
transferred 44.5 GiB of 60.0 GiB (74.19%)
transferred 45.1 GiB of 60.0 GiB (75.20%)
transferred 45.7 GiB of 60.0 GiB (76.20%)
transferred 46.3 GiB of 60.0 GiB (77.20%)
transferred 46.9 GiB of 60.0 GiB (78.20%)
transferred 47.5 GiB of 60.0 GiB (79.21%)
transferred 48.1 GiB of 60.0 GiB (80.21%)
transferred 48.7 GiB of 60.0 GiB (81.21%)
transferred 49.3 GiB of 60.0 GiB (82.21%)
transferred 49.9 GiB of 60.0 GiB (83.22%)
transferred 50.5 GiB of 60.0 GiB (84.22%)
transferred 51.1 GiB of 60.0 GiB (85.22%)
transferred 51.7 GiB of 60.0 GiB (86.22%)
transferred 52.3 GiB of 60.0 GiB (87.23%)
transferred 52.9 GiB of 60.0 GiB (88.23%)
transferred 53.5 GiB of 60.0 GiB (89.23%)
transferred 54.1 GiB of 60.0 GiB (90.23%)
transferred 54.7 GiB of 60.0 GiB (91.24%)
transferred 55.3 GiB of 60.0 GiB (92.24%)
transferred 55.9 GiB of 60.0 GiB (93.24%)
transferred 56.5 GiB of 60.0 GiB (94.24%)
transferred 57.1 GiB of 60.0 GiB (95.25%)
transferred 57.8 GiB of 60.0 GiB (96.25%)
transferred 58.3 GiB of 60.0 GiB (97.25%)
transferred 59.0 GiB of 60.0 GiB (98.26%)
transferred 59.6 GiB of 60.0 GiB (99.26%)
transferred 60.0 GiB of 60.0 GiB (100.00%)
transferred 60.0 GiB of 60.0 GiB (100.00%)
TASK OK

which took 1h 36m 24s
 
the whole clone took 1,5h? that's around 500mbit/s of throughput, depending on your network setup that might just be the bottle neck.. how is the connectivity to both storages?
 
the whole clone took 1,5h? that's around 500mbit/s of throughput, depending on your network setup that might just be the bottle neck.. how is the connectivity to both storages?
The NetApp storage device is connected via 2x 100G > 4x 25G cables, and all proxmox nodes are connected via 2x10G to the switch
 
sorry, I didn't read the log closely enough!

could you try benchmarking the performance of an LV on your LVM storage (directly on the host) and then of a volume allocated to a VM from within the VM?

keep in mind that a clone will incur a double network penalty, as it has to read from the source volume and write to the target volume, so all data flows from your storage to your PVE node and back again..
 
LV benchmark:

sequential read/write

Code:
root@pve01:~# fio --name=seqtest --filename=/dev/nvme_prd/testlv --rw=readwrite --bs=1M --size=4G --numjobs=1 --direct=1 --iodepth=16 --runtime=60 --time_based
seqtest: (g=0): rw=rw, bs=(R) 1024KiB-1024KiB, (W) 1024KiB-1024KiB, (T) 1024KiB-1024KiB, ioengine=psync, iodepth=16
fio-3.33
Starting 1 process
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
Jobs: 1 (f=1): [M(1)][100.0%][r=152MiB/s,w=195MiB/s][r=152,w=195 IOPS][eta 00m:00s]
seqtest: (groupid=0, jobs=1): err= 0: pid=430767: Wed Oct 22 10:01:15 2025
  read: IOPS=250, BW=251MiB/s (263MB/s)(14.7GiB/60001msec)
    clat (usec): min=1163, max=355105, avg=1768.26, stdev=6043.47
     lat (usec): min=1163, max=355106, avg=1768.68, stdev=6043.47
    clat percentiles (usec):
     |  1.00th=[  1221],  5.00th=[  1254], 10.00th=[  1270], 20.00th=[  1287],
     | 30.00th=[  1319], 40.00th=[  1336], 50.00th=[  1385], 60.00th=[  1631],
     | 70.00th=[  1729], 80.00th=[  1778], 90.00th=[  1893], 95.00th=[  2311],
     | 99.00th=[  4752], 99.50th=[  6652], 99.90th=[ 17695], 99.95th=[145753],
     | 99.99th=[274727]
   bw (  KiB/s): min=51200, max=344064, per=100.00%, avg=256980.97, stdev=53818.92, samples=119
   iops        : min=   50, max=  336, avg=250.96, stdev=52.56, samples=119
  write: IOPS=250, BW=250MiB/s (263MB/s)(14.7GiB/60001msec); 0 zone resets
    clat (usec): min=1721, max=29308, avg=2158.65, stdev=878.32
     lat (usec): min=1747, max=29400, avg=2215.54, stdev=880.54
    clat percentiles (usec):
     |  1.00th=[ 1795],  5.00th=[ 1844], 10.00th=[ 1860], 20.00th=[ 1893],
     | 30.00th=[ 1926], 40.00th=[ 1958], 50.00th=[ 1991], 60.00th=[ 2040],
     | 70.00th=[ 2073], 80.00th=[ 2147], 90.00th=[ 2343], 95.00th=[ 2933],
     | 99.00th=[ 5473], 99.50th=[ 7242], 99.90th=[14091], 99.95th=[17695],
     | 99.99th=[24249]
   bw (  KiB/s): min=40960, max=325632, per=100.00%, avg=256447.46, stdev=53860.06, samples=119
   iops        : min=   40, max=  318, avg=250.44, stdev=52.60, samples=119
  lat (msec)   : 2=71.59%, 4=26.52%, 10=1.62%, 20=0.21%, 50=0.02%
  lat (msec)   : 100=0.01%, 250=0.01%, 500=0.01%
  cpu          : usr=1.83%, sys=5.35%, ctx=30094, majf=0, minf=1055
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=15047,15023,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=16

Run status group 0 (all jobs):
   READ: bw=251MiB/s (263MB/s), 251MiB/s-251MiB/s (263MB/s-263MB/s), io=14.7GiB (15.8GB), run=60001-60001msec
  WRITE: bw=250MiB/s (263MB/s), 250MiB/s-250MiB/s (263MB/s-263MB/s), io=14.7GiB (15.8GB), run=60001-60001msec

random I/O
Code:
root@pve01:~# fio --name=randtest --filename=/dev/nvme_prd/testlv --rw=randrw --bs=4k --size=4G --numjobs=4 --iodepth=32 --runtime=60 --time_based --direct=1
randtest: (g=0): rw=randrw, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=psync, iodepth=32
...
fio-3.33
Starting 4 processes
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
Jobs: 4 (f=4): [m(4)][100.0%][r=49.2MiB/s,w=49.8MiB/s][r=12.6k,w=12.8k IOPS][eta 00m:00s]
randtest: (groupid=0, jobs=1): err= 0: pid=433881: Wed Oct 22 10:03:15 2025
  read: IOPS=3090, BW=12.1MiB/s (12.7MB/s)(724MiB/60001msec)
    clat (usec): min=57, max=6868, avg=129.87, stdev=73.01
     lat (usec): min=57, max=6869, avg=130.07, stdev=73.02
    clat percentiles (usec):
     |  1.00th=[   76],  5.00th=[   84], 10.00th=[   90], 20.00th=[   99],
     | 30.00th=[  109], 40.00th=[  117], 50.00th=[  123], 60.00th=[  129],
     | 70.00th=[  137], 80.00th=[  149], 90.00th=[  169], 95.00th=[  202],
     | 99.00th=[  269], 99.50th=[  297], 99.90th=[  766], 99.95th=[ 1303],
     | 99.99th=[ 3490]
   bw (  KiB/s): min= 9856, max=13792, per=25.27%, avg=12364.98, stdev=787.96, samples=119
   iops        : min= 2464, max= 3448, avg=3091.21, stdev=196.98, samples=119
  write: IOPS=3087, BW=12.1MiB/s (12.6MB/s)(724MiB/60001msec); 0 zone resets
    clat (usec): min=101, max=12812, avg=190.51, stdev=122.09
     lat (usec): min=102, max=12812, avg=190.78, stdev=122.10
    clat percentiles (usec):
     |  1.00th=[  126],  5.00th=[  137], 10.00th=[  145], 20.00th=[  153],
     | 30.00th=[  163], 40.00th=[  172], 50.00th=[  180], 60.00th=[  188],
     | 70.00th=[  196], 80.00th=[  206], 90.00th=[  229], 95.00th=[  260],
     | 99.00th=[  388], 99.50th=[  562], 99.90th=[ 1713], 99.95th=[ 3032],
     | 99.99th=[ 4686]
   bw (  KiB/s): min=10280, max=13576, per=25.18%, avg=12346.25, stdev=737.76, samples=119
   iops        : min= 2570, max= 3394, avg=3086.52, stdev=184.42, samples=119
  lat (usec)   : 100=10.58%, 250=85.68%, 500=3.35%, 750=0.17%, 1000=0.07%
  lat (msec)   : 2=0.10%, 4=0.04%, 10=0.01%, 20=0.01%
  cpu          : usr=1.91%, sys=6.51%, ctx=370706, majf=0, minf=145
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=185438,185239,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32
randtest: (groupid=0, jobs=1): err= 0: pid=433882: Wed Oct 22 10:03:15 2025
  read: IOPS=3060, BW=12.0MiB/s (12.5MB/s)(717MiB/60001msec)
    clat (usec): min=59, max=6157, avg=130.36, stdev=72.06
     lat (usec): min=59, max=6158, avg=130.57, stdev=72.07
    clat percentiles (usec):
     |  1.00th=[   76],  5.00th=[   85], 10.00th=[   91], 20.00th=[  100],
     | 30.00th=[  110], 40.00th=[  117], 50.00th=[  123], 60.00th=[  129],
     | 70.00th=[  137], 80.00th=[  149], 90.00th=[  169], 95.00th=[  202],
     | 99.00th=[  265], 99.50th=[  293], 99.90th=[  766], 99.95th=[ 1237],
     | 99.99th=[ 3490]
   bw (  KiB/s): min=10080, max=13776, per=25.02%, avg=12242.76, stdev=828.50, samples=119
   iops        : min= 2520, max= 3444, avg=3060.63, stdev=207.14, samples=119
  write: IOPS=3074, BW=12.0MiB/s (12.6MB/s)(721MiB/60001msec); 0 zone resets
    clat (usec): min=106, max=10615, avg=192.01, stdev=121.40
     lat (usec): min=106, max=10615, avg=192.30, stdev=121.42
    clat percentiles (usec):
     |  1.00th=[  127],  5.00th=[  139], 10.00th=[  145], 20.00th=[  155],
     | 30.00th=[  165], 40.00th=[  174], 50.00th=[  182], 60.00th=[  188],
     | 70.00th=[  196], 80.00th=[  208], 90.00th=[  231], 95.00th=[  260],
     | 99.00th=[  392], 99.50th=[  562], 99.90th=[ 1729], 99.95th=[ 3032],
     | 99.99th=[ 4621]
   bw (  KiB/s): min=10192, max=13520, per=25.08%, avg=12298.17, stdev=767.49, samples=119
   iops        : min= 2548, max= 3380, avg=3074.49, stdev=191.86, samples=119
  lat (usec)   : 100=9.86%, 250=86.39%, 500=3.36%, 750=0.18%, 1000=0.07%
  lat (msec)   : 2=0.10%, 4=0.04%, 10=0.01%, 20=0.01%
  cpu          : usr=2.04%, sys=6.50%, ctx=368112, majf=0, minf=146
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=183618,184479,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32
randtest: (groupid=0, jobs=1): err= 0: pid=433883: Wed Oct 22 10:03:15 2025
  read: IOPS=3076, BW=12.0MiB/s (12.6MB/s)(721MiB/60001msec)
    clat (usec): min=58, max=5614, avg=129.86, stdev=69.94
     lat (usec): min=58, max=5615, avg=130.07, stdev=69.95
    clat percentiles (usec):
     |  1.00th=[   76],  5.00th=[   85], 10.00th=[   90], 20.00th=[  100],
     | 30.00th=[  110], 40.00th=[  117], 50.00th=[  123], 60.00th=[  129],
     | 70.00th=[  137], 80.00th=[  149], 90.00th=[  169], 95.00th=[  202],
     | 99.00th=[  269], 99.50th=[  302], 99.90th=[  693], 99.95th=[ 1139],
     | 99.99th=[ 3425]
   bw (  KiB/s): min= 9672, max=13656, per=25.16%, avg=12307.88, stdev=787.14, samples=119
   iops        : min= 2418, max= 3414, avg=3076.92, stdev=196.78, samples=119
  write: IOPS=3077, BW=12.0MiB/s (12.6MB/s)(721MiB/60001msec); 0 zone resets
    clat (usec): min=105, max=9872, avg=191.66, stdev=117.74
     lat (usec): min=105, max=9872, avg=191.95, stdev=117.75
    clat percentiles (usec):
     |  1.00th=[  127],  5.00th=[  137], 10.00th=[  145], 20.00th=[  155],
     | 30.00th=[  163], 40.00th=[  174], 50.00th=[  182], 60.00th=[  188],
     | 70.00th=[  196], 80.00th=[  208], 90.00th=[  229], 95.00th=[  260],
     | 99.00th=[  392], 99.50th=[  578], 99.90th=[ 1762], 99.95th=[ 2835],
     | 99.99th=[ 4293]
   bw (  KiB/s): min=10248, max=13643, per=25.10%, avg=12308.88, stdev=782.06, samples=119
   iops        : min= 2562, max= 3410, avg=3077.18, stdev=195.48, samples=119
  lat (usec)   : 100=10.24%, 250=86.03%, 500=3.32%, 750=0.19%, 1000=0.06%
  lat (msec)   : 2=0.10%, 4=0.04%, 10=0.01%
  cpu          : usr=2.00%, sys=6.53%, ctx=369259, majf=0, minf=144
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=184598,184635,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32
randtest: (groupid=0, jobs=1): err= 0: pid=433884: Wed Oct 22 10:03:15 2025
  read: IOPS=3003, BW=11.7MiB/s (12.3MB/s)(704MiB/60001msec)
    clat (usec): min=59, max=9499, avg=136.18, stdev=87.06
     lat (usec): min=59, max=9499, avg=136.39, stdev=87.07
    clat percentiles (usec):
     |  1.00th=[   79],  5.00th=[   88], 10.00th=[   94], 20.00th=[  104],
     | 30.00th=[  113], 40.00th=[  120], 50.00th=[  126], 60.00th=[  135],
     | 70.00th=[  145], 80.00th=[  157], 90.00th=[  180], 95.00th=[  208],
     | 99.00th=[  277], 99.50th=[  318], 99.90th=[ 1004], 99.95th=[ 1811],
     | 99.99th=[ 3752]
   bw (  KiB/s): min= 9912, max=13624, per=24.55%, avg=12013.56, stdev=720.62, samples=119
   iops        : min= 2478, max= 3406, avg=3003.36, stdev=180.16, samples=119
  write: IOPS=3018, BW=11.8MiB/s (12.4MB/s)(707MiB/60001msec); 0 zone resets
    clat (usec): min=103, max=12212, avg=192.24, stdev=125.57
     lat (usec): min=104, max=12213, avg=192.53, stdev=125.58
    clat percentiles (usec):
     |  1.00th=[  129],  5.00th=[  141], 10.00th=[  147], 20.00th=[  157],
     | 30.00th=[  165], 40.00th=[  172], 50.00th=[  180], 60.00th=[  188],
     | 70.00th=[  196], 80.00th=[  206], 90.00th=[  229], 95.00th=[  262],
     | 99.00th=[  404], 99.50th=[  603], 99.90th=[ 1827], 99.95th=[ 2999],
     | 99.99th=[ 4178]
   bw (  KiB/s): min= 9512, max=13472, per=24.62%, avg=12069.19, stdev=728.16, samples=119
   iops        : min= 2378, max= 3368, avg=3017.27, stdev=182.03, samples=119
  lat (usec)   : 100=8.02%, 250=88.01%, 500=3.52%, 750=0.21%, 1000=0.07%
  lat (msec)   : 2=0.11%, 4=0.06%, 10=0.01%, 20=0.01%
  cpu          : usr=2.04%, sys=6.73%, ctx=361345, majf=0, minf=97
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=180222,181091,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32

Run status group 0 (all jobs):
   READ: bw=47.8MiB/s (50.1MB/s), 11.7MiB/s-12.1MiB/s (12.3MB/s-12.7MB/s), io=2867MiB (3006MB), run=60001-60001msec
  WRITE: bw=47.9MiB/s (50.2MB/s), 11.8MiB/s-12.1MiB/s (12.4MB/s-12.6MB/s), io=2873MiB (3012MB), run=60001-60001msec
 
VM Benchmark:

sequential read/write

Code:
root@tools:~ # fio --name=vm-test --filename=/tmp/testfile --rw=readwrite --bs=1M --size=1G --numjobs=1 --direct=1 --iodepth=16 --runtime=60 --time_based
vm-test: (g=0): rw=rw, bs=(R) 1024KiB-1024KiB, (W) 1024KiB-1024KiB, (T) 1024KiB-1024KiB, ioengine=psync, iodepth=16
fio-3.40
Starting 1 process
vm-test: Laying out IO file (1 file / 1024MiB)
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
Jobs: 1 (f=1): [M(1)][100.0%][r=383MiB/s,w=376MiB/s][r=383,w=376 IOPS][eta 00m:00s]
vm-test: (groupid=0, jobs=1): err= 0: pid=986: Wed Oct 22 10:09:19 2025
  read: IOPS=352, BW=352MiB/s (369MB/s)(20.6GiB/60001msec)
    clat (usec): min=67, max=17665, avg=350.36, stdev=322.84
     lat (usec): min=67, max=17665, avg=350.72, stdev=322.86
    clat percentiles (usec):
     |  1.00th=[  139],  5.00th=[  153], 10.00th=[  163], 20.00th=[  182],
     | 30.00th=[  200], 40.00th=[  225], 50.00th=[  310], 60.00th=[  375],
     | 70.00th=[  416], 80.00th=[  474], 90.00th=[  562], 95.00th=[  644],
     | 99.00th=[ 1004], 99.50th=[ 1827], 99.90th=[ 4293], 99.95th=[ 5473],
     | 99.99th=[10159]
   bw (  KiB/s): min=217088, max=706560, per=100.00%, avg=360928.64, stdev=54770.17, samples=119
   iops        : min=  212, max=  690, avg=352.45, stdev=53.49, samples=119
  write: IOPS=352, BW=352MiB/s (369MB/s)(20.6GiB/60001msec); 0 zone resets
    clat (usec): min=85, max=14098, avg=2457.00, stdev=559.67
     lat (usec): min=103, max=14113, avg=2483.75, stdev=561.71
    clat percentiles (usec):
     |  1.00th=[  668],  5.00th=[ 1893], 10.00th=[ 2024], 20.00th=[ 2147],
     | 30.00th=[ 2245], 40.00th=[ 2311], 50.00th=[ 2376], 60.00th=[ 2474],
     | 70.00th=[ 2573], 80.00th=[ 2737], 90.00th=[ 2999], 95.00th=[ 3359],
     | 99.00th=[ 4424], 99.50th=[ 4686], 99.90th=[ 5538], 99.95th=[ 6980],
     | 99.99th=[11731]
   bw (  KiB/s): min=256000, max=761856, per=100.00%, avg=360911.81, stdev=48532.41, samples=119
   iops        : min=  250, max=  744, avg=352.45, stdev=47.40, samples=119
  lat (usec)   : 100=0.13%, 250=23.23%, 500=18.94%, 750=7.11%, 1000=0.87%
  lat (msec)   : 2=4.01%, 4=44.78%, 10=0.92%, 20=0.02%
  cpu          : usr=1.17%, sys=21.39%, ctx=55982, majf=0, minf=0
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=21137,21123,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=16

Run status group 0 (all jobs):
   READ: bw=352MiB/s (369MB/s), 352MiB/s-352MiB/s (369MB/s-369MB/s), io=20.6GiB (22.2GB), run=60001-60001msec
  WRITE: bw=352MiB/s (369MB/s), 352MiB/s-352MiB/s (369MB/s-369MB/s), io=20.6GiB (22.1GB), run=60001-60001msec

random I/O
Code:
root@tools:~ # fio --name=rand --filename=/tmp/testfile --rw=randrw --bs=4k --size=1G --numjobs=4 --iodepth=32 --runtime=60 --time_based --direct=1
rand: (g=0): rw=randrw, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=psync, iodepth=32
...
fio-3.40
Starting 4 processes
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
note: both iodepth >= 1 and synchronous I/O engine are selected, queue depth will be capped at 1
Jobs: 4 (f=4): [m(4)][100.0%][r=12.1MiB/s,w=12.4MiB/s][r=3103,w=3184 IOPS][eta 00m:00s]
rand: (groupid=0, jobs=1): err= 0: pid=992: Wed Oct 22 10:10:55 2025
  read: IOPS=753, BW=3015KiB/s (3087kB/s)(177MiB/60001msec)
    clat (nsec): min=1644, max=31595k, avg=42007.67, stdev=205464.47
     lat (nsec): min=1712, max=31595k, avg=42188.75, stdev=205467.34
    clat percentiles (usec):
     |  1.00th=[    3],  5.00th=[    4], 10.00th=[    8], 20.00th=[   34],
     | 30.00th=[   37], 40.00th=[   38], 50.00th=[   40], 60.00th=[   41],
     | 70.00th=[   44], 80.00th=[   47], 90.00th=[   54], 95.00th=[   65],
     | 99.00th=[  102], 99.50th=[  133], 99.90th=[  635], 99.95th=[ 1270],
     | 99.99th=[ 3621]
   bw (  KiB/s): min= 2088, max= 6480, per=25.25%, avg=3017.55, stdev=469.90, samples=119
   iops        : min=  522, max= 1620, avg=754.39, stdev=117.47, samples=119
  write: IOPS=749, BW=2998KiB/s (3070kB/s)(176MiB/60001msec); 0 zone resets
    clat (usec): min=4, max=27253, avg=1288.48, stdev=400.24
     lat (usec): min=4, max=27253, avg=1288.69, stdev=400.26
    clat percentiles (usec):
     |  1.00th=[   53],  5.00th=[ 1029], 10.00th=[ 1090], 20.00th=[ 1139],
     | 30.00th=[ 1188], 40.00th=[ 1221], 50.00th=[ 1254], 60.00th=[ 1303],
     | 70.00th=[ 1352], 80.00th=[ 1418], 90.00th=[ 1549], 95.00th=[ 1696],
     | 99.00th=[ 2008], 99.50th=[ 2278], 99.90th=[ 4817], 99.95th=[ 5800],
     | 99.99th=[12780]
   bw (  KiB/s): min= 1976, max= 6840, per=25.10%, avg=2998.39, stdev=436.58, samples=119
   iops        : min=  494, max= 1710, avg=749.60, stdev=109.15, samples=119
  lat (usec)   : 2=0.01%, 4=2.52%, 10=3.06%, 20=0.23%, 50=37.76%
  lat (usec)   : 100=6.63%, 250=0.52%, 500=0.25%, 750=0.11%, 1000=1.02%
  lat (msec)   : 2=47.37%, 4=0.42%, 10=0.09%, 20=0.01%, 50=0.01%
  cpu          : usr=0.39%, sys=6.57%, ctx=57249, majf=0, minf=0
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=45219,44973,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32
rand: (groupid=0, jobs=1): err= 0: pid=993: Wed Oct 22 10:10:55 2025
  read: IOPS=748, BW=2994KiB/s (3066kB/s)(175MiB/60004msec)
    clat (nsec): min=1577, max=63625k, avg=42889.33, stdev=318564.37
     lat (nsec): min=1632, max=63626k, avg=43072.26, stdev=318566.98
    clat percentiles (usec):
     |  1.00th=[    3],  5.00th=[    5], 10.00th=[    8], 20.00th=[   35],
     | 30.00th=[   37], 40.00th=[   38], 50.00th=[   40], 60.00th=[   41],
     | 70.00th=[   44], 80.00th=[   47], 90.00th=[   55], 95.00th=[   65],
     | 99.00th=[  102], 99.50th=[  133], 99.90th=[  766], 99.95th=[ 1401],
     | 99.99th=[ 3490]
   bw (  KiB/s): min= 1848, max= 7920, per=25.05%, avg=2993.61, stdev=554.39, samples=119
   iops        : min=  462, max= 1980, avg=748.40, stdev=138.60, samples=119
  write: IOPS=753, BW=3013KiB/s (3086kB/s)(177MiB/60004msec); 0 zone resets
    clat (usec): min=3, max=26199, avg=1281.35, stdev=378.13
     lat (usec): min=3, max=26199, avg=1281.56, stdev=378.16
    clat percentiles (usec):
     |  1.00th=[   41],  5.00th=[ 1020], 10.00th=[ 1090], 20.00th=[ 1139],
     | 30.00th=[ 1188], 40.00th=[ 1221], 50.00th=[ 1254], 60.00th=[ 1303],
     | 70.00th=[ 1352], 80.00th=[ 1418], 90.00th=[ 1549], 95.00th=[ 1696],
     | 99.00th=[ 2008], 99.50th=[ 2278], 99.90th=[ 4621], 99.95th=[ 5473],
     | 99.99th=[11863]
   bw (  KiB/s): min= 1976, max= 8496, per=25.22%, avg=3013.92, stdev=567.24, samples=119
   iops        : min=  494, max= 2124, avg=753.48, stdev=141.81, samples=119
  lat (usec)   : 2=0.01%, 4=2.47%, 10=2.94%, 20=0.21%, 50=38.07%
  lat (usec)   : 100=6.49%, 250=0.49%, 500=0.21%, 750=0.11%, 1000=0.98%
  lat (msec)   : 2=47.50%, 4=0.41%, 10=0.09%, 20=0.02%, 50=0.01%
  lat (msec)   : 100=0.01%
  cpu          : usr=0.40%, sys=6.56%, ctx=57377, majf=0, minf=0
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=44917,45204,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32
rand: (groupid=0, jobs=1): err= 0: pid=994: Wed Oct 22 10:10:55 2025
  read: IOPS=748, BW=2993KiB/s (3065kB/s)(175MiB/60001msec)
    clat (nsec): min=1580, max=29939k, avg=41642.59, stdev=211598.81
     lat (nsec): min=1639, max=29940k, avg=41825.67, stdev=211601.70
    clat percentiles (usec):
     |  1.00th=[    3],  5.00th=[    4], 10.00th=[    7], 20.00th=[   34],
     | 30.00th=[   37], 40.00th=[   38], 50.00th=[   39], 60.00th=[   41],
     | 70.00th=[   44], 80.00th=[   47], 90.00th=[   55], 95.00th=[   65],
     | 99.00th=[   99], 99.50th=[  125], 99.90th=[  255], 99.95th=[  457],
     | 99.99th=[10552]
   bw (  KiB/s): min= 1824, max= 5192, per=25.04%, avg=2991.39, stdev=404.27, samples=119
   iops        : min=  456, max= 1298, avg=747.85, stdev=101.07, samples=119
  write: IOPS=747, BW=2989KiB/s (3060kB/s)(175MiB/60001msec); 0 zone resets
    clat (usec): min=3, max=25767, avg=1292.00, stdev=357.52
     lat (usec): min=3, max=25768, avg=1292.21, stdev=357.54
    clat percentiles (usec):
     |  1.00th=[  306],  5.00th=[ 1029], 10.00th=[ 1090], 20.00th=[ 1139],
     | 30.00th=[ 1188], 40.00th=[ 1221], 50.00th=[ 1254], 60.00th=[ 1303],
     | 70.00th=[ 1352], 80.00th=[ 1418], 90.00th=[ 1549], 95.00th=[ 1696],
     | 99.00th=[ 2008], 99.50th=[ 2311], 99.90th=[ 4752], 99.95th=[ 5407],
     | 99.99th=[11731]
   bw (  KiB/s): min= 1984, max= 5248, per=25.01%, avg=2988.77, stdev=330.12, samples=119
   iops        : min=  496, max= 1312, avg=747.19, stdev=82.53, samples=119
  lat (usec)   : 2=0.01%, 4=2.54%, 10=3.09%, 20=0.21%, 50=37.30%
  lat (usec)   : 100=6.68%, 250=0.56%, 500=0.36%, 750=0.13%, 1000=1.06%
  lat (msec)   : 2=47.53%, 4=0.45%, 10=0.08%, 20=0.02%, 50=0.01%
  cpu          : usr=0.43%, sys=6.49%, ctx=57603, majf=0, minf=0
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=44897,44830,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32
rand: (groupid=0, jobs=1): err= 0: pid=995: Wed Oct 22 10:10:55 2025
  read: IOPS=736, BW=2945KiB/s (3016kB/s)(173MiB/60005msec)
    clat (nsec): min=1579, max=614980, avg=39020.91, stdev=18442.70
     lat (nsec): min=1647, max=615060, avg=39206.49, stdev=18475.75
    clat percentiles (usec):
     |  1.00th=[    3],  5.00th=[    5], 10.00th=[    8], 20.00th=[   34],
     | 30.00th=[   37], 40.00th=[   38], 50.00th=[   40], 60.00th=[   41],
     | 70.00th=[   44], 80.00th=[   47], 90.00th=[   55], 95.00th=[   64],
     | 99.00th=[   96], 99.50th=[  114], 99.90th=[  176], 99.95th=[  239],
     | 99.99th=[  392]
   bw (  KiB/s): min=   56, max= 3984, per=24.65%, avg=2945.61, stdev=427.71, samples=119
   iops        : min=   14, max=  996, avg=736.40, stdev=106.93, samples=119
  write: IOPS=736, BW=2945KiB/s (3016kB/s)(173MiB/60005msec); 0 zone resets
    clat (usec): min=339, max=28103, avg=1304.50, stdev=299.04
     lat (usec): min=340, max=28103, avg=1304.71, stdev=299.07
    clat percentiles (usec):
     |  1.00th=[  922],  5.00th=[ 1057], 10.00th=[ 1090], 20.00th=[ 1156],
     | 30.00th=[ 1188], 40.00th=[ 1221], 50.00th=[ 1254], 60.00th=[ 1303],
     | 70.00th=[ 1352], 80.00th=[ 1418], 90.00th=[ 1549], 95.00th=[ 1696],
     | 99.00th=[ 2008], 99.50th=[ 2245], 99.90th=[ 4555], 99.95th=[ 5080],
     | 99.99th=[ 6652]
   bw (  KiB/s): min=   64, max= 3336, per=24.65%, avg=2945.14, stdev=368.87, samples=119
   iops        : min=   16, max=  834, avg=736.29, stdev=92.22, samples=119
  lat (usec)   : 2=0.01%, 4=2.44%, 10=2.97%, 20=0.24%, 50=37.35%
  lat (usec)   : 100=6.60%, 250=0.37%, 500=0.06%, 750=0.10%, 1000=1.06%
  lat (msec)   : 2=48.29%, 4=0.42%, 10=0.09%, 20=0.01%, 50=0.01%
  cpu          : usr=0.34%, sys=6.51%, ctx=57300, majf=0, minf=0
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=44180,44181,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=32

Run status group 0 (all jobs):
   READ: bw=11.7MiB/s (12.2MB/s), 2945KiB/s-3015KiB/s (3016kB/s-3087kB/s), io=700MiB (734MB), run=60001-60005msec
  WRITE: bw=11.7MiB/s (12.2MB/s), 2945KiB/s-3013KiB/s (3016kB/s-3086kB/s), io=700MiB (734MB), run=60001-60005msec
 
could you try without sync? but still, those numbers don't look super good..
 
could you try without sync? but still, those numbers don't look super good..
without sync:
https://pastebin.com/mqiXpF8h

I will try netapp support... When is native proxmox support coming for NVMe/TCP? Now I connected our netapp via: "apt -y install nvme-cli"

Any clues why the performance is bad?

Btw: During a clone, the cloning stops at 100 percent of each disk and waits for like 15-30minutes (depending on how large the disk is I think) and then continues
 
Last edited:
Btw: During a clone, the cloning stops at 100 percent of each disk and waits for like 15-30minutes (depending on how large the disk is I think) and then continues

that sounds like flushing the pending writes at the end..
 
  • Like
Reactions: waltar and UdoB
flushing just means that those writes are not done sync (which is faster) so there is a backlog of them that still needs to be written out. for bulk data like this this is the best way to do it normally, I don't think the async-ness is the cause of your slowness here, rather the huge backlog is a symptom of it.
 
  • Like
Reactions: waltar