Help Windows Server 2012 R2 error IO disk

[hibrid0]

Hi guys, I'm using Proxmox on a Dedicated server from OVH.
I have 3 Virtual Machines.
1. Pfsense.
1. Windows Server 2012 R2 (Domain Controller).
1. Windows Server 2012 R2 ( Web Application Server).

On Application Server I have installed with Virtio hard drive and virtio network card.
2 disk qcow2.
1. 200GB.
2. 1.2TB.

On info I have around 600GB.
But start to slow when I try to upload info. I check event viewer and see this errror:

The IO operation at logical block address 0x113467531 for Disk 2 (PDO name: \Device\00000002b) was retried."

And I don't know what I need to do.
I check Windows disk with chkdsk /r.
On proxmox I see smart state and say is so good.

Help.

 

[manu]

Have a look in pve host in /var/log/kernel.log if you see I/O errors related messages
they look like;
end_request: I/O error, dev sdb, sector 1261069669

 

[hibrid0]

I have Proxmox Virtual Environment 4.3-1/e7cdc165

I have this file on the log:
.

And this is the images folder for this virtual machine and disk DF on Host

I have Soft-Raid.

I'm not sure the format size, because If I see the image size is more bigger than the disk size.




The file size set on Proxmox is 1.2TB.
The real file size is 2TB.
And I have 2x2TB. on Soft-Raid.
How can I make more smaller the qcow2 file. Maybe this is the error.

 

[hibrid0]

I see similar problem here:
http://serverfault.com/questions/534864/qcow2-disk-size-much-bigger-than-virtual-size
Any Ideas?

 

[manu]

Disk errors:
A) soft raid is not recommended by PVE. Many people have experience file corruption after using this. I advise to make a backup of your VM as soon as possible
disk image size:
can you post the port the output of qemu-img /path/to/mydiskimage ?

 

[hibrid0]

Hi bro.

This is part of the output.

24 errors were found on the image.
Data may be corrupted, or further writes to the image may corrupt it.

38 leaked clusters were found on the image.
This means waste of disk space, but no harm to data.
19660800/19660800 = 100.00% allocated, 38.01% fragmented, 0.00% compressed clusters
Image end offset: 2194719899648


:/var/lib/vz/images/104# qemu-img check -r all vm-104-disk-2.qcow2 >qemu-error.txt
ERROR cluster 2273346 refcount=0 reference=1
ERROR cluster 2273352 refcount=0 reference=1
ERROR cluster 2273353 refcount=0 reference=1
ERROR cluster 2273354 refcount=0 reference=1
ERROR cluster 2273355 refcount=0 reference=1
ERROR cluster 2273356 refcount=0 reference=1
ERROR cluster 2273359 refcount=0 reference=1
ERROR cluster 2273360 refcount=0 reference=1
ERROR cluster 2273361 refcount=0 reference=1
ERROR cluster 2273364 refcount=0 reference=1
ERROR cluster 2273365 refcount=0 reference=1
ERROR cluster 2273366 refcount=0 reference=1
ERROR cluster 2273372 refcount=0 reference=1
ERROR cluster 2273374 refcount=0 reference=1
ERROR cluster 2273378 refcount=0 reference=1
ERROR cluster 2273381 refcount=0 reference=1
ERROR cluster 2273388 refcount=0 reference=1
ERROR cluster 2273389 refcount=0 reference=1
ERROR cluster 2273390 refcount=0 reference=1
ERROR cluster 2273392 refcount=0 reference=1
ERROR cluster 2563950 refcount=0 reference=1
ERROR cluster 2563951 refcount=0 reference=1
Leaked cluster 21654367 refcount=1 reference=0
Leaked cluster 21654368 refcount=1 reference=0
Leaked cluster 21654388 refcount=1 reference=0
Leaked cluster 21654389 refcount=1 reference=0
Leaked cluster 21654546 refcount=1 reference=0
Leaked cluster 21654547 refcount=1 reference=0
Leaked cluster 21654663 refcount=1 reference=0
Leaked cluster 21654664 refcount=1 reference=0
Leaked cluster 21654704 refcount=1 reference=0
Leaked cluster 21654705 refcount=1 reference=0
Leaked cluster 23932158 refcount=1 reference=0
Leaked cluster 23932159 refcount=1 reference=0
Leaked cluster 23932160 refcount=1 reference=0
ERROR cluster 29630626 refcount=0 reference=1
ERROR cluster 29630628 refcount=0 reference=1
Leaked cluster 32159593 refcount=1 reference=0
Leaked cluster 32159594 refcount=1 reference=0
Leaked cluster 32505856 refcount=1 reference=0
Leaked cluster 32538624 refcount=1 reference=0
Leaked cluster 32538625 refcount=1 reference=0
Leaked cluster 32538626 refcount=1 reference=0
Leaked cluster 32579245 refcount=1 reference=0
Leaked cluster 32579247 refcount=1 reference=0
Leaked cluster 32579248 refcount=1 reference=0
Leaked cluster 32800768 refcount=1 reference=0
Leaked cluster 32800769 refcount=1 reference=0
Leaked cluster 32800770 refcount=1 reference=0
Leaked cluster 32833537 refcount=1 reference=0
Leaked cluster 32833538 refcount=1 reference=0
Leaked cluster 32931840 refcount=1 reference=0
Leaked cluster 32997376 refcount=1 reference=0
Leaked cluster 32997377 refcount=1 reference=0
Leaked cluster 32997378 refcount=1 reference=0
Leaked cluster 33095680 refcount=1 reference=0
Leaked cluster 33095681 refcount=1 reference=0
Leaked cluster 33292288 refcount=1 reference=0
Leaked cluster 33292289 refcount=1 reference=0
Leaked cluster 33488765 refcount=1 reference=0
Leaked cluster 33488766 refcount=1 reference=0
Leaked cluster 33488767 refcount=1 reference=0
Rebuilding refcount structure
ERROR writing refblock: File too large
qemu-img: Check failed: File too large
root@:/var/lib/vz/images/104# qemu-img --version
qemu-img version 2.6.1pve-qemu-kvm_2.6.1-6, Copyright (c) 2004-2008 Fabrice Bellard
root@:/var/lib/vz/images/104# qemu-img version 2.6.1pve-qemu-kvm_2.6.1-6, Copyright (c) 2004-2008 Fabrice Bellard
-bash: error sintáctico cerca del elemento inesperado `('
root@:/var/lib/vz/images/104#

 

[manu]

make a backup to an external storage as soon as you can

 

[hibrid0]

was made a Backup yesterday, I'm migrating to a better server with more space, 2x4TB.
I have installed today the new Proxmox Server and restore 2 old-machines, but this machine with problems is really big.
800GB on proxmox Backup, but I have just 250GB on second disk a report I have 2TB.
How can I fix this image?

 

[hibrid0]

Ok I try to fix the image with this command:
qemu-img -f all MyQcow2Image.qcow2.

And show this nice message:

ERROR writing refblock: File too large.
qemu-img: Check failed: File too large.

 

[manu]

As long as you did not do a backup of this VM (yes it's big, maybe you did snaphots ? ) I do not advocate to try to fix the disk image.
What you're trying right now is risky at best.
Does the data in this VM matters to you ? Then ask your provider for a backup space where you can put a backup of the disk image.

 

[hibrid0]

I have 5TB of Backup Space.
Yesterday I made the full proxmox server backup and transfer to my Storage Backup.
Today I restore 2 Machines, and transfering the Problematic machine to the new server 800GB (37% of progress).
I dont have any snapshot, this machine is not in production, in theory this machine is the backup for the real machine.

The better way is create a new RAW disk and transfer all información from qcow2 to new disk? or What's the better way?.
Or just convert on RAW with qemu-img?