[SOLVED] qmrestore / write zeros

I

infogram

New Member
Jan 11, 2016
5
0
1
25
Hi,

First of all, apologies if this question has been answered already but I haven't found the answer in the archives.

Also I'm rather a noob :-/

Here is my situation :

I'm running PVE on Debian 8 on 2 hosts :
I have HOST1 with 3 VMs (.raw) running, and HOST2 where I backup those 3 VMs every night (via NFS), then I restore them (but keep them stopped).
So if HOST1 fails, I'll just have to start the VMs in HOST2.

So I backup the 3 VMs using a PVE task (snapshot). Everything is working has expected.

Then, on the backup host, I'm using a cron script which :
- erases the old stopped VMs if found (qm destroy)
- restores some new VMs (qmrestore, no fancy options)​

This has beeen working like a charm for months.

Then, 2 days ago I made a Debian update, starting with HOST2
I went from :
- pve-manager 4.1-4/ccba54b0 to 4.1-13/cfb599fb
- kernel 4.2.6-1-pve to 4.2.8-1-pve​

And now my qmrestore is waaaaay more slow, so I checked the logs and I found :
map 'drive-virtio0' to '/var/lib/vz/images/101/vm-101-disk-1.raw' (write zeros = 1) ​
whereas before it was (write zeros = 0)
Then I checked the restored VMs sizes, and yes, they now have the size I gave to their HD, whereas before the restored VMs kept the dynamic size I gave to the original VMs.
So my guess is some default settings has changed in qmrestore but I can't find where (and I don't see why).

Could anybody give me some help please ?
 
Hi Dominik,
Thanks very much for your quick response.
I've just updated to pve 4.3-10.
I will update the thread after tonight's qmrestore.
 
Hi,
So everything went back to normal after the PVE upgrade.
Thank you Dominik for your help !
 
I have a raw disk image with preallocation=full and when i do a restore the disk of the VM return to defaults
 
You must log in or register to reply here.
Top Bottom