Hi all,
I used Proxmox 1.6 for a while and I am now on the move for the 2.0 waiting (with excitation for the HA feature... congratulation to the dev team for that).
To migrate my VM from my old 1.6 server to the new 2.0 I did a backup of the VMs on the old server, transfer my .tgz on the new backup storage of the new server then I restore them on the new server.
The restoring time was aroud (in my case) 3/4 hour quite for all my VM (+/- 1/4 hour between them). One of my VM was 1To disk size and others were around 100Go but at that time they are all occuying less than 20Go in "real data".
Today I restored one of them (the smallest (100Go) on the server 2.0 and it took me 2hours and a half for the restore process. It ends with a message telling me it had written all the declared size (100Go).
This make me becoming very afraid to restore the 1To size machine as it seems to be very much more machine size dependant as before ??
I just wanted to know if there was some modification in the backup / restore process that could explain this diffrence (since my old version I wouldn't be surprised but there is could one of them explain this difference ?) or do you had some clue to investigate what could be the reason.
PS: the only installation difference for me is that in the new case I use a lvm on top on DRBD storage (for the upcoming HA option) rather than lvm on top on physical device before. Can this be an explanation and if it is should this additionnal time being a constant or related to the content ??? (so 1To will take a full day restoring ??? )
Many thanks for information
Snarf
I used Proxmox 1.6 for a while and I am now on the move for the 2.0 waiting (with excitation for the HA feature... congratulation to the dev team for that).
To migrate my VM from my old 1.6 server to the new 2.0 I did a backup of the VMs on the old server, transfer my .tgz on the new backup storage of the new server then I restore them on the new server.
The restoring time was aroud (in my case) 3/4 hour quite for all my VM (+/- 1/4 hour between them). One of my VM was 1To disk size and others were around 100Go but at that time they are all occuying less than 20Go in "real data".
Today I restored one of them (the smallest (100Go) on the server 2.0 and it took me 2hours and a half for the restore process. It ends with a message telling me it had written all the declared size (100Go).
This make me becoming very afraid to restore the 1To size machine as it seems to be very much more machine size dependant as before ??
I just wanted to know if there was some modification in the backup / restore process that could explain this diffrence (since my old version I wouldn't be surprised but there is could one of them explain this difference ?) or do you had some clue to investigate what could be the reason.
PS: the only installation difference for me is that in the new case I use a lvm on top on DRBD storage (for the upcoming HA option) rather than lvm on top on physical device before. Can this be an explanation and if it is should this additionnal time being a constant or related to the content ??? (so 1To will take a full day restoring ??? )
Many thanks for information
Snarf
