Changing backup algorithm

[GoogleCodeExporter]

Reason:
Too long suspend time(especially with "fat" VE).

Description:
Now:
1) suspend VE(timestamp1)
2) tar /var/lib/vz/private/VEID
3) continue VE(timestamp2)

Want(vzdump algorithm):
1) rsync /var/lib/vz/private/VEID to tmp location
2) suspend VE(timestamp3)
3) rsync /var/lib/vz/private/VEID one more time
4) continue VE(timestamp4)

timestamp4-timestamp3 < timestamp2-timestamp1

man vzdump:
***
         For OpenVZ, this mode uses rsync to copy the VM to a temporary location (see option --tmpdir). Then the VM is suspended and a second rsync copies changed files. After that, the VM is started (resume) again. This results in a minimal downtime, but needs additional space to hold the VM copy.
***

PS
using 2.0 release at debian squeeze

Original issue reported on code.google.com by natan.f...@gmail.com on 12 Dec 2011 at 6:15

[GoogleCodeExporter]

Ooops, lost little part:

after 
3) rsync /var/lib/vz/private/VEID one more time
4) continue VE(timestamp4)

we should add 
5) tar VE copy

Original comment by natan.f...@gmail.com on 12 Dec 2011 at 7:13

[GoogleCodeExporter]

Original comment by sibprogrammer on 13 Mar 2012 at 12:57

• Added labels: Milestone-Release2.5