[PVE-User] WG: difference b/w backup types

Alain Péan alain.pean at lpp.polytechnique.fr
Wed Sep 28 18:22:37 CEST 2011


Hi Jeremiah,

That is what I do myself with BackupPC, for some VMs (so backup files 
inside the VM).

The use of rsync to backup the vzdump(ed) images has the advantage that 
you backup a file that does not change during the backup, not a live VM.

Thanks for all these informations.

Alain

Le 28/09/2011 17:55, obj at jltechinc.com a écrit :
> Hello Alan,
>
> We use BackupPC to backup the files within the VMs, since these are 
> what the users will accidentally delete and want restored.
>
> We use a separate rsync backup to grab the vzdump(ed) images, but only 
> after a significant O/S update, else we have everything else we need 
> from the incremental backups.
>
> -J
>
> Alain Péan said the following on 09/28/2011 11:21 AM:
>> Hi Jeremiah,
>>
>> Just a question, just to know what is the best policy. Do you use 
>> BackupPC to backup the image files inside the host, or the files 
>> inside the VM ? The image files are big files, that change at least 
>> slightly (logs, updates...), so backupPC needs to backup the whole 
>> (big) file each time.
>>
>> If you backup the files inside the VM, you won't backup the files 
>> that don't change, it is less ressource intensive for your network, 
>> but it is less easy to restore than an image file...
>>
>> Alain
>>
>> Le 28/09/2011 17:12, obj at jltechinc.com a écrit :
>>> I concur, we use the high-performance, enterprise-grade, openSource 
>>> *BackupPC* with data-deduplication.
>>>
>>> http://backuppc.sourceforge.net/
>>>
>>> -Jeremiah
>>>
>>> Stefan Sänger said the following on 09/28/2011 11:05 AM:
>>>> Hi Muhammad,
>>>>
>>>> so I see that your case is a bit more production related and harder 
>>>> to judge. Have you considered other backup solutions?
>>>>
>>>> Because I'm not sure if proxmox VM backup is the right choice for 
>>>> your problem. Basically all it provides is a disk image, and if you 
>>>> don't take care of backing up these images all that you have left 
>>>> when you
>>>> run into problems with your VM is the disk image of the last backup 
>>>> run.
>>>>
>>>> It is definitely not the best choice for a reliable backup - if you 
>>>> run a database and just perform snapshots, you may not be able to 
>>>> restore tha latest database status.
>>>>
>>>> I prefer to look deeper into this and choose a real backup solution
>>>> to cope with this. That even gives me the possibility to do 
>>>> incremental backups with regular full backups and weekly kvm disk 
>>>> images.
>>>>
>>>> But of course all that depends on what you are running inside the VM.
>>>>
>>>>
>>>> Stefan
>>>>
>>>> On 28.09.2011 16:10, Muhammad Yousuf Khan wrote:
>>>>>
>>>>> Thank for the info.well some time backup got failed while copying 
>>>>> to the
>>>>> network folder. i have asked this question previously, people have 
>>>>> their
>>>>> own perception some say it is because you have 100mbps LAN card  so i
>>>>> changed it to 1gb and now still having the same problem. so i 
>>>>> taught if
>>>>> i can manage to minimize some process burden during backup that 
>>>>> would be
>>>>> fine i can bear 1 or 2 hours down time for a single VM at night 
>>>>> when no
>>>>> one is in office. so that was the reason i asked thus i could 
>>>>> minimize
>>>>> the failure ratio of backup. i am backing up single VM every week and
>>>>> this is a 4th time that it couldnt copy. therefore i was kinda
>>>>> troubleshooting also .
>>>>>
>>>>> Thank you,
>>>>>
>>>>>
>>>>> On Wed, Sep 28, 2011 at 5:33 PM, Stefan Sänger 
>>>>> <stsaenger at googlemail.com
>>>>> <mailto:stsaenger at googlemail.com>> wrote:
>>>>>
>>>>>     On 28.09.2011 14:16, Muhammad Yousuf Khan wrote:
>>>>>
>>>>>         still not an answer to my question. i already read this
>>>>>         information on
>>>>>         wikki. but i wanted to know which one is more faster and low
>>>>>         resource
>>>>>         intensive.
>>>>>
>>>>>
>>>>>     Well - there is no general answer to that question, because 
>>>>> you have
>>>>>     to consider this with different aspects in mind.
>>>>>
>>>>>     So, if you compare these different backup scenarios, obviously
>>>>>     - snapshot is fastest concerning downtime of your VM, as there is
>>>>>       virtually no downtime - but it will need the most resources 
>>>>> since the
>>>>>       VM itself is running as well as the backup process.
>>>>>     - suspend and stop need significantly less resources since during
>>>>>       backup only the bakup task is running
>>>>>     - stop will need more time than suspend since the VM is 
>>>>> shutdown before
>>>>>       performing the backup and afterwards has to boot up again
>>>>>     - suspend will need little more resources since the VM is not 
>>>>> shut down
>>>>>       and hence its status has to be kept in memory during backup
>>>>>
>>>>>     But - these conclusions are fairly simple. I really wonder why 
>>>>> you
>>>>>     ask these questions, are you going to start a de3tailed 
>>>>> discussion
>>>>>     on backing up VMs?
>>>>>
>>>>>
>>>>>     Stefan
>>>>>
>>>>>     _________________________________________________
>>>>>     pve-user mailing list
>>>>> pve-user at pve.proxmox.com <mailto:pve-user at pve.proxmox.com>
>>>>> http://pve.proxmox.com/cgi-__bin/mailman/listinfo/pve-user
>>>>> <http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user>
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> pve-user mailing list
>>>> pve-user at pve.proxmox.com
>>>> http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user
>>>>
>>>
>>>
>>> _______________________________________________
>>> pve-user mailing list
>>> pve-user at pve.proxmox.com
>>> http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user
>>
>>
>> -- 
>> ==========================================================
>> Alain Péan - LPP/CNRS
>> Administrateur Système/Réseau
>> Laboratoire de Physique des Plasmas - UMR 7648
>> Observatoire de Saint-Maur
>> 4, av de Neptune, Bat. A
>> 94100 Saint-Maur des Fossés
>> Tel : 01-45-11-42-39 - Fax : 01-48-89-44-33
>> ==========================================================
>>
>>
>> _______________________________________________
>> pve-user mailing list
>> pve-user at pve.proxmox.com
>> http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user


-- 
==========================================================
Alain Péan - LPP/CNRS
Administrateur Système/Réseau
Laboratoire de Physique des Plasmas - UMR 7648
Observatoire de Saint-Maur
4, av de Neptune, Bat. A
94100 Saint-Maur des Fossés
Tel : 01-45-11-42-39 - Fax : 01-48-89-44-33
==========================================================

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pve.proxmox.com/pipermail/pve-user/attachments/20110928/f565a8e3/attachment-0014.html>


More information about the pve-user mailing list