Re: Some Data Recovery questions

Eggert382 · ‎08-19-2010

Hey All,

i'm totally to this communitie and i'm getting started with VMware Data Recovery.

I use ESX 4.1 and VDR 1.2 - everythings seems to work fine, backup jobs run and i can restore the VMs - but i'm having 3 questions:

1: when i mark restorepoints as "ready/prepare to delete" in restoring, when will these marked restorepoints be deleted?

2: where can i choose between a full or an incremental/differetal backup?

3: in vdr_10_admin.pdf (attached) on page 8 there stands:

NOTE These optimizations do not apply to virtual machines created with VMware products prior to vSphere
4.0. For example, change tokens are not used with virtual machines created with Virtual Infrastructure 3.5 or
earlier. As a result, virtual machines created with earlier VMware versions take longer to back up.

Nearly all our VMs were created on ESX 3.5 - but they are all upgraded to VM version 7 - what does that mean to/for me, do i have to live with any restrictions or is all fine because of VM version 7.

I hope my english is not that bad and you unterstand what i was trying to say.

Kind Regards,

Eggert382

Paul11 · ‎08-19-2010

1: when i mark restorepoints as "ready/prepare to delete" in restoring, when will these marked restorepoints be deleted?

You Can start an Integrity Check manualy, than the marked restore points get deleted. Otherwise you had to wait until the Reclaiming Task is running.

2: where can i choose between a full or an incremental/differetal backup?

The first backup will be a full backup, all other backups are than incremental for the life of your datastore. When you create a new datastore, you will get a new Full Backup for the first backup, after than alle backups are incrementals.

3: in vdr_10_admin.pdf (attached) on page 8 there stands:

...

If the VM is convertet to HW 7, you use the "Change Tracking log", which will make your backups fast. It dosn't matter, what the VM was bevor. So your convertet VM's will be fine.

Paul

Eggert382 · ‎08-20-2010

THX for your answers Paul.

You write:

If the VM is convertet to HW 7, you use the "Change Tracking log", which
will make your backups fast. It dosn't matter, what the VM was bevor.
So your convertet VM's will be fine.

Where do i exactly find "Change Tracking log"?!

Eggert

Paul11 · ‎08-20-2010

In the home-directory of your VM you will see a file named 'name of the VM'-ctk.vmdk. This is the file, where the changes of the disk-blocks are saved.

In Virtual Center you can look at Edit Settings - Options - General - Configuration Parameters: There ist a parameter named ctkEnable, which is true, if you have ever backuped your VM with VDR. (You can only look at this, when the VM is powered off).

But you will also see this parameter in the .vmx file of the VM. Hope this is the answer for you question.

Paul

Eggert382 · ‎08-23-2010

THX, for your answer - two more questions:

#1: the *-ctk.vmdk file will only be created if the VM is backed up?

#2: If i add a hdd to my VM which is backed up, the hdd will be automatically backed up. But in the log it's just saying "incremental back up" - is this correct and if this is correct, why is there no full back up required?

regards

Eggert

Paul11 · ‎08-23-2010

#1: the *-ctk.vmdk file will only be created if the VM is backed up?

The file will be created at the first power on time for every disk of the VM. You can delete this file, when the VM is powered off. Than you should see a Full backup again.

#2: If i add a hdd to my VM which is backed up, the hdd will be automatically backed up. But in the log it's just saying "incremental back up" - is this correct and if this is correct, why is there no full back up required?

I don't know this, may be azmir can answer this question.

Eggert382 · ‎08-24-2010

Ok, i think i figered it out with the "*-ctk.vmkd"-files.

More questions:

#1:

I started backing up VMs on our producticv environment with some 'tests'. It works, but it is always saying "Incremental-Backup" - but when i look i the time it takes, it should be a Full-Backup - because the first run takes 17min and the second run only 2min. Am I missing something or doing something wrong?

Here the protocol from the konfiguration tab:

23.08.2010 15:05:19: Normales Backup verwendet Initial
23.08.2010 15:05:25: Kopiervorgang wird ausgeführt VM016061
23.08.2010 15:05:40: Inkrementell-Sicherung der Festplatte "[http://vm02-ul.local-store] VM016061/VM016061-flat.vmdk" wird unter Verwendung von "Netzwerk" durchgeführt
23.08.2010 15:22:31: Aufgabe wurde erfolgreich abgeschlossen
23.08.2010 15:22:31: Abgeschlossen: 5 Dateien, KB
23.08.2010 15:22:31: Durchsatz: 7290 MB/Minute
23.08.2010 15:22:31: Dauer: 00:17:12 (00:00:20 Leerlauf/Warten/Vorbereiten)
23.08.2010 19:00:01: Normales Backup verwendet Täglich
23.08.2010 19:00:09: Kopiervorgang wird ausgeführt VM016061
23.08.2010 19:00:18: Inkrementell-Sicherung der Festplatte "[http://vm02-ul.local-store] VM016061/VM016061-flat.vmdk" wird unter Verwendung von "Netzwerk" durchgeführt
23.08.2010 19:01:55: Aufgabe wurde erfolgreich abgeschlossen
23.08.2010 19:01:55: Abgeschlossen: 5 Dateien, KB
23.08.2010 19:01:55: Durchsatz: 75214 MB/Minute
23.08.2010 19:01:55: Dauer: 00:01:55 (00:00:16 Leerlauf/Warten/Vorbereiten)
24.08.2010 08:00:31: Normales Backup verwendet Initial
24.08.2010 08:00:41: Kopiervorgang wird ausgeführt VM016059
24.08.2010 08:00:51: Inkrementell-Sicherung der Festplatte "[http://vm02-ul.local-store] VM016059/VM016059-flat.vmdk" wird unter Verwendung von "Netzwerk" durchgeführt
24.08.2010 08:23:20: Aufgabe wurde erfolgreich abgeschlossen
24.08.2010 08:23:20: Abgeschlossen: 5 Dateien, KB
24.08.2010 08:23:20: Durchsatz: 11386 MB/Minute
24.08.2010 08:23:20: Dauer: 00:22:49 (00:00:19 Leerlauf/Warten/Vorbereiten)
24.08.2010 09:00:00: Normales Backup verwendet Initial
24.08.2010 09:00:19: Kopiervorgang wird ausgeführt svc04-ul
24.08.2010 09:00:50: Inkrementell-Sicherung der Festplatte "[store03-ul LUN1] svc04-ul/svc04-ul-flat.vmdk" wird unter Verwendung von "Netzwerk" durchgeführt
24.08.2010 09:13:50: Aufgabe wurde erfolgreich abgeschlossen
24.08.2010 09:13:50: Abgeschlossen: 5 Dateien, KB
24.08.2010 09:13:50: Durchsatz: 7753 MB/Minute
24.08.2010 09:13:50: Dauer: 00:13:51 (00:00:37 Leerlauf/Warten/Vorbereiten)

#2: So the only way to do a new Full-Backup of VMs is to delete the "*-ctk.vmdk"-file? Is there no other way to do a Full-Backup - for examaple every sunday --> Full-Backup?

#3: The target-store where the VMs are backed up to: If i have 5 VMs with one disk having 10GB, will the non deduplicated space take ~50GB (the first full back ups) the deduplicated space will grow with the incremental back ups?

Sorry for these noobie questions....

Eggert

Edit: something terribly went wrong when i was trying to edit the post...

Message was edited by: Eggert382

Paul11 · ‎08-24-2010

#2: So the only way to do a new Full-Backup of VMs is to delete the "*-ctk.vmdk"-file? Is there no other way to do a Full-Backup - for examaple every sunday --> Full-Backup?

Sorry for my last post, the information was wrong. Deleting the -ctk.vmdk file will only trigger a full scan of your devices instead of reading the information of changed blocks from the -ctk-file. So you will not get a new Full backup. Every Backup you made is like a full backup, but only the first backup needs to write every blocks of your disk to the deduplication-store. The following backups only transfers the changes since the last backup, but for restoring a disk it's like a full backup every day.

#3: The target-store where the VMs are backed up to: If i have 5 VMs with one disk having 10GB, will the non deduplicated space take ~50GB (the first full back ups) the deduplicated space will grow with the incremental back ups?

Also the blocks of the other VM's get deduplicated, so your backup store will not need 50 GB for the fist Full Backup when you have similar VM's which were backuped to the same datastore. After that, your datastore will grow a little bit every day due to the incremental backups, until the Reclaim Job is running and deletes your old backups. So after a while, your datastore will not grow very much, because old blocks were deleted and the new blocks can use that space again. (In theory, because you can read about other experiencies in this forum).

Eggert382 · ‎08-24-2010

THX once again Paul.

You can't help my with #1?

#2: So the only way to do a new Full-Backup of VMs is to delete the
"*-ctk.vmdk"-file? Is there no other way to do a Full-Backup - for
examaple every sunday --> Full-Backup?
Sorry for my last post, the information was wrong. Deleting the
-ctk.vmdk file will only trigger a full scan of your devices instead of
reading the information of changed blocks from the -ctk-file. So you
will not get a new Full backup. Every Backup you made is like a full
backup, but only the first backup needs to write every blocks of your
disk to the deduplication-store. The following backups only transfers
the changes since the last backup, but for restoring a disk it's like a
full backup every day.

But when i now mark the first backup as ready todelet and the reclaim job deletes it, how can i then restore my VM? Should not then be "a new Full Backup" generated because it doesn't have the original restorepoint?

Paul11 · ‎08-24-2010

#1 But when i now mark the first backup as ready todelet and the reclaim job deletes it, how can i then restore my VM? Should not then be "a new Full Backup" generated because it doesn't have the original restorepoint?

I don't think that there is a problem deleting the first backup (which is the Full Backup), because also the blocks of this full backup are maybe used by many other backups, and so VDR can only delete these blocks, which are unique to your backup you want to delete. Don't forget, the deduplication is on block level over all your backups (and VM's) and deduplicated blocks are only pointers to the disk-blocks, that's why VDR can deduplicate so much.

Eggert382 · ‎08-24-2010

THX - i think i will do some long term tests an check some things - if i'll have more questions, i know where/how to find you...

Regards

Eggert