vcbmounter to create simple backup

[mike-p]

i have 2 esx3.5 server without shared storage and only 4-6 vm's.

now i plan to create a simple cron controlled vcbmounter script to backup the vm's vice versa online to the local storages of the 2 hosts.

i found out that if i want to use these backups directly i have to edit some VM-names, the vmx and the 1st vmdk file.

but this is only possible if i use sparse files. if i use the "full" option the 1st vmdk cannot be edited ( because it's too large ).

what shall happen if i use a 100gb original vm and the backup consits of 50 2gb files ? i think this will end in very bad performance.

[kjb007]

The -flat.vmdk files are the actual binary data that is stored in the vm disks. This should never need to be edited. If any edits need to take place, they would be to the .vmdk file that is used to describe the -flat.vmdk file. If you are on ESX host, you can scp the vm files directly, or use vmware converter inside of the vm's, and convert the vm to the opposite esx server. That would be much safer than trying to edit vmdk files directly for any reason.

-KjB

[mike-p]

if i use converter for restore don't i need the same amount of disk space again ?

[kjb007]

You should be able to select "thin" format for stand-alone vm files.

-KjB

[mike-p]

do you mean 2gb sparse files ?

the target is to get a backup which can be used directly to create a new vm or to add to inventory ( not to use it as source for a restore option - this uses too much disk space ).

if i use vcbmounter with the scp-option i tested 2gb sparse and full vm. which option do you suggest ?

[Texiwill]

Hello,

2gbsparse is not the same as thin.

2gbsparse splits the VMDK into 2GB chunks. However the files will only contain as much data as there is in the 2GB segment of the disk, so in essence you could be storing less data than the entire VMDK. This is the format best used for moving files around different file systems. I.e. NTFS, ext3, etc.

Thin is a disk that has only those inodes that are in use allocated and the end inode. This is also considered to be a 'sparse' file in UNIX terms. This file can only live on a VMFS generally or where it was created. Once moved the sparse parts are not necessarily copied (scp will copy it properly btw). It depends on the copy tool you use.

2gbsparse can be used within ESX as is or imported using converter.

Best regards,

Edward L. Haletky

VMware Communities User Moderator

====

Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education.

CIO Virtualization Blog: http://www.cio.com/blog/index/topic/168354

As well as the Virtualization Wiki at http://www.astroarch.com/wiki/index.php/Virtualization

[mike-p]

ok,

which type should i use in the vcbmounter command :

vcbMounter -h localhost -u root -p nopasswd -a ipaddr:192.168.1.10 -r scp://root@192.168.1.2:/vmfs/volumes/esx02\:storage1/vm01

in order to use the created output directly to start the vm in case of a failure of the 1st esx according this modifications :

Step 1: Review Catalog File Inventory list of disknames to avoid duplicate <name>.vmdk filenames Step 2: Edit .vmx File Remove -NNNNNN inserted into <name>.vmdk filename instances Ensure all <name>.vmdk filename instances are unique Step 3: Rename .vmdk Files to Match .vmx Listings Remove scsiN-N-N- pre-extensions from <name>.vmdk files Ensure <name>.vmdk files match those in .vmx Step 4: Edit .vmdk Index File (Not -sNNN Disk Files) Correct names for <name>.vmdk and <name>-sNNN.vmdk files accordingly

if i use the fullvm paramter i cannot find the .vmdk index file.