Personal Backup Appliance

[VMTN_Admin]

http://www.vmware.com/vmtn/appliances/directory/321

Backup, Restore, or Clone the Hard Disk Drive of any Computer or Virtual Machine over the Network!

[rxhui]

Thanks for the feedback and suggestions! I appreciate it.

I think adding webmin would be a great idea. In future versions (post-1.0.2), I plan to do a major overhaul and upgrade the distro to ubuntu dapper drake. I'm also thinking of stripping out the X server at that time to reduce the size of the VM and then implementing webmin for remote web administration. The support for multiple HDDs is a bit trickier, but I will add it to my growing feature request list.

That's strange. I never saw any problems with XP backups and restores. I'll have to do more testing.

Does anyone want to share your experience with backing up and restoring XP using PBA? I'm interested in getting a sense of whether or not XP backups and restores generally work. (please also mention what version of PBA you are using)

Thanks

[rxhui]

I didn't intend to do another release candidate so quickly, but I spent a bunch of time on PBA over the last couple of days and overhauled the user interface on the client end, and fixed a few corner case bugs.

Screenshots of new interface:

http://sourceforge.net/project/screenshots.php?group_id=171414

Here's a download link for the patch to v1.0.2 Release Candidate #3.

http://prdownloads.sourceforge.net/pba-vm/pba-1.0.2-rc3.zip?download

The procedure is the same as with RC2 and is documented in the readme.txt file in the zip file.

Enjoy.

[kwizi]

I'm a moron when it comes to Linux. Help me out here. I want to run this to get images for servers, but I want the images to save on a different disk than the ones created in the appliance.

I have a USB hard drive attached to the laptop I'm running the VM on. The servers drives are MUCH larger than the HD on the laptop, so I want to use an external usb disk. How can I redirect the image files to a folder on this hard drive (not inside a vmware disk)?

[kaso]

Hi

It's very nice new interface.

Unfortunatelly new version still doesn't solve very strange problem with restoring suse origin physical disk to vmware new guest with scsi disk.

the restore process goes well, says "restoring original /dev/sda1 to /dev/sda1" and "sda2 to sda2"

but when booting os after retore - it fails with "waitig for sda1 to appear not found"

I'll send screenshot.

beside that other stuff works fine. I'll keep on testing different configuration of hw and sw later.

thanks a lot

[sidimustafa]

Hi there,

I tested this product, (latest version) and realised it did a pretty good job,

but after a few hours of playing around with the restored image, i noticed that i had no swap space.

here's what i did,

Created a Virtual Machine in Vmware Workstation 5.5.1.

This image was then backup with PBA, and then restored to a physical laptop.

Once the machine was started kudzu did it's magic, by detecting the hardware that was removed, and detected the new hardware,

Just what i expected..

i started playing around, and this was the conclusion..

Seems like something went wrong with the restoring of the swap partition, or is it something that i didn't do?

Seems like when u do an fdisk the swap partition is listed as normal, but when u run a "free" the swap space shows up as 0

Original Image (Virtual Machine) base on Centos 4.3

======================================

\[root@vmw01 /]# fdisk /dev/hda

The number of cylinders for this disk is set to 7179.

There is nothing wrong with that, but this is larger than 1024,

and could in certain setups cause problems with:

1) software that runs at boot time (e.g., old versions of LILO)

2) booting and partitioning software from other OSs

(e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): p

Disk /dev/hda: 59.0 GB, 59055800320 bytes

255 heads, 63 sectors/track, 7179 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System

/dev/hda1 * 1 13 104391 83 Linux

/dev/hda2 14 144 1052257+ 83 Linux

/dev/hda3 145 275 1052257+ 82 Linux swap

/dev/hda4 276 7179 55456380 5 Extended

/dev/hda5 276 1580 10482381 8e Linux LVM

/dev/hda6 1581 7179 44973936 8e Linux LVM

Command (m for help):

--

\[root@vmw01 /]# free -m

total used free shared buffers cached

Mem: 249 51 198 0 7 24

-/+ buffers/cache: 19 230

Swap: 1027 0 1027

\[root@vmw01 /]#

\----

---

Restored Image (Physical Machine) base on Centos 4.3

======================================

\[root@vmw01 /]# fdisk /dev/hda

The number of cylinders for this disk is set to 7296.

There is nothing wrong with that, but this is larger than 1024,

and could in certain setups cause problems with:

1) software that runs at boot time (e.g., old versions of LILO)

2) booting and partitioning software from other OSs

(e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): p

Disk /dev/hda: 60.0 GB, 60011642880 bytes

255 heads, 63 sectors/track, 7296 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System

/dev/hda1 * 1 13 104391 83 Linux

/dev/hda2 14 144 1052257+ 83 Linux

/dev/hda3 145 275 1052257+ 82 Linux swap

/dev/hda4 276 7179 55456380 5 Extended

/dev/hda5 276 1580 10482381 8e Linux LVM

/dev/hda6 1581 7179 44973936 8e Linux LVM

--

\[root@vmw01 /]# free -m

total used free shared buffers cached

Mem: 1003 169 833 0 25 75

-/+ buffers/cache: 68 934

Swap: 0 0 0

\[root@vmw01 /]#

Message was edited by:

sidimustafa

[rxhui]

Thanks for trying PBA. PBA doesn't backup linux swap partitions by design (because it isn't necessary). Swap partitions are created on restore. If partition type is 82, then mkswap is run to format the partition. Upon first boot up of the linux distro that you restored, it's assumed that your linux distro sets up the swap space by looking in /etc/fstab and running "swapon -a" or some equivalent command in the startup scripts.

What does your /etc/fstab say about /dev/hda3?

You can try typing 'swapon -a' or 'swapon /dev/hda3', and then type 'free -m' to see if the swap space is being used.

Does the same thing happen if you restore it to another blank VM?

Message was edited by:

rxhui

[sidimustafa]

this is the /etc/fstab

thanks for the tip, i'll run the swap command manually, and see what goes on.

\[root@vmw01 init.d]# more /etc/fstab

\# This file is edited by fstab-sync - see 'man fstab-sync' for details

LABEL=/ / ext3 defaults 1 1

LABEL=/boot /boot ext3 defaults 1 2

none /dev/pts devpts gid=5,mode=620 0 0

none /dev/shm tmpfs size=2G 0 0

/dev/sysvg/home /home ext3 defaults 1 2

none /proc proc defaults 0 0

none /sys sysfs defaults 0 0

/dev/sysvg/tmp /tmp ext3 defaults 1 2

/dev/sysvg/usr /usr ext3 defaults 1 2

/dev/sysvg/var /var ext3 defaults 1 2

/dev/vmwvg/vmachines /v_machines jfs defaults 1 2

/dev/vmwvg/images /v_images ext3 defaults 1 2

LABEL=SWAP-hda3 swap swap defaults 0 0

/dev/hdc /media/cdrecorder auto pamconsole,exec,noauto,managed 0 0

btw: thanks for the quick response

Message was edited by:

sidimustafa

[sidimustafa]

Ok, running it manually did the trick

i guess my so limited knowledge of Linux, didn't prepare me enough to know how to manually activate the swap partition,

but i guess, this is the tool i need to make images of my linux pc's,

previous tools didn't allow me to do LVM's...

I had no problem with the grub work-around, only thing was that i lost the CentOS splash screen during the the loading of grub..

thanks, for the help

[rxhui]

it's strange that Centos didn't activate the swap partition on bootup. I never had to manually activate the swap partition before. Does the swap stay enabled after a reboot or do you need to issue swapon everytime? What does your /etc/fstab say? (nevermind, I see that it's been posted already - edited by rxhui)

about the grub work-around...I think I may have a solution in the next version which will not require the workaround but it's pending further investigation.

Message was edited by:

rxhui

[rxhui]

nevermind about the /etc/fstab, I saw that it's been posted already. thanks

[rxhui]

okay, the reason, centos doesn't activate the swap file automatically is that /etc/fstab uses labels to reference the partitions instead of the actual partition. Since the swap partition was created fresh on the restore, it has no label, and thus does not activate on bootup.

here's a link on how you can add the appropriate label to your partition, so that references in /etc/fstab will be meaningful.

http://www.redhat.com/docs/manuals/linux/RHL-7.3-Manual/ref-guide/s1-filesystem-ext3-create.html

I suppose I can add a feature to detect this scenario and run e2label accordingly. I'll look into that.

[The_Fiddler]

Hi there,

Great app, wich will possibly be used to make image-backups of the windows_XP clients in my network. Did som testing today with verson 1.01, and everything went well.

The reasen of wanting to use it is that we give 24/24, 7/7 support, and everyone of us has to be able to give support. The simplicity of this program

is making that very usefull for this task if a restore is nessecery.

Backup and restore of a drive having one partition went well. After restore the sytem worked fine (WinXP Pro SP2 on ntfs).

For the second test i resized the the drive and made two partitions, to find out what PBA does. Partition magic complained about some "0 length files", but the XP drive test said nothing.

Restoring a drive with two ntfs partitions go's well allso.

A question here. How big is the (growing) backup partition by default in the vmware-machine ?

Will it in the future be possible to only backup the first partition of a drive (its on my wichlist)?

Thanks for the graet work.

[sidimustafa]

Ok, just for your information - just incase u want to incorporate the Swap fix for Centos / RHEL in the upcoming version.

swapon /dev/hda3 did not survive the reboot, upon reboot, i had no swap space.

e2label, only works on ext2/ext3 partions, and not on swap space.

so with the help of google, i came across the trick

mkswap to the rescue.

mkswap, is the command that is use to assign the swap parttion..

it also has the switch to assign a Label

command was as follows:

"mkswap -L SWAP-hda3 /dev/hda3"

whereas SWAP-hda3 was the label generated during the original installation.

after the reboot, swapspace was automatically allocated

\[sidi]

Message was edited by:

sidimustafa

[rxhui]

Cool! Thanks for clarifying that. I overlooked the fact that e2label is for ext2/ext3. At a minimum, I will add mkswap -L to cover the Centos case to do a cheap fix. A better solution would be to eventually scan through /etc/fstab to determine what the label should be. This is a great find that will make PBA much better in the future.

[rxhui]

Thanks for the feedback.

By default it's 20GB, but in the hdd_upgrade folder, there's a readme that describes how you can upgrade the capacity to to 250GB.

Backing up only one partition or selected partitions is on the list of things I will consider in the future. For now, the emphasis is on simplicity (as few steps as possible) and at some point, there will be an advanced menus with options to toggle.

[sidimustafa]

Ok, not in a rush, but would be nice for you to add on your todo list.

Progress Bar.

\----

---

as i read in the hand book, that u use different part. tools for different file systems.

when i did my clone, it was done on ext3/lvm/jfs

it seems to me that partitions that was ext3 (could be i'm wrong) had a nice display progress bar, where i could see that it was in progress.

then, it looked like when the LVM's cat to bat, a next tool came to the mount, (i think it was the split/unsplit tool) well he came with no style what soever,

just telling me that he was doing something and no progress bar.

well, if it was a small Logical Volume then it was cool, but a large 40Gb+ Log. Vol, took some time, good thing it was one of those laptops that had a nice HD LED, so i could have seen it chunning away, letting me know that something was happing.

Would be nice to see the progress display in one of the to come versions.

Crazy Me.

Well, i know the book said, DON"T USE HARDRIVES SMALLER THAN THE CLONE IMAGE..

i had to try it out to see what would happen.

this is what went down:

the last partition (LVM) was blank, so wasn't no problem if it went wrong.

i cloned a 60Gb disk, and restore it back on a 40Gb disk.

well, all went good, until it reach the end of the 40Gb disk.

I was just told that there was an error message, and it coldn't write to the file.

and then it stood there till thy kindom come, ain't move a next inch, nor a nice a friendly message stating, HEY MORRAN, i told u no clonning from large disk to small disk, go and fix it so come back.

--Best of all, even thou it hung, when i rebooted, i had a fully cloned image, working like nobody business, only thing was the last partition /dev/hda6 (LVM / jfs) was giving lots of error messages during the boot process, and better yet, it was still readable

Well, this was just a test, but someone might just miss that statement, when reading the manual (especially if already custom of symantec ghost) ghost automaticall adjust the size of the disk (it's up to u to make sure, that u don't have no data on the partition that will get shrink)

so, would be nice to give a friendly message, at the end of the process, when it goes wrong, stating that I will no exit due to error message.

Or better yet, at the begin of the process, it does a checksum, and make sure the size of the clone image is the same size as the disk to be clone, and can exit the process at that point.

Ok, that was just my little 0.02 cents.

I love this product, fast/simple, and best yet, IT DOES LVM's.....

i'll keep testing, and reporting as i go by.

Gonna try some more crazy stuff, going to check that NAS thing, with syncbak or the M$ synctool, and see how that work.

once again, keep the good work up.

Message was edited by:

sidimustafa

[rxhui]

Another release candidate for 1.0.2, release candidate #4. 1.0.2 will be released on 9/1.

The procedure to patch your 1.0.1 installation to 1.0.2-RC4 is the same as with previous release candidates and is described in the readme.txt file in the zip.

http://prdownloads.sourceforge.net/pba-vm/pba-1.0.2-rc4.zip?download

In addition to all the changes/fixes from previous release candidates, 1.0.2-RC4 fixes the following:

1. Fixed some potential bootup problems when restoring to a blank hard disk. Some users reported bootup issues when restoring to a blank, zeroed-out hard disk. I believe this issue was due to the undocumented use of the area between the MBR and the start of the first partition. The bootup problem on my notebook went away after I made a change to backup and restore this region. I also hope that this will do away with the need to run fixmbr from the Windows recovery console in scenarios where it was required.

2. added swap file label for centos (mkswap -L)

Please help me test. there will likely be an RC5 in about a week to address other issues, and after that I guess I should refrain from doing anything major to allow for enough time to test before 1.0.2 is cut on 9/1.

[kaso]

Hi

Just a suggestion /question

Can you add an option of "delete image" to main menu? so I can delete images somehow from the list and from the disk?

Now I'm doing a lot of testing an a lot of images - so I need to manually delete them from disk and vi images.toc when some of them are no lnger needed.

It can be nice to have this option when booted from cd or/and from PBA server.

Thanks a lot

[rxhui]

Hi gbjbaanb:

You may not need to run fixmbr anymore when restoring winxp to a blank hard disk. I had a similar problem while testing on my notebook (thinkpad x41), and I fixed the problem for my situation, and there's a good chance that it will fix your situation also. Please see my previous post for details on the problem & fix. The fix is in 1.0.2-rc4 which is a testing version released as a patch for PBA 1.0.1. There's also a good chance that re-installing grub may also not be necessary anymore, but I haven't tested this, and it doesn't seem to impact anything by leaving it in.

[rxhui]

this is definitely on my list of things to implement