Preliminary System Restoration Test

Preliminary System Restoration Test - ATI Echo Server / Win2K3

All,

My trial period for ATI Echo Server is fading rapidly and I'd like to run a preliminary system restoration test to check that at least the fundamentals are working. I do not have a test server so I only want to go to the step before actually writing any data to the hard drive.

I've created but have not used the rescue CD. I understand I can boot from the rescue CD into the Acronis TI Echo Server environment. From there I should be able to navigate to either the disk images held in the secure zone or the images held on the external USB drive.

What more, if anything, can be checked at this point? I don't have spare HDs to replace into the RAID to run a real restoration although I know it is best to do that.

 

All,

I've completed the first test and I simply want to post my observations before going any further. The plan was to restore a file from a disk (D: partition) image which was saved to the external USB drive.

I rebooted the server with the Rescue CD and the Acronis environment appeared as expected. I chose Recovery. My RAID1 drives were listed as separate disks which I assume is normal; my C: was listed twice as was D:. The Acronis Secure Zone was also listed twice. I could navigate through the directories of all drives.

I chose a file several folders down in the disk image saved on the external USB drive - the image was orginally taken from the D: partition. I created a new folder to restore the file into and Acronis restored the file into the new folder without any problems.

Now something unexpected happened. I exited Acronis and the server auto rebooted. After booting up I received a message that said my new device had been installed and I had to reboot. Before I rebooted I checked the D: partition and my new folder with the restored file were there. The server rebooted and did a CHKDSK of D:. I again checked D: and the new folder and restored file were ok.

What was the new device that was installed? I don't think the CHKDSK would have been normal? and regarding my RAID1 drives being listed twice, can I assume the first listing corresponds to the first HD (channel 1 on the RAID card) and the second listing is the 2nd HD (channel 2 on the RAID card)? Otherwise it would be possible to restore to the wrong HD which will be overwritten by the first drive when the RAID1 starts the rebuild.

Any comments are welcome.

 

This is not correct. If the RAID setup is seen correctly, it should show up as one drive. Showing up as two drives means the RAID setup is not being detected properly (most likely due to the lack of Linux driver support).

Does the Server version have a Safe Mode version? If it does, you might boot to it and see if it detects your RAID correctly (as one drive).

In any case, you should probably submit a support request to Acronis and ask for a custom/updated ISO that contains the necessary RAID drivers.

Trying to restore any files to a RAID setup that is detected as single drives is asking for problems. It may be what triggered the chkdsk and also the device detection. I think the only reason this half-way worked for you was because it was RAID 1. If it had been RAID 5 or RAID 0, I suspect it would have corrupted the drives.

---------

One other test you can do from the rescue environment is to run a Validation on the backup image. While this doesn't guarantee a successful restore, it does tell you if TI can successfully read the backup image.

 

This concerns me. The RAID card is an Adaptec 1220SA with latest drivers and BIOS - we are not talking about generic no-brand hardware.

Yes it does and I'll try that tonight. I recall seeing that option on the RescueCD.

I'll do that today - it appears the product cannot be used for restorations in its current state.

That would have been an extremely unattractive outcome - possibly corrupting the sytem while attempting to restore one file.

To be honest I think I should stop testing at this point, particularly if the RAID continues to show as separate drives.

 

If the image file isn't on the RAID, testing if you can run a successful Validation should be risk-free since you'd be reading from a completely different drive.

See how the Safe Mode does, though. It may work fine.

 

MudCrab,

Thanks for your assistance mate, forgot to mention that above

I logged onto the server after I posted before to find that all three scheduled tasks also failed this morning. The log indicates that the tasks started properly but 6 or 7 entries down they all report "Can not find a partition with the path \\local\hd_sign ... ".

I started the Adaptec Storage Manager which didn't indicate any issues. However I started an array verification which is currently at 17%, no issues as yet. I'll try manually running the Acronis tasks when the RAID verification has finished.

I've logged a support job with Acronis, #1328229, with a link to this thread. My main focus right now is the stability of the server.

Cheers.

 

Update..

The Adaptec Storage Manager completed the RAID verification without errors. I've attempted to manually start each Acronis scheduled backup task and each report the same error as stated above ie "Can not find a partition with the path \local\hd_sign(EDA8E9A1)\...". I created a new backup task which again produced the same error.

It seems a reinstall of Acronis is required due to some form of corruption to the Acronis software. I suspect myown actions are responsible for this after continuing with the RAID not correctly recognised.

I've been browsing the threads on this forum and, until I hear from Acronis support, I think I'll try the Acronis BartPE RescueCD first, simply to provide a fast restoration solution (I'm dead in the water if I require a system restoration right now! - although I do have a Windows ASR on stanby), then perhaps investigate Mustang's plug-in depending on how this all pans out.

 

I've uninstalled TI Echo Server and deleted the Safe Zone returning the space back to D: which is reporting its new size correctly. Adaptec Storage Manager is reporting the RAID to be fine.

I've rebooted the server twice and each time CHKDSK has run on D:, there are no reported errors. I want to resolve this before pushing forward and using Mustang's BartPE plugin.

Any suggestions?

[Edit - I've found the culprit. The folder I created and restored into refuses to be deleted. When I try I receive "cannot delete attack-drop.sdf file as directory is corrupted and unreliable". The specified file is simply the first in the list, a file from a Cisco IOS. I also tried deleting the folder in Safe Mode. How do I get around this little gem?]

[Edit 2 - this is now a Windows issue but any help from this forum is welcome. MS article suggests the file system is possibly corrupt and to run "chkdsk /f" on the volume. When I attempt that I receive a warning that the volume needs to be first unmounted and handles will be lost. What's the implication of doing that? MS also suggest backing up the volume first!]

 

If I may be so bold, what I believe mudcrab means is that if ATI sees the member disks of a RAID as individual drives rather than seeing the RAID as jsut one logical drive, then something is wrong and a backup or restore is bound to have problems.

 

That's exactly what I meant.

 

If you had a BartPE or VistaPE CD, then I would suggest you boot to that at try and delete the folder. Another option is to find a program that deletes "undeletable" files.

Also, I think I remember reading a post where someone had restored another file (with the same name) over the "corrupted" file and then they could delete it. You could try this in Windows and see if it works.

 

Hello all,

Roman1, I've checked this support case and found that you was provided with updated ISO version of Acronis Bootable Rescue Media.

Could you please check if this media is recognizing your RAID setup correctly and answer the e-mail from Acronis Support Team?

Please clarify if issue with scheduled task was fixed by reinstallation. If the issue still exist then please provide them with Acronis True Image log.

Thank you.
--
Michael Levchenko

 

Hi Michael,

Thanks for the reply. Yes I received the update ISO and have tested it on my server. The RAID1 still appears as individual disks when booted from the new ISO. The server has an Adaptec 1220SA RAID controller with the latest drivers and BIOS with 2 x WD3200KS 320Gb SATA II drives.

I have reinstalled TI Echo Server but have not scheduled any backup tasks so I do not have a log to send. I am still attempting to delete the corrupt folder on the D: partition.

I did use the new ISO CD to validate the existing images on the external USB drive so it seems I can attempt a restore of D: if I can get the RAID to be correctly identified and I'm confident enough to try it - although I have the corrupt folder the server is otherwise functioning normally except for the CHKDSK of D: every reboot.

I'll e-mail support separately.

Cheers

 

Assuming you finally get an iso that shows yours drives correctly when you try to pic a source drive for a restoration (i.e., a drive with a tib file on it) then also be sure that it sees your drives correctly when you try to select a target for restore (i.e., the target you are intend to restore to.). If you get that far and everything works, it's 99% that you can restore wtihout probs.

100% certainty comes from doing an actual restore

 

All,

I hope I'm close to resolving this. I believe I've exhausted all options with regard to deleting the corrupted folder. I'd like to post my plan here first because I'm depending on Acronis to come through. The RAID recognition from RescueCD is not resolved yet. I'm planning to image D: to an external USB drive, format D:, then restore D: from external USB while remaining in Windows (Server Win2k3 R2).

Here is why http://support.microsoft.com/kb/246026

I don't appear to be able to exclude folders if I run a disk image of D: (as suggested by MS) but it seems I can exclude folders in the Restore Disk wizard during the restoration using Restore Specified Files or Folders instead of Disk. I'm hoping this will resolve the corrupt MFT on D:.

My D: partition holds all user data including redirected folders and user roaming profiles. NTFS permissions have been maintained in the image. I've successfully validated the image and I have tested stepping through the wizard and deselecting the problem folder at the root of D: on the image.

What do people think and is there anything I can do better? <nervous smile>

 

I'm happy to report that the corrupt folder on D has been repaired - by restoring the same folder from an Acronis image (within Windows) which then allowed me to delete it!

Also, Acronis support has been excellent. I've posted my sysinfo to them and all going well I'm sure my RAID will be recognised soon.

 

Good to hear. This seems to be one of the easiest ways to get rid of this problem (if it works).

 

All,

After several weeks, and as many ISOs, the Acronis Support team have come through with flying colours. My server's adaptec RAID card is fully recognised by the Acronis Rescue CD and I've restored an image of D: including user shares, user profiles and redirected folders with all NTFS permissions in tact.

Nice work guys! You've got one hell of a product with a support team that is serious about fixing the inevitable issues that arise due to the wide range of different products available on the market.

Your persistence with my issue is appreciated and I hope readers of this thread who may have issues will now exercise a little patience with the knowledge that you are serious about customer support and identifying solutions.

Cheers - ticket closed.