MS Server 2012 R2 BMR backup failed to complete
-
I have an MS Server 2012 R2 with a usb 3.0 disk attached for nightly full image backup using Windows Backup. It has started to complete with warnings, quoting there are 4096 bytes that are unreadable. I have done multiple checkdisk /r and rebooted the machine but the backup still fails to complete. The current OS is on a pair of 1TB WD Red's in a RAID 1 (Intel Raid). I am attempting to image the entire machine and install 4 new WD Gold's in a Raid 10 configuration.
I also installed Veeam Endpoint and tried to do a full image but that fails with a different error. I reached out to Veeam support and they tried a few things but were unable to assist.
Any ideas or would the best plan of action at this point to install Hyper-V server on the new Raid 10, and disk to vhd the old server and run it virtually?
-
Backup your data (instead of the entire system) and pray.
-
I'm with JB, don't worry about windows. Only worry about your data. But I'm guessing you probably have things like AD and GPOs, etc on that server - so you'll need to backup more things to be able to recover.
-
@i3 said in MS Server 2012 R2 BMR backup failed to complete:
... action at this point to install Hyper-V server on the new Raid 10, and disk to vhd the old server and run it virtually?
You have a second server? One that you are prepping to put Hyper-V 2016 on?
If yes, then do that, install whatever licensed version of Windows server you have, join it to the domain (assuming you have one), made it an AD server (assuming you have one) then create a migration plan for the data on the current server.
-
On a more serious note, I had a similar failure in 2015 at a client.
The RAID array failed 2 hours before I arrived on site from a scheduled week long visit to migrate things to Hyper-V 2012 R2.
Turns out the RAID array was RAID 0 when it was supposed to have been a RAID 5. I got lucky and was able to force the array online, but backups and images failed due to bad blocks.
The only VM we were unable to get a full image of was the DC, so it is obvious where the bad sectors were.
I had week old backups but as I had the system online (more or less), we decided to backup the data itself and just blowup the system.
I spun up a temp Hyper-V server on a desktop and installed Server 2012 R2, made it the DC and transferred the roles and remade the shares and copied the data over. Found 5 folders in the shares that were unable to be copied and none of them were deemed important so the were just dropped.
Dropped the old Server 2008 R2 DC from the domain and then powered down the hardware.
Installed the replacement drive, and remade it as a RAID 10, installed Hyper-V Server 2012 R2 and transferred the new DC over to it.
Migrated the SQL server (Server 2008 R2) from their second server to the new host, and then nuked that box to become a Veeam replication target for fail over.
Took two weeks, to get it all done, but the client lost nothing (they chose not to have me dig those 5 folders out of backup) and everything is still running now on this hardware.
-
It's not a seconds server. I plan on using the same hardware but 4 completely new drives.
-
@i3 said in MS Server 2012 R2 BMR backup failed to complete:
It's not a seconds server. I plan on using the same hardware but 4 completely new drives.
Nuke a user's desktop and install your Hypervisor on it as a temporary host.
-
@jaredbusch said in MS Server 2012 R2 BMR backup failed to complete:
@i3 said in MS Server 2012 R2 BMR backup failed to complete:
It's not a seconds server. I plan on using the same hardware but 4 completely new drives.
Nuke a user's desktop and install your Hypervisor on it as a temporary host.
And again - yep. Now is the time to change your network to be able to move to new VMs as needed with little to no impact to users. use DFS, use a DNS name that's not attached to a server name for network shares, etc.
-
I don't understand what's failing... writing to the backup drive, or unable to read a sector on the OS?
Since its a RAID1, you could pull a drive and clone it to a good one, then try just running only from that one. Maybe then it'll back up.
-
@i3 - Is this an HP server ? If so what raid controller does it have?
-
@syko24 It is a SuperMicro server with Intel Raid.
-
@tim_g Unable to read a sector on the OS. I tried writing to multiple usb disks and they both have the same warning. I may have discovered more info, I loaded a gparted Live CD to take a look at the partition layout of the server and received a message:
“Can’t have a partition outside the disk!”Failed to read last sector (1952800766): Invalid argument HINTS: Either the volume is a RAID/LDMbut it wasn’t’…… Unable to read the contents of the filesystem.
When I run fdisk –l I get
/dev/sdb
/dev/sda
/dev/md126
/dev/loop0Fdisk lists the Last sector of all three devices being smaller than the End Sector.
This is a SuperMicro server using Intel RAID. Two 1tb disks in a RAID 1.
I was hoping to use NTFSfix in Ubuntu to repair the NTFS partition but do not know which device I should run the command on.Should I issue the NTFSfix on the /dev/md126 device? Is that the RAID volume? Or should I perform the command on each of the /dev/sdb and /dev/sda devices?
-
@i3 said in MS Server 2012 R2 BMR backup failed to complete:
@syko24 It is a SuperMicro server with Intel Raid.
The reason I asked was there was an issue on certain HP servers where the raid controller did not properly size the disk. The fix was to go into disk management and shrink the volume a couple GB and then resize it back. Might be worth a try.
-
@i3 - just to be clear you would resize using windows disk manager not the raid controller.
-
@syko24 said in MS Server 2012 R2 BMR backup failed to complete:
@i3 said in MS Server 2012 R2 BMR backup failed to complete:
@syko24 It is a SuperMicro server with Intel Raid.
The reason I asked was there was an issue on certain HP servers where the raid controller did not properly size the disk. The fix was to go into disk management and shrink the volume a couple GB and then resize it back. Might be worth a try.
SmartArray did that? Any additional details?
-
@scottalanmiller - This is not official from HP although one of the posters mentions he got the info from HP support.
-
This was a similar situation that I ran into with a client a year or two ago. The fix was to shrink the volume a little bit and then extend it.
-
@syko24 said in MS Server 2012 R2 BMR backup failed to complete:
@scottalanmiller - This is not official from HP although one of the posters mentions he got the info from HP support.
Odd, taht thread is tough as it almost never mentions what storage platform is being used, just stuff that doesn't mean anything like the chassis model
-
@scottalanmiller - The link I posted was just something I pulled up on my phone. Again, this is something I experienced in the past and was hoping it may help resolve @i3's issue. There are a lot of similar threads out there with similar suggestions.
Here's another one. Again does not mention specific hardware but mentions HP servers:
-
Thank you to everyone who replied. So the issue was that the partition exceeded the disk layout. I ended up adding a 2TB blank disk to the system and booted into Acronis TrueImage 2014. It discovered both the RAID and the newly added disk. I performed a clone and choose the proportional option which allowed Acronis to change the partition layout on the new disk. This completed successfully. I then took out the original two WD Red drives and set them aside as an "undo" plan. I was then able to boot from the newly cloned disk and used Veeam to successfully take an image. I then inserted the 4 new disks, created the new Raid 10 virtual disk and restored the Veeam image to the Raid 10 without issue.
I realize that @Dashrender suggestion of installing the Hypervisor on the new drives would have been a better route to go, however, I am still having issues reliably managing Hyper-V Server installations in a non-domain environment. It seems like I can get it working initially, but then when I want to connect to it again from the same machine, some other part fails to connect. I feel until I can get whatever issue I am having resolved, I don't feel comfortable moving to that scenario.
