Hi,
This morning I woke up to a red Health light on my HP EX475. The homeserver console gave the error message "Backup Service not running" and the Backup tab under settings showed "Backup settings are not available."

Rummaging through the event viewer logs I found the little telling "Faulting application whsbackup.exe, version 6.0.2423.0, faulting module whsbackup.exe, version 6.0.2423.0, fault address 0x00063604." in the Applications tab. But better info was in the HomeServer tab, where it read:

Unexpected error 0x45d from ReadFile on D:\folders\{00008086-058D-4C89-AB57-A7F909A47AB4}\Index.4096.dat: The request could not be performed because of an I/O device error.

Client Backup server failed at d:\wssg_src\whs_pp3\qhs\src\backup\util\diskfile.cpp(449)

So I tried to manually copy the file in the error message (Index.4096.dat), but the server hung. Obviously the disk is damaged. Interestingly, all disks show as "Healthy" in the console. My immediate step, which is already underway is the run chkdsk as recommended at Microsoft social web site:

chkdsk /r
for /d %1 in (c:\fs\*) do chkdsk /x /r %1

I would like to get some input as to what to pursue next if the chkdsk cannot fix the file. Can the Index.4096.dat file be rebuilt somehow? Or do I need to give up on the backup and restore my BDBB backup? I would like to avoid the latter as the last BDBB backup is about 2 months old. Any input is greatly appreciated.

Thanks,
rev

I would also look at the client backups to see which client may also have failed backups and look for the reasons why.

I think your off to a good start and depending on how many drives your dealing with also a time consuming one but a necessary effort. If all the drives check out then you maybe forced to perform a repair of the backup database which could cause you to lose some or all of your client backups. Its at this point where you will have to determine if using BDBB to restore the database is necessary or not. At this point there is no need to jump to conclusions until you determine the condition of the drives and if CHKDSK is able to resolve issues if found.

The Storage Tab on the console indicating drives as being healthy is not a very good indicator as to the health of the drives in your server but there is an Add-In called Home Server Smart that does a great job reading the Smart Data from the drives to provide you a more accurate picture of the health of the individual drives and can be downloaded from the link below.

http://www.dojonorthsoftware.net/Freebi ... SMART.aspx

The backup database and it integrity can be reported bad if a drive pinches a fit or for some reason goes missing. The drive could very well be experiencing intermittent issues or be on its way out or could have an issue that can be corrected by using CHKDSK as your doing now. There are many variables here but you are off to a good start.

Since you did not include the drive make and model its difficult to determine if there might be other issues. For example if you are running the newer WD Advanced Format Drives known as EARS drives they would of needed to be jumperd before use as storage pool drives or aligned. If these drives are in use in your server you may want to review the information in the link below.

http://homeservershow.com/wd-ears-drives.html

When you finish your task at hand please let us know how you make out and if you require additional help.

_________________
Exploring the possibilities!

Migrated from WHS2011 to WS2012E: HIGHLANDER

~ Norco 4220 Enclosure
~ Gigabyte GA-990FXA-UD5
~ AMD Phenom II X4 995 3.2 GHz
~ 8 GB Corsair Vengeance DDR3 1600 (PC3 12800)
~ 3 Supermicro AOC-SASLP-MV8

Thanks for the quick response. The drives in my server are all WD EADS green drives. These are not advance format drives that have been reported to cause trouble to some WHS users. So the drive models should not be a problem per se, but the aging of the drives is certainly a concern.

I have the SMART addin installed. Should have looked at it today, but did not. As of a few weeks ago, no drive was reported as problematic. Of course, this could have changed in the mean time. Will check when chkdsk completed... by tomorrow??? This will give me at least insight into the internal pool drives. I also have an external enclosure connected via USB and the SMART addin cannot access the drives SMART information from the drives in this enclosure.

I have tried the DB repair earlier when I was following the MS troubleshooting instructions. It quit after 1-2 minutes. I guess reading the index file prior to repair is mandatory. I can try this again if chkdsk does not solve the issue.

My intend was to get a sense for the inner working of the backup service to estimate how radical of an approach I will likely have to make. Maybe Alex will have some input as to whether the index can be recreated from the backup data themselves. If the index file is a single point of failure, that would be disappointing.

If the USB connected drive is in an enclosure where the drive can be easily removed it can be swapped out with a drive in the servers enclosure and then you can read the Smart Data on that drive but if the external drive is not part of the storage pool it can easily be removed to rule out any affect it may have.

I have not played with the Backup Database enough to know how it fully functions but what I do know is that Alex's BDBB and its restore functions works well. I know this from using it after performing a factory recovery and then used BDBB to restore the database and was pleased with its results.

I run BDBB only when there is a significant change that I need to preserve and to protect and then I make a redundant copy of it on another server and those are set for duplication for additional protection.

_________________
Exploring the possibilities!

Migrated from WHS2011 to WS2012E: HIGHLANDER

~ Norco 4220 Enclosure
~ Gigabyte GA-990FXA-UD5
~ AMD Phenom II X4 995 3.2 GHz
~ 8 GB Corsair Vengeance DDR3 1600 (PC3 12800)
~ 3 Supermicro AOC-SASLP-MV8

I did some further research and it does not look good. From what I find it appears that if chkdsk cannot repair the index files, I will have to go with the delete option and restore the 2-month-old database from BDBB. Apparently, Microsoft provides no tools for repairing the backup set in my case. Bad design...

On the bright side, at least i do have a backup and thanks got to Alex for creating this useful BDBB tool!

Just to wrap this topic up, here are my findings from the last week. Chkdsk completed and found a problem with one file on one disk. This file was the index file with the io error above. I thought chkdsk marked the respective sector as bad and I could continue to use the disk. I was wrong. After restoring the backup database and attempting to clean the database, the same error occurred in a different file. That was the time when I ordered and new drive. Thankfully 2 TB drives have significantly come down in price over the past year. With the replacement drive life is good again.

Still puzzled about the old drive though as to why it would not work after the chkdsk repair. I am now trying some lower level HD tools to see whether I can properly mark the one bad sector found by chkdsk. It is disappointing that chkdsk could not take care of this job.