Hi,

I've got a couple of clients setup for continuous incrementals to a permanently connected USB HDD and all collapses etc are performed on that drive. I have a couple of additional 2.5" USB drives for use as offsite storage and run a robocopy script each day so that the the client can take the offsite disk away at night.  Incrementals are taken on the hour, Mon-Fri, between 8am-6pm.  The robocopy process runs at 2:15pm Mon-Fri so the client can take it home at night after work.

From an image continuity perspective, this is my question.  Suppose my permanent disk experiences a catastrophic failure at 1:15pm. I replace the disk and copy back from one of my offsite drives to the new permanent disk.  This would effectively take the images back to 2pm the previous day.  However what about all the incrementals taken from 8am-1pm the day of the failure that were not included in the copy the day before.  I guess in this case the chain has been broken.  Is SP smart enough to recover from this and carry on or do I need to start a new continuous incremental chain (beginning with a new full backup)?

Thanks in advance for any help given.

Actually the chain will not be broken but will be continued based on the incremental prior to the image restored.  From your example you could have restored the incremental taken 15 minutes before the failure since it is not stored on the drive that failed.  Supposing no local incrementals are available and you have to restore a file from the previous day, the chain will be continued by doing a differential against the incremental prior to the restored file.

Thanks Doug,

Note that there was no need to recover the SERVER from an image, the server did not experience any outage.  Rather it's about copying back the complete "Continuous Incremental" chain from an offiste USB HDD to the new onsite USB HDD.  So in my scenario, the server did complete a number of hourly images before the onsite USB HDD failed.  Since the hourly incrementals are not contained on the offsite disk used to copy back to the onsite disk, I'm assuming the chain will  be broken?

 Cheers.

Ok, now I understand your scenario.  This will not result in a broken chain either because of the Advanced option under the Image tab "Enable self-healing incremental recovery" (this is checked by default for all incremental jobs).  When the next job runs it will detect that the last incremental created is no longer in the directory and force a differential backup with as many files in the current chain that it can find.  This differential will take as long or longer than a full backup but the resulting backup image will be much smaller, only containing the changes since the last incremental found in the chain.  Once it completes, incremental tracking is restored.