Resolved what just happened, should I feel confident moving forward?

[Rye-fi]

I have a hard drive I use in a media streaming server on the internet. Because it's on the internet I have it isolated in a DMZ network. So my media management strategy is to synchronize (push) my media folders from my LAN based file server to the media server in the DMZ. Lately I was getting I/O errors in robocopy when synchronizing one folder. So I upgraded to Spinrite 6.1 and ran level 2 for about an hour on the DMZ media server's storage disk. Dynastat tried to do its thing but the block was marked with a U (Uncorrectable). It reported a bunch of other blocks as fine so I terminated the run and rebooted into the OS.

After that the file that was getting the sync errors in robocopy disappeared but on the next synchronization run it was recreated and completely successful. Did Spinrite somehow inform the file system or Windows that the problem area of the drive is no longer to be used? It's been 3 days, do you think its safe to carry on or should I replace the drive? I'd run a full level 3 on it to completion but the drive is 24TBs and that'd take, uhm, a while to complete while taking my media server offline.

 

[DanR]

After that the file that was getting the sync errors in robocopy disappeared but on the next synchronization run it was recreated and completely successful. Did Spinrite somehow inform the file system or Windows that the problem area of the drive is no longer to be used? It's been 3 days, do you think its safe to carry on or should I replace the drive? I'd run a full level 3 on it to completion but the drive is 24TBs and that'd take, uhm, a while to complete while taking my media server offline.

I believe it is safe to carry on. When level 2 recovers a sector, that sector will be re-written/refreshed. When Level 2 can only recover some but not all the data in a sector, the recovered data will be rewritten along with zeros for the unrecoverable data (losing the unrecoverable data) and the sector is marked with a U. The file that sector was part of will be changed/damaged. Re-syncing will rewrite /refresh/restore the entire file.

While a full Level 3 is desirable it is as you noted time consuming. A faster approach would be a full Level 2 as only slow reading sectors will be rewritten. A few U's may occur. A re-sync will fix those by rewriting/refreshing/restoring any changed/damaged files, thus avoiding future read errors for those files.

 

[DanR]

While a full Level 3 is desirable it is as you noted time consuming. A faster approach would be a full Level 2 as only slow reading sectors will be rewritten. A few U's may occur. A re-sync will fix those by rewriting/refreshing/restoring any changed/damaged files, thus avoiding future read errors for those files.

Level 2 does one normal read on a sector. If successful, go to the next sector.

If the normal read fails, retry -> DynaStat. Vary the read speed (slower!), come at the sector from different directions, and so on. If successful, rewrite/refresh the sector. SpinRite is a versatile tool that can often do an amazing job of data recovery from slow/hard to read sectors. But it is not perfect.

If 100% of a sectors data cannot be read/recovered, then Level 2 rewrites the data recovered and zeros for the data not recovered. That unrecovered data is lost, the file is damaged. In your case, re-sync restores/rewrites/refreshes the entire file from a presumably good source, eliminating the problem.

And yes, data recovery can be time consuming.

Level 3 does all that Level 2 does PLUS Level 3 rewrites EVERY sector, fully refreshing the entire drive and providing optimum drive performance. Thus L3 typically takes much longer than L2 due to rewriting the entire drive.

This should always raise the question "Why did these U sectors occur in the first place?" If they are few and randomly located it may be nothing more than bit rot, in which case a rewrite/refresh of the data is all that is needed. But if they are mostly located in one area of the drive, that could well be an indication of possible drive failure - replace it!

 

[Rye-fi]

Level 2 does one normal read on a sector. If successful, go to the next sector.

If the normal read fails, retry -> DynaStat. Vary the read speed (slower!), come at the sector from different directions, and so on. If successful, rewrite/refresh the sector. SpinRite is a versatile tool that can often do an amazing job of data recovery from slow/hard to read sectors. But it is not perfect.

If 100% of a sectors data cannot be read/recovered, then Level 2 rewrites the data recovered and zeros for the data not recovered. That unrecovered data is lost, the file is damaged. In your case, re-sync restores/rewrites/refreshes the entire file from a presumably good source, eliminating the problem.

And yes, data recovery can be time consuming.

Level 3 does all that Level 2 does PLUS Level 3 rewrites EVERY sector, fully refreshing the entire drive and providing optimum drive performance. Thus L3 typically takes much longer than L2 due to rewriting the entire drive.

This should always raise the question "Why did these U sectors occur in the first place?" If they are few and randomly located it may be nothing more than bit rot, in which case a rewrite/refresh of the data is all that is needed. But if they are mostly located in one area of the drive, that could well be an indication of possible drive failure - replace it!

Thank you for that explanation. Based on what you're saying I think I'll get a replacement drive lined up and start preparing it. After that I'll replace it and run a full level 3 on the old drive and see if I need to do a warranty replacement.