Add option to scrub new blocks immediately after sync, update messages to reflect sync failures due to add/delete ratio.

[tehniemer]

This will scrub anything new that was added to the array after a sync, it solves the issue of some percentage of the array that hasn't reached the age threshold remaining unscrubbed for those arrays that are changing often.

[auanasgheps]

Can I get a bit more background from this scenario? Have you experienced parts of the array never being scrubbed?

[tehniemer]

I had status reports that were consistently reporting that some percentage of my array remained unscrubbed, from 10-15% on average. This is what they looked like:

SnapRAID Status
Self test...  
Loading state from /srv/dev-disk-by-uuid-7c66e24c-1ae0-4cd6-a727-f9a42bfe09e3/snapraid.content...  
WARNING! With 5 disks it's recommended to use two parity levels.  
Using 2597 MiB of memory for the file-system.  
SnapRAID status report:  
   Files Fragmented Excess  Wasted  Used    Free  Use Name  
            Files  Fragments  GB      GB      GB  
   34348    5632   21151       -    7057    2711  72% D2  
   79691    4208   20232       -    7116    2705  72% D1  
   37094    5318   22152       -    7135    2691  72% D3  
   42469    3936   20359       -    7126    2703  72% D4  
   47209    2807   14956   -96.6    9081    2716  76% D5  
  240811   21901   98850     0.0   37517   13527  73%  
 18%|                       o       *                                     o  
    |                       *       *                                     o  
    |o                      *       *                                     o  
    |*      o               *       *                                     *  
    |*      *               *       *                                     *  
    |*      *               *       *                                     *  
    |*      *               *       *                                     *  
  9%|*      *               *       *                                     *  
    |*      *       o       *       *                                     *  
    |*      *       *       *       *                                     *  
    |*      *       *       *       *                              o      *  
    |*      *       *       *       *                              o      *  
    |*      *       *       *       *                              o      *  
    |*      *       *       *       *      o                       o      *  
  0%|*______*_______*_______*______o*______o_______o_______o_______o______*  
     9                    days ago of the last scrub/sync                 0  
The oldest block was scrubbed 9 days ago, the median 6, the newest 0.  
No sync is in progress.  
The 15% of the array is not scrubbed.  
You have 29 files with zero sub-second timestamp.  
Run the 'touch' command to set it to a not zero value.  
No rehash is in progress or needed.  
No error detected.

My array undergoes fairly significant change on a daily basis, and I figured out that the unscrubbed percentage was what had changed and was not included in the default scrub options, adding the option to additionally scrub new blocks brought my array into fully scrubbed status, this is the status after adding the additional option:

SnapRAID Status
Self test...  
Loading state from /srv/dev-disk-by-uuid-7c66e24c-1ae0-4cd6-a727-f9a42bfe09e3/snapraid.content...  
WARNING! With 5 disks it's recommended to use two parity levels.  
Using 2611 MiB of memory for the file-system.  
SnapRAID status report:  
   Files Fragmented Excess  Wasted  Used    Free  Use Name  
            Files  Fragments  GB      GB      GB  
   34457    5647   21370       -    7074    2683  72% D2  
   80403    4237   20724       -    7146    2671  72% D1  
   37430    5223   22510       -    7210    2691  72% D3  
   42585    3935   20578       -    7163    2661  72% D4  
   47421    2823   15350   -91.9    9140    2662  77% D5  
  242296   21865  100532     0.0   37734   13370  73%  
 17%|   *                                                                    
    |   *                                                                    
    |   *                                                                    
    |   *                                                                    
    |   *                                                                    
    |   *                                                                    
    |   *   *                                                                
  8%|   *   *                                                                
    |   *   *                               *                  *   *         
    |   *   *                               *                  *   *   *  *  
    |   *   *                       *   *   *  *   *   *       *   *   *  *  
    |*  *   *                       *   *   *  *   *   *       *   *   *  *  
    |*  *   *                       *   *   *  *   *   *       *   *   *  *  
    |*  *   *                       *   *   *  *   *   *       *   *   *  *  
  0%|*__*___*___*___*___*___*___*___*__**___*__*___*___*___*___*___*___*__*  
    18                    days ago of the last scrub/sync                 0  
The oldest block was scrubbed 18 days ago, the median 7, the newest 0.  
No sync is in progress.  
The full array was scrubbed at least one time.  
You have 6 files with zero sub-second timestamp.  
Run the 'touch' command to set it to a not zero value.  
No rehash is in progress or needed.  
No error detected.

After adding this I've also been able to set the scrub percent and age to lower and higher values respectively and drastically reduced my nightly scrub time.

[auanasgheps]

Interesting. My array is only 2% not scrubbed, but mine does not changes as frequently as yours.

What is the time taken to scrub the new parity compared to the whole process?

[tehniemer]

The whole scrub process including both new and old takes about 1:30 on average. I have my percent and age set to 5% and 15 days. Previously, without scrubbing new, I was using about 15% and 5 days and it was taking about 4 hours.

[auanasgheps]

Thanks for the latest updates. This really needs to be merged! PS: I want to release v3.2 in main so I will adjust the version later

[auanasgheps]

Hey @tehniemer, where are all of your changes? I finally have some time to look at them, but there's only a doc correction to make. Have you moved them to your dev branch?

[tehniemer]

@auanasgheps Odd, not sure what happened, I'll open a new PR.