[[6.8.3] Unnecessary overwriting of JSON-files in docker.img every 5 seconds]

No official word from the Lime Tech folks if this is going to be officially fixed? 

[[6.8.3] Unnecessary overwriting of JSON-files in docker.img every 5 seconds]

This is amazing work mgutt. This issue has plagued me for a long time and has destroyed two of my nice SSDs already.

 

Hoping to see this officially fixed in an unraid update...

 

Anyone have any idea if they'll officially reply?

[[6.8.3] docker image huge amount of unnecessary writes on cache]

That's unfortunate to hear. Can you share your results of going back to BTRFS when you have them in a few days?

Also, what's the thought process behind going to XFS? Additionally, how many cache drives did you have when you were using BTRFS?

[[6.8.3] docker image huge amount of unnecessary writes on cache]

Hey everyone, I wanted to report that I believe i'm seeing this bug demonstrated on a 6.9.2 Unraid box.

I had a cache pool of two 480 GB SSDs in RAID 1 that stopped working, which I believe it was due to excessive writes. I replaced the hardware just this morning and put only `appdata`, `domains` and `system` shares on the cache using the setting `prefer`.

Being concerned about the number of writes, I checked thees and with the server being online for ~26 minutes, the cache has experienced already 110,519 writes (~55,000 per disk).

Installing `iotop` with Nerdpack allowed me to run `iotop -ao` which showed that `[loop2]` is responsible for the majority of the writes.

 

```

Linux 5.10.28-Unraid.

root@tower:~# tmux new -s cache

Total DISK READ :       0.00 B/s | Total DISK WRITE :       0.00 B/s

Actual DISK READ:       0.00 B/s | Actual DISK WRITE:       0.00 B/s

  TID  PRIO  USER     DISK READ  DISK WRITE  SWAPIN     IO>    COMMAND

13149 be/0 root        564.00 K    135.61 M  0.00 %  0.44 % [loop2]

```

I've read that some people have been unable to have their cache drives me unencrypted and experience less writes. That's not something I'd like to do...

 

I searched online for any advice on how to fix this, found this threads:

which pointed me to this bug report.

Any advice on how to resolve this?

Thanks,

Greg