willow1959

I did figure this out with a lot of digging. My solution was to go to disk settings and adjust the values for: Tunable(md_num_stripes) - Played around a lot but set it at 131072. Make sure you have enough RAM once you start going higher. Tunable(md_queue_limit) - Sat it at 256. Any higher and my rate dropped drastically. Tunable(md_sync_limit) - 20. I read that really any higher than these numbers can improve the quality of the rebuild. With these numbers I jumped from 5-10mb/s up to 110-120mb/s.

So long story short - I messed up when adding a new drive and am now having to rebuild a disk. Still can't find the new disk but that's another issue. Yesterday the rebuild was flying along at 120mb/s but from last night it has significantly dropped. Ranging from 5-15mb/s. I shut down all of my dockers except Plex so no writing from any of those. All SMART tests reported no issues on any of the other drives. I have 7 drives connected via mobo and one with an LSI. (One of the 7 isn't currently recognised - separate problem. Probably separate post once rebuild is over - it has no data). Its gone from getting finished in 18 hours to taking 5 days. Any insights would be appreciated. I've checked the logs but to be honest - not too sure what I'm looking for in terms of the dropping speed indicators. tower-diagnostics-20240908-1130.zip

I'm getting this same atomicparsley error. Just installed the docker today. Strange thing is - its not in usr/sbin Any way to manually pull atomicparsley ?