Leaderboard

I found when copying to the pool it defragments files quite badly. My solution to it is ultimatedefrag. Although any folder/file layout defragger will do just fine. I'm old school and obsessive about track to track seeks, power usage and head flying hours. One of the things i've read often mentioned here is how old your drive is. Really the question should be is the drive under 40,000 flying hours thats my benchmark for an old drive where failure rates increase to a significant level. One of the useful things i use is in the power settings, the ability to ignore burst traffic for 1 minute, and HIPM/DIPM/DEVSLP. It keeps the drives spun down until i actually need them. I tend to use a pair of seagate powerbalance SAS drives for frequent data, as i love the timer control, and running Idle_C for extended periods. Copying 8TB over to a pair of 12TB drive i had about 3 million file fragments (mostly film/tv). So i sort to end of disk- File/Folder Descending. It takes 2 days on a new drive, but it has the advantage of the free space being at the front of the disk, and the head doesnt have to seek any further than 33% into the platters unless its a film (in which case its a sequential read and even on the final sectors not alot of work even for a 4k film). It should be pointed out that i do this yearly at most. Doing a consolidation defrag every 3 months to push the new films etc on the outer tracks to backfill from the new end of freespace at 4TB. It also has the advantage of being a sequential read to duplicate the disk. Of course you could defrag to the outside of the disk but that has the disadvantage of the free space being on the end of the platter, which is a full stroke anytime you want to write data. If you really needed file recovery, and can read the platters, sequential data is much easier too. Another issue is rated workload, i tend to prefer SAS drives, so i can get away with sustained use in heavy periods. But a way to age your drives fast is to scan them on a 30 day basis for sector errors. Thats 12x per year or a workload of 144TB on a 12TB drive (Ironwolf have 155TB/year usage). A better solution is to get the new drive, allow it to warm up a couple of hours. Then do the smart tests quick then extended, then use a sector regenerator to scan the disk (theres usually a point on the platter where theres a slowdown at least on the drives i buy 12TB+ (NAS or SAS)). Scans dont need to be more than 180 days less 10days per drive age in years(min 90 days). Also as an old timer, i like the sound of noisy drives I can hear what they're doing better. But mostly the measures herein reduce most of the seek noise, so usually all i hear is parking, or a long seek (and the occasional crunching over an area of weak sectors i forgot to regen). I'm sure i forgot to mention other stuff, but its enough for an intro to me, and my personal choices in how i organzie my data. Of course i welcome any thoughts.