Copy speeds are super slow on certain files?

[Britishblue]

I'm transferring my jellyfin media library to a ZFS pool. Previously I was using 3 drives adding up to 5tb of space. Now I have another 6tb drive and i want to add the other drives to the pool once i transfer everything to the 6tb drive. However some of the media is transferring painfully slowly. One episode of a show was transferring at 800kb/s. However it sped up a little later into copying. Other episodes on the same drive copy at 60 - 140mb/s just fine?

I thought it might be the file being corrupted but I watched the file around where the copying was taking ages and there were no glitches that I could see. I was using linux file system on the drives im copying from. Does anyone know what might be the problem? There are no smart errors on the drive.

[Britishblue]

I looked in the kernel log

[Fri May 16 18:15:07 2025] ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
[Fri May 16 18:15:07 2025] ata5.00: failed command: READ DMA EXT
[Fri May 16 18:15:07 2025] ata5.00: cmd 25/00:00:f8:0f:db/00:08:60:00:00/e0 tag 19 dma 1048576 in
                                    res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[Fri May 16 18:15:07 2025] ata5.00: status: { DRDY }
[Fri May 16 18:15:07 2025] ata5: hard resetting link
[Fri May 16 18:15:07 2025] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[Fri May 16 18:15:07 2025] ata5.00: configured for UDMA/33
[Fri May 16 18:15:07 2025] sd 4:0:0:0: [sdc] tag#19 FAILED Result: hostbyte=DID_TIME_OUT driverbyte=DRIVER_OK cmd_age=30s
[Fri May 16 18:15:07 2025] sd 4:0:0:0: [sdc] tag#19 CDB: Read(10) 28 00 60 db 0f f8 00 08 00 00
[Fri May 16 18:15:07 2025] I/O error, dev sdc, sector 1624969208 op 0x0:(READ) flags 0x80700 phys_seg 42 prio class 0
[Fri May 16 18:15:07 2025] ata5: EH complete

I see these errors consistently happening every so often. So its experiencing read errors, but it seems to be eventually successfully copying to the new location as once the files copy they do actually play ok. It just takes hours.

Is it safe to use this older drive for redundancy or should i just chuck it and get a new one? It has about 19000 power on hours.

[OddOod]

34 minutes ago, Britishblue said:

its experiencing read errors

You could tray swapping sata cables. Sometimes they die for some unholy reason

 

34 minutes ago, Britishblue said:

Is it safe to use this older drive for redundancy

Safe is a relative term. Nothing is perfect. If you have 1:1 backup you have ~0 risk. If not, but you have 2 drive redundancy, you should be totally fine. If not but you have 1 drive redundancy, you could be running a slight risk. If you have no backups and no redundancy, you play with fire. 
Honestly, as long as SMART is not throwing fits and the disk doesn't click, I'd trust it.

And 20k hours ain't nothing, I built an 8x4TB array 12 years ago with shucked WD drives. It has long since been decommissioned, but in doing so I scattered the drives across my life. Scratch disks in main and workshop rigs, one to my NVR and a couple to my dad's, 4 to friends. All daily drive ~24/7 systems. I think one of the drives has failed. 
 

[Britishblue]

1 hour ago, OddOod said:

You could tray swapping sata cables. Sometimes they die for some unholy reason

 

Safe is a relative term. Nothing is perfect. If you have 1:1 backup you have ~0 risk. If not, but you have 2 drive redundancy, you should be totally fine. If not but you have 1 drive redundancy, you could be running a slight risk. If you have no backups and no redundancy, you play with fire. 
Honestly, as long as SMART is not throwing fits and the disk doesn't click, I'd trust it.

And 20k hours ain't nothing, I built an 8x4TB array 12 years ago with shucked WD drives. It has long since been decommissioned, but in doing so I scattered the drives across my life. Scratch disks in main and workshop rigs, one to my NVR and a couple to my dad's, 4 to friends. All daily drive ~24/7 systems. I think one of the drives has failed. 
 

The very same drive copied other files extremely fast, so it's not the sata cable. Its for a jellyfin media server, and I have all my optical media stored safely so I just decided to rerip and recompress the files that were taking forever. Maybe ill use the drive that had all the problems for redundancy then so I wont have to go through the problem of re ripping my media in future. Its not as if its critical anyway as I always have a secure backup on the dics

[kokosnh]

Can you post full smart table of that drive? 
 

[OddOod]

On 5/16/2025 at 2:27 PM, Britishblue said:

The very same drive copied other files extremely fast

I should have asked earlier: is the behavior consistent? Like, the same exact file will always be slow while others are always fast?