1

I bought a group of Seagate Expansion portable 4TB drives to harvest and install as a low-power RAID group. I've been testing them before breaking their cases open in case one has an infant mortality failure (since I'll lose warranty once opened).

They all seem to work fine except one drive has wildly fluctuating write speeds - at times it hums along at 120MB/sec, but more often it slows way down - anywhere between 10MB and 10kB/sec. During a bulk copy, I had a 600MB file take ~4minutes. It also may hang at the end of a copy operation taking several minutes to stop being busy (trying to shut down Ubuntu caused it to hang for several minutes while I'm assuming the drive or driver-SW timed out).

In a sense the drive seems to be technically working fine - it never reports an error, SMART only shows some historical timeout issues, you can read/write from it, no noises, etc. Not sure if "very slow" is grounds for a return. But obviously something is going on. Any suggestions? I've been doing a bulk copy of 150k files/~1TB which looks it will take about 20hours (been going for 14.5 hours so far with 200GB remaining). Once it finishes I'll do the same operation on another drive in the group to provide a baseline.

Thanks.

Notes: - I've tried this on 2 different PCs (one Windows using File explorer, one Ubuntu using Nautilus) - both PCs have USB3 controllers (used the same setup to test the other drives in the goup I'm planning to harvest) - also tried multiple USB3 cables.

Improve this question
3
  • Could it be connected to a slow USB port or have USB port driver issues? (ie did you plug it into a different port to the other disks - some motherboards have multiple controllers, not always the same chipset)
    – davidgo
    Commented May 3, 2017 at 21:17
  • This could be caused by the USB to (I assume) SATA interface in the enclosure as well. Normally you would want to run manufacture diagnostics on the drive as they often require a failure code to process the warranty claim (however you probably can just return to place of purchase without going through this). This said, you should not run desktop drives in a RAID array. If the drives encounter even the slightest problem they'll retry up to 30 seconds, causing the array to kick the drive out as failed.
    – I say Reinstate Monica
    Commented May 3, 2017 at 23:56
  • Thanks for the suggestions. I ran Seagate tests on the drive - the short self-test and the long generic (read) test both pass, while the short generic test fails (I get the same results testing all the drives of this type). But I guess a failed test may be good enough to return the unit to Seagate. I also used CrystalDiskInfo / SMART testing which also reports Good. I repeated a near-identical copy test on another drive of the same type (first deleting and then recopying the same group of files - roughly 1TB) and it took about 1/3 the time - about 6.5 vs 20 hours.
    – M Szil
    Commented May 5, 2017 at 11:33

1 Answer 1

Reset to default
0

Just a note in case it helps others - returned the drive without hassle and the replacement appears to be normal. Seagate tools was relatively useless at diagnosis in the end the only specific symptom I could establish was using Crystal Disk Info that the timeout counter in Smart (0xBC) had a non-zero value indicating timeouts were taking place. Would be nice if drives provided more feedback in these types of scenarios as I assume some internal errors must be getting logged.

Fingers crossed the group will work reasonably well in a soft-RAID setup.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .