S.M.A.R.T. tests aborted by host

[dkhath]

System: F4-212
TOS: 5.1
Drives: WD Red Pro 20TB WD201KFGX (4)

This is a new install and setup. Nothing is stored on the drives, yet. I am trying to test them fully before using them.

I've been trying to run smartctl extended tests on new hard drives, and they keep getting aborted by the host. At first, I tried running the test through the TOS GUI, and it aborted after 5 hours. Then, I thought I found the solution on running the test through SSH by watching the server to make sure it doesn't go to sleep.

Specifically, I tried this in two SSH Putty sessions: (x stands for the drive: a, b, c, or d)

First Window (watch)
watch -d--cumulative -n 30 smartctl -a /dev/sdx

Second Window (test)
smartctl -t long /dev/sdx

All four drives consistently aborted. They all stated this with "smartctl -a"

Self-test execution status: (25) The self-test routine was aborted by the host.

They all aborted an estimated 2 hours or more after I started them; however, the log structure lifetime hours do not seem correct. I am probably misreading them.

What am I doing wrong? How can I run a complete test without the NAS aborting it? I'd hate to return the NAS and get a more expensive brand just to be able to successfully run an extended SMART test.

Thanks everyone!

[dkhath]

A quick follow-up: The first test I ran late last night may have aborted itself after ~2 hours as well.

[dkhath]

Tried it again just now, and it looks like it gets aborted by host after one hour-ish.

[Gremlin]

FYI:
I ran a 'long' smart test yesterday (Tos5.1.145) on 6TB Hitachi drive. It took approx 11 hours to complete. No issues with running this and Log shows Start and Completed. {Drive is mounted as single volume - ext4}

[dkhath]

Thanks, @Gremlin. It looks like the four brand new HDs from Best Buy are defective, according to WD support. Four? Brand new? I've turned off the default 30 minute sleep mode for hard drives, then tried to re-run the extended test via ssh. It kept aborting immediately. Then, I restarted the NAS, and now it won't boot at all, and all I hear are the heads reading heavily every 2-3 seconds. I doubt I will have any HDs to re-test for another week after I return these since everyone is sold out, but I will report back with new drives (and a new brand likely) when I do.

[Gremlin]

That is tough. My eyebrows would be raised........all from same batch?
Of course, my test was on a single drive whereas your's was on a raid array 'element'. I don't know if that makes a difference.

[dkhath]

I am using all drives as single drives. Here's an interesting find: I got it to boot up and found the initiator of the repetitive "read grind" sound loop I kept hearing: TOS 5.1 > Hardware & Power > Hard Drive Sleep > Never. The default was set to 30 minutes. I had changed it in case it was causing the long test to abort. If I change it back to 30 minutes and apply the change, then the read grind stops. I waited several minutes to confirm. Then I changed the setting back to "Never", and the repetitive "read grind" sound returned. I will try Seagate drives next when I return these, and I will report back on progress. Not many people run their HDs through a long test (it usually takes a day or more), and I am glad I started doing this years ago after a new drive had become defective too soon.

[Gremlin]

"Never" setting was a known issue previously. Sorry I missed the connection. I still have my main machine running Tos5.1 ('up' for 75 days at present) but most of my 'playing' is now on Tos6 versions.

[dkhath]

I am familiar with the QNAP family and have never had any problem running tests with them. I've been running QNAP NASs for seven years without any issues, and I wanted to try something more economical since I basically only need single drives. One or two drives had to be replaced after usage and one was dead on arrival after adding it (June 2023 - 18TB Red Pro NAS - WD181KFGX)--it too failed the system SMART test and threw the red flag, but I never had any problem running the long HD tests on any of the others through their QNAP's GUI. Historically, I've had issues long ago with both major brands of HDs going bad after some time--this trained me to testing drives before using them. I vaguely recall circa late 1990s early 2000s, I ran into the click-of-death on one or two--not knowing how to handle those issues back then.

Something is fishy, though. The Western Digital Dashboard (their HD utility--current version) would not finish the long test on any of the new drives. I pulled out some older 8TB Red Pro drives (unmarked, so they could have been ones that failed, but I know I just replaced some with bigger drives recently, so one of them should be good), and it also failed the long test on both of those.

Since then, I've been trying GSmartControl to run the SMART tests from my HD docking bay on Win10. For one of the new drives, it failed the long test the first time I tried, but the second time it was still going strong passing 30% before I cancelled it. The scientist in me wants to test the new HDs in one of my QNAP NASs and run the same tests there to see how they report. I will definitely report back with those results to compare.

[dkhath]

As a quick follow-up, I discovered the aborted SMART tests are not an indicator of a bad drive and are easily interrupted by hardware or software (motherboard, power settings, drivers, etc.). I confirmed this with the team for the software Hard Disk Sentinel--a thorough testing and repair solution. The mistake I made was assuming the SMART test in the new TerraMaster NAS was a substitute for the Bad Blocks test (a Linux utility, I just now discovered) that QNAP provides in its GUI. I am concerned about my F4-212 and its software, though, in comparison to my QNAP devices because their tests are not interrupted, so I am cautiously proceeding as an educational opportunity to prove my concerns unfounded.

I will be running the read + write test from Hard Disk Sentinel on all four drives to determine if they are defective. That may take quite a while for four 20TB drives, and I will post those results here for anyone else who stumbles upon how to test new drives before using them or has similar questions and concerns. By the way, the testing for new drives larger than 8TB that Hard Disk Sentinel recommends are these two in their software: 1) Disk -> Surface test -> Read test or a full wipe test 2) Disk menu -> Surface test -> Write + Read test. The process I will use is the Write + Read test on new drives, but I will use the Read test for older drives. The good thing about either test is that it can repair issues along the way that are not physical.