Hard Drive SMART Raw_Read_Error_Rate

[MurphyZoo]

I'm having an issue with my new FreeNAS system. I've just recently built the system and am going through the burn-in testing.

I have a ASRock C2550D4i based system with 32GB ECC memory and 6 @ 4TB hard drives.
3 of the hard drives are Western Digital RED's and 3 are Seagate NAS drives.
(Here's a link to my build thread: https://forums.freenas.org/index.php?threads/new-to-freenas-please-review-my-build-plan.40269/)

I've run the smartctl short, conveyance and long tests and am now running the badblocks tests.
Everything is looking good on the WD hard drives, however, I've noticed that the Seagate drives have high Raw_Read_Error_Rate numbers.

Here's an example. The raw value is 15,196,360 for the Raw_Read_Error_Rate
1 Raw_Read_Error_Rate 0x000f 108 100 006 Pre-fail Always - 15196360

Also, the Seek_Error_Rate is high:
Seek_Error_Rate 0x000f 100 253 030 Pre-fail Always - 22201

I've checked the smart values for all 3 Seagate drives and they are all equally high (not the same value, but >10million for the Raw_Read_Error_Rate.

Does this mean the drives are bad?
The hard drive burn in guide indicates some of the more important fields are Reallocated_Sector_Ct, Current_Pending_Sector, and Offline_Uncorrectable, however, I'm not seeing errors there.

Should I go ahead and return the Seagates and get different drives (would probably buy WD RED's again).

Thanks

 

[SweetAndLow]

If the tests are passing them they are OK but I wouldn't trust them especially if they are new.

 

[danb35]

IIRC, Seagates use a funky inverse value for these values--something like, for a value of n, the error rate is 1:n. Thus larger values are better. I think I've seen @jgreco describe it a bit more precisely.

 

[MurphyZoo]

Thank you. That prompted me to do a little more research here. I had originally searched for raw_read_error rate and other SMART errors before posting, but not specifically for Seagate drives. This is a good explanation of the data: http://www.users.on.net/~fzabkar/HDD/Seagate_SER_RRER_HEC.html

It looks like I have some math to do to determine if I have a real problem or not, but given many of the other SMART values are ok (0), I'm becoming less worried about these drives.

I'm still waiting for the badblocks tests to finish (they take a while) and need to still figure out how to work the "tmux attach" command to get to the results from a disconnected ssh connection, but I'm hoping the test results will be ok (no errors). If so, I'll do some math based on the information in that link to see if there is more to be concerned about with those drives.

Thanks

 

[titan_rw]

https://forums.freenas.org/index.ph...ard-drive-with-smart-errors.12396/#post-57928

 

[joeschmuck]

It's simple, ID's 5, 196, 197, 198, and possibly 200 (Multizone) are the ones to look for a count on and if they are ZERO after a SMART Extended test, then your drives are operating fine and I wouldn't exchange them if it will cost you more money. I do however prefer WD Reds because I've had great luck with them. Read the Wiki (in my tag line) and look for the IDs in red. And ignore ID1, it is generally meaningless to us common folk.

 

[MurphyZoo]

Thank you! That information will help a lot.