I have a pile of HDD’s, mostly old ones ranging in age from 6-13 years old. Just going through them in case I can re-use them.
3 x 2TB drives are not trusted by TrueNAS! Two different makes and models, but they are far from proper NAS/Enterprise drives. They are models such as:
Seagate “Desktop HDD”, dated 2016 (given, never used)
2 x WD Caviar Green, dated 2010 (ceased use in 2017)
I am so glad I tried ZFS a few years ago, and then eventually put it into production. Oh, and bought decent NAS drives…phew.
I’ve had mixed results with older seagate drives and almost complete bad luck with WD green drives. I don’t use them in NAS arrays but I do put them in desktops I intend to install an OS on
Interesting to hear. Unfortunately I’ve only been ‘hard drive failure aware’ during the last 3 years or so. Before then, god knows what data has been corrupted/lost.
Agreed that’s also why I love ZFS not only for hard drive failure tolerance but I have experienced plenty of data corruption transferring files over the years from old computer to new computer enough times. Back then it was just something like xcopy in windows terminal and my backups were only my school files on Zip disks (which actually survived even to this day!)…
Whether copy errors or drive errors that caused corruption I guess I will never know. But a lot of data was permanently lost I feel. The worst part is finding out the file is corrupted only when you actually need it is infuriating…the idea of a ZFS checking data for errors with scrubs gives me more peace of mind than a hardware controller RAID ever could.
Amazing that those zip disks survived eh! I used to use those in a company I worked for, always a life saver…or rather a money saver because of the file value!
One day I’ll have to merge my files, so many disk images and zip folders with duplicates, it would save me a good TB or 2 I reckon. Absolutely though, you only know if it’s corrupted when you try to access a file, I really like TrueNAS’s scrub feature, I make sure it happens every 2 weeks at the very least.
So far I’ve been very fortunate with data loss (that I know of). There was a time when I visited a girlfriends country (Slovakia) and the airport scanners killed the cameras sd card. Very luckily there was a copy of the photos on her brothers computer…really glad to return there 6 months later and retrieve them. Even so, I still just had one copy for a very long time…luck more than good judgement/knowledge!
i don’t know i’ve lucky to date , been putting aside older 2TB + drive from " to recycle / dod wiped drive" and i’ve got at least 40 good ones… all 5 years old +
i’ve got 2 truenas. one “production with new drives”
and 1 with recycled drive for project.
Hi @sy5tem , well I think you’ve been a bit lucky, but perhaps I haven’t stored these faulty drives as well as I could have (re temp/humidity). I am better these days! #RespectTheSpinningRust
Hard disk lifetimes are almost random and almost always a gamble in specific cases(large variance in lifetime data). Only in large quantities do the statistics become more significant(You might have luck with a particular brand with a low sample size, that doesn’t mean it is better or worse than a different brand if you haven’t tested 100’s of drives.).
For example, I have two old low-end 5400rpm notebook drives, and they still run fine in my desktop system(I don’t trust them, I just store steam games etc. on them)
One has 43703 hours power-on hours, and the other one has 15193 hours.
Other drives have failed me after not even a year of regular use.
Test the drives using smartctl. For desktop usecases you can use old drives. Don’t trust them with important data. Don’t trust any one drive with important data. Always plan on a drive failing in the most inconvenient moment.
I am considering to swap out the spinning 2.5inch disk from my pfsense. 51358 hours based on smart data. On the 2nd thought, since it is working, don’t fix it.
smartctl 7.2 2021-09-14 r5236 [FreeBSD 12.3-STABLE amd64] (local build)
Copyright (C) 2002-20, Bruce Allen, Christian Franke, www.smartmontools.org
=== START OF INFORMATION SECTION ===
Model Family: Hitachi Travelstar 5K320
Device Model: Hitachi HTS543232L9A300
Serial Number: 080606FB0400LEG2VT9A
LU WWN Device Id: 5 000cca 55dc14d49
Firmware Version: FB4OC40C
User Capacity: 320,072,933,376 bytes [320 GB]
Sector Size: 512 bytes logical/physical
Rotation Rate: 5400 rpm
Device is: In smartctl database [for details use: -P show]
ATA Version is: ATA8-ACS T13/1699-D revision 3f
SATA Version is: SATA 2.6, 3.0 Gb/s
Local Time is: Sat Mar 11 20:20:53 2023 +04
SMART support is: Available - device has SMART capability.
SMART support is: Enabled
AAM level is: 254 (maximum performance), recommended: 128
APM level is: 128 (minimum power consumption without standby)
Rd look-ahead is: Enabled
Write cache is: Enabled
DSN feature is: Unavailable
ATA Security is: Disabled, frozen [SEC2]
Wt Cache Reorder: Enabled
=== START OF READ SMART DATA SECTION ===
<omitted>
SMART Attributes Data Structure revision number: 16
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME FLAGS VALUE WORST THRESH FAIL RAW_VALUE
1 Raw_Read_Error_Rate PO-R-- 100 100 062 - 0
2 Throughput_Performance P-S--- 100 100 040 - 0
3 Spin_Up_Time POS--- 253 253 033 - 0
4 Start_Stop_Count -O--C- 001 001 000 - 9193034
5 Reallocated_Sector_Ct PO--CK 100 100 005 - 0
7 Seek_Error_Rate PO-R-- 100 100 067 - 0
8 Seek_Time_Performance P-S--- 100 100 040 - 0
9 Power_On_Hours -O--C- 001 001 000 - 51358
10 Spin_Retry_Count PO--C- 100 100 060 - 0
12 Power_Cycle_Count -O--CK 100 100 000 - 1212
191 G-Sense_Error_Rate -O-R-- 100 100 000 - 0
192 Power-Off_Retract_Count -O--CK 098 098 000 - 569
193 Load_Cycle_Count -O--C- 001 001 000 - 14879494
194 Temperature_Celsius -O---- 122 122 000 - 45 (Min/Max 6/52)
196 Reallocated_Event_Count -O--CK 100 100 000 - 0
197 Current_Pending_Sector -O---K 100 100 000 - 0
198 Offline_Uncorrectable ---R-- 100 100 000 - 0
199 UDMA_CRC_Error_Count -O-R-- 200 200 000 - 0
223 Load_Retry_Count -O-R-- 100 100 000 - 0
||||||_ K auto-keep
|||||__ C event count
||||___ R error rate
|||____ S speed/performance
||_____ O updated online
|______ P prefailure warning
Absolutely, although there’s a good chance that a HDD will be fine, it is a risk considering all the goings on inside the box!
For storage machines I do smartctl SHORT, LONG and the other one that I’ve forgotten. I then do a test dump of data (quarter full) and move it to other machines.
Like you, it appears I’ve had what I thought were good drives, and yet TrueNAS spots faults immediately, while OS’s didn’t seem to care.
Blimey, only last week I threw some of them away!
I guess so long as you have a config backup, might as well run it into the ground eh?
You have to be careful with WDC RED NAS drives; they do not pay well with ZFS. This may also be the same for older HDs.
Search for ’ WD Red for NAS – Now More Choices for Customers’ and it should point you to the WDC blog post. I cannot add a link here.
And, of course, I moved a few WD40EFRX HDs from a hardware RAID to ZFS and suffered badly with ZFS reporting errors that were not due to a failing HD. Until I came up with a solution and edited crontab for each of the HDs:
