[SOLVED] Can I use NAS drive in Desktop PC?

860lacov · 2021-01-28 11:13:09

I wanted to buy WD Red PRO because for me longer warranty and speed is worth extra money compared to WD Blue.
I started to look for disadvantages of NAS drives in regular PC. I have found something like this:

TLER technology allows the drive to mark sectors as bad sectors when it can't access the sectors after continuous attempt for 3-7 seconds. This reduced the failure rate of raid array and improved reliability and stability.
However, if you use NAS hard drive as a regular drive, some sectors may be marked as bad sectors incorrectly. Because common desktop drives without TLER technology usually mark sectors as bad sectors when they can't access the sectors after continuous attempt for 30 seconds to 1 minute.

I use SSD as my system drive - btrfs
HDD I would like to use as my data drive. I edit photos in darktable for example.
I'll probably not mount this drive permanently but only when needed. File system for HDD - ext4 probably.

I have no idea if TLER technology is a big deal / big disadvantage for non array use (I may build RAID in the future but I'm not able to determin when).

Or maybe there is a way to control TLER time?

Last edited by 860lacov (2021-01-31 12:06:56)

graysky · 2021-01-28 12:38:33

Did you google this one?

860lacov · 2021-01-28 14:21:00

graysky wrote:

Did you google this one?

I googled title of my topic.
Usually therea are answers like. It ahoukd be ok.
And sometimes answers with TLER inside

Thats is the source of my worries connected to TLER

When I tried to find something about changing drives settings I managed to find that there was WD software (windows software) to change TLER settings.

I decided to ask here because my experience show that Arch users know everything

P.s.
I started on pc hardware forum.
After asking for the best hdd I was told that It doesn't matter because drives are pretty much the same.
When I asked about TLER the topic died

seth · 2021-01-28 15:09:58

The whole thing is geared towards (hw) raids.
You /can/ use it on a desktop (single disk system/raid0) but it's not advisable.

Maybe your desired HDD can toggle the feature, "smartctl -l scterc,0,0 /dev/sdX" or "smartctl -l scterc,70,70 /dev/sdX"

860lacov · 2021-01-31 00:47:19

seth wrote:

The whole thing is geared towards (hw) raids.
You /can/ use it on a desktop (single disk system/raid0) but it's not advisable.
Maybe your desired HDD can toggle the feature, "smartctl -l scterc,0,0 /dev/sdX" or "smartctl -l scterc,70,70 /dev/sdX"

I have now WD Red Pro
it seems that scterc future is here.

I checked and dafault is:

SCT Error Recovery Control set to:
Read: 70 (7.0 seconds)
Write: 70 (7.0 seconds)

I couldn't find this on Arch wiki, but I was able to find on a "regular" Wikipedia something like this:
Standalone_vs._RAID_considerations

Recommendation from there is to set "smartctl -l scterc,0,0 /dev/sdX"

When I was looking on the web I've found an information that around 7 seconds is standard for RAID and in a desktop the default is around 30-60 seconds, so drive can longer check for an error.
I tried, and I can set only what you said so:

"smartctl -l scterc,0,0 /dev/sdX" or "smartctl -l scterc,70,70 /dev/sdX"

I tried to change it to different values but there was an error

SCT (Set) Error Recovery Control command failed
Retry with: 'scterc,70,70' to enable ERC or 'scterc,0,0' to disable

Does it mean - in case of an error drive will try to fix it for 7s or for infinite time?
Can 0,0 result in PC freeze or something?

Last edited by 860lacov (2021-01-31 00:50:42)

seth · 2021-01-31 08:49:30

0,0 should™ simply disable the (time limiting) feature what should™ lead to the normal™ behavior which is some™ amount of time - depending on the firmware.
It should™ not be "forever", but longer™ (I couldn't find a tech sheet for the drive beyond "it's great!!" but it's hardly gonna be longer than 60s - unless a bug in the firmware… because of the "should" thing ;-)

Also this is only relevant when a block fails and needs to be re-mapped.

Ropid · 2021-01-31 09:16:56

@860lacov: You should keep TLER enabled and at 7 seconds! It's a good feature! It's not a bad one. The normal desktop HDDs that lock up for several minutes are bad, your drive with its short 7 seconds lock up is good.

seth · 2021-01-31 09:46:40

If you reduce the error recovery timeout the drive will age faster as block will be marked "bad" (and re-mapped) more easily - and the data on the block will be lost if it happens on a read.
It's a good raid feature because you have redundant data anyway and you want speed and not one disk stall the entire stack and so you just throw money against the problem.

I'd really like to hear an explanation why you think this is a sane idea on a single disk desktop system (where the backup strategy is likely "I have a strategy to somewhen make a backup") - esp. since every advise you'll find anywhere is to turn that off in a desktop scenario because see above. But maybe you've got a good reason beyond "FASTAAA!!! IS BETTAAA!!" that nobody thought of?

860lacov · 2021-01-31 12:04:06

I have internal drive and as a backup external drive.
I wanted red drive mostly because of longer warranty and I assume that in the future I may build some kind of nas or raid inside pc.

I'll go with 0,0

I know what to do know so thank you.

Last edited by 860lacov (2021-01-31 12:06:20)

Ropid · 2021-01-31 12:06:48

@seth:

I had around eight HDDs develop bad sectors and fail over the decades. I remember names Conner, Maxtor, IBM, Samsung, Seagate, Toshiba.

When those drives were failing, it made the PC virtually useless because the drives stayed locked up for minutes for each bad sector they ran into. There was never just one bad sector, it was always several. I never felt like the minutes of retrying actually helped. If reading the sector doesn't work after the drive does its scary "tack.. tack.. tack.." a bunch of times, there's just no hope. My bet is that the 7 seconds rule has a very similar chance to rescue data compared to the usual behavior. The usual behavior just turns working with the drive into torture.

When I had those drives start to develop bad sectors, I kept using them for fun just to see how things develop. I turned them into drives for unimportant stuff like bittorrent. Basically all of the drives died pretty fast. The bad sectors kept increasing and within days the drives were useless. There was just one drive from IBM that had its bad sector numbers increase for a while but then stop. That one drive then kept working fine for years and then one day was just dead.

Last edited by Ropid (2021-01-31 12:08:20)

860lacov · 2021-01-31 12:26:57

I have 1 TB disk since 2011
It's full right now.
It's standard 7200 samsung drive.

If setting 0,0 can give me similar effect like in this old drive then it's ok. Nevertheless I think that its possible that if hdd can't handle error in 7 seconds then 30 may not help.

seth · 2021-01-31 15:26:19

If a drive is really falling apart this doesn't matter anymore.
You need a backup and gonna spend as much time as it takes to replace the disk and restore the backup. Which is going to be more than you were listening to frightening clicks ;-)

The long delay is usually to distinct a bad sector from "well, we're just very busy down here, ok?!" where the drive can't respond because of the backlog or it's performing a self-test etcetc.
In the context of a server RAID this will make the raid figure "ok, drive dead - we drop it" what means the data now reads and writes slower because your raid shrunk - which is bad.
So instead the drive can say "Can't respond - sector may be good or bad, but we're too busy to figure that and there's a deadline to make the call. So we just guess it's damaged" and discard an actually perfectly fine sector to avoid being dropped out of the RAID.

If that happens too often, you run out of spare sectors and at this point the drive is virtually dead.

860lacov · 2021-02-02 17:51:31

I noticed that after reboot when I check drive parameters then hard disk is set to default
I checked a while ago and there is:

SCT Error Recovery Control:
Read: 70 (7,0 seconds)
Write: 70 (7,0 seconds)

Can I change this permanently or do I have to change settings every time?

seth · 2021-02-02 19:33:36

No, transient setting.
You'll have to use eg. a udev rule to auto-set it.

Arch Linux

#1 2021-01-28 11:13:09

[SOLVED] Can I use NAS drive in Desktop PC?

#2 2021-01-28 12:38:33

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#3 2021-01-28 14:21:00

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#4 2021-01-28 15:09:58

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#5 2021-01-31 00:47:19

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#6 2021-01-31 08:49:30

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#7 2021-01-31 09:16:56

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#8 2021-01-31 09:46:40

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#9 2021-01-31 12:04:06

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#10 2021-01-31 12:06:48

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#11 2021-01-31 12:26:57

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#12 2021-01-31 15:26:19

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#13 2021-02-02 17:51:31

Re: [SOLVED] Can I use NAS drive in Desktop PC?

#14 2021-02-02 19:33:36

Re: [SOLVED] Can I use NAS drive in Desktop PC?

Board footer