SSD wear poll (Gentoo root disk only!)

eccerr0r

I suspect a lot of people have Solid State Drives already as main storage.

On your current most often used SSD running Gentoo, what is your media wearout indicator at? I don't care that you used a tmpfs for PORTAGE_TMPDIR, just that the machine must run Gentoo Linux most of the time. And yes, this number is dependent on how many emerges you've run on the disk, but let's ignore that for now...

This is SMART field 233 typically, and is a number from 100 to 0 where 100% is full life available (new disk) to 0% when all write cycles are consumed.

This is a limited time poll because alas, over time, SSDs will wear out and your answers will change over time.

Of the few SSDs I have, one has a really crappy SMART info that doesn't clearly tell what the wear indicator is at, and one that has finally hit 1% used and have hit swap storms before. The others are still at 100% even after many years of use.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

dmpogo · Advocate Joined: 02 Sep 2004 Posts: 3267 Location: Canada

I had SSD on my 2008 thinkpad laptop, running Gentoo exclusively. Laptop still functions, but CPU fan died and it overheats. SSD still showed (as I recall) almost no wear. This is after 9 years of use.

Muso · Posted: Mon Feb 04, 2019 5:15 am Post subject:

I use an identical 500gb pair of SSDs for my Gentoo desktop. I have / & ccache one, and /home on the other. For music, videos, pictures, documents, etc... I have a 4TB "normal" spin drive.

The wear is minimal. By the time the SSDs wear out, I'll already be on a new desktop.
_________________
"You can lead a horticulture but you can't make her think" ~ Dorothy Parker
2021 is the year of the Linux Desktop!

Ant P. · Watchman Joined: 18 Apr 2009 Posts: 6920

99% remaining on my router/fileserver/caching proxy/game server/distcc box, in service since 2012. smartctl says 55889 hours of use.

mike155 · Posted: Mon Feb 04, 2019 11:16 am Post subject:

Smartctl output of the system disk on my NFS server:

Anon-E-moose · Posted: Mon Feb 04, 2019 11:52 am Post subject:

Mine is similar to mike155

saellaven · l33t Joined: 23 Jul 2006 Posts: 646

root device (installed Feb 17, 2016):

eccerr0r · Posted: Mon Feb 04, 2019 8:19 pm Post subject:

Yeah, I suspect it depends on the manufacturer. Sandforce seems to be using 233 but it varies. But anyway I think that 4 just means 4 average erases per erase block... but it is less than 1% of the total erase endurance cycles (as witnessed by the "WORST" SMART field data) so you should be picking 1% ! (that is, unless you know your ssd will only handle 100 erase cycles...)

So far it seems most people haven't really been eating into their disks life. It is a bit unfortunate my theoretically most worn SSD doesn't seem to clearly mention its wear level. It's a 32GB mPCI so it should be easy to hit GC. I *know* I've dumped a lot of erase cycles on it but it doesn't seem to show on the SMART logs, and even worse, the numbers it does report doesn't seem to correspond to anything consistent after doing 256MB and 1GB test writes.

I've been thinking about redoing that SSD with FDE which would write yet another full erase cycle...

Anyway, I don't know if the person who voted 0% was just being facetious or someone who deliberately tried to kill SSDs for a review, but wonder what the story behind that is.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

cboldt · Veteran Joined: 24 Aug 2005 Posts: 1046

dmpogo · Advocate Joined: 02 Sep 2004 Posts: 3267 Location: Canada

Checked my desktop SSD that houses / , but not /var (where portage results) nor /home.

On since 2011, Media_Wearout_Indicator 99

Etal · Veteran Joined: 15 Jul 2005 Posts: 1931

My 128 GB bcache disk, running Gentoo since 2015.

eccerr0r · Posted: Tue Feb 05, 2019 5:12 am Post subject:

392 average erases, 87% remaining = 13% used

392/.13 = 3015

This calculates out reasonably. There are many MLC SSDs (2 bits per cell) that are rated for 3000 erase cycles. You win so far for the most provably worn out SSD :D
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

Proinsias · Posted: Tue Feb 05, 2019 3:18 pm Post subject:

eccerr0r · Posted: Tue Feb 05, 2019 8:28 pm Post subject:

93% left = 7% used
76 erase divided by 7% = about 1085

Wow, I suppose this one is one of those 3 bit per cell SSDs that wears out faster than MLC 2-bit. But it still has a lot of life left.

NeddySeagoon · Posted: Tue Feb 05, 2019 9:42 pm Post subject:

After a few years this 512G drive shows

Proinsias · Posted: Tue Feb 05, 2019 10:25 pm Post subject:

Elleni · Veteran Joined: 23 May 2006 Posts: 1270

NeddySeagoon · Posted: Wed Feb 06, 2019 10:50 pm Post subject:

Elleni,

No.

eccerr0r · Posted: Wed Feb 06, 2019 11:48 pm Post subject:

98 is 2% used... raw erase average of 21 ... so 21/.02 = about 1000 as well => You also have a three bit per cell SSD.

When the RAW_VALUE hits 1000 you've used up the guaranteed erase cycles of the chips used. SMART will have your VALUE of ID 177 hit 0 at the same time, which will trigger a lifetime warning "Drive failing now". You technically don't need to look at RAW_VALUE, as long as WORST is larger than THRESH(old) you're good.

Not sure if this is an "old age" field versus "impending failure" field... it sort of means both.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

Elleni · Veteran Joined: 23 May 2006 Posts: 1270

Thanks guys, wow thats pretty cool for a drive that is about 5 years old or even more - dont rembember exactly when I bought it and it was used gentoo root for my main desktop boxes all the time. User homes are on 2x1 TB raid0 array

The_Document · Apprentice Joined: 03 Feb 2018 Posts: 275

Very decent they dont seem to wear out very much even with the amount of utilization by portage and building packages.

https://paste.pound-python.org/show/NaOnhF6efrd9bGT4TVm5/

bunder · Bodhisattva Joined: 10 Apr 2004 Posts: 5934

post got way too big, so have a github gist. :lol:

https://gist.github.com/bunder2015/8f25de4a1ee3ad6f9476d461f38a7683
_________________

ppurka · Advocate Joined: 26 Dec 2004 Posts: 3256

I haven't got much idea how much time is left in mine.

eccerr0r · Posted: Mon Feb 11, 2019 4:37 pm Post subject:

I'd trust what SMART says and it will be fine for a long while. Unless you know the controller, that 340K number doesn't mean anything.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

cwc · Posted: Sun Feb 24, 2019 5:06 pm Post subject:

samsung ssd - about 4 years

smartctl -A /dev/sda
smartctl 6.6 2017-11-05 r4594 [x86_64-linux-4.14.83-gentoo] (local build)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF READ SMART DATA SECTION ===
SMART Attributes Data Structure revision number: 1
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
5 Reallocated_Sector_Ct 0x0033 100 100 010 Pre-fail Always - 0
9 Power_On_Hours 0x0032 094 094 000 Old_age Always - 27159
12 Power_Cycle_Count 0x0032 099 099 000 Old_age Always - 101
177 Wear_Leveling_Count 0x0013 098 098 000 Pre-fail Always - 36
179 Used_Rsvd_Blk_Cnt_Tot 0x0013 100 100 010 Pre-fail Always - 0
181 Program_Fail_Cnt_Total 0x0032 100 100 010 Old_age Always - 0
182 Erase_Fail_Count_Total 0x0032 100 100 010 Old_age Always - 0
183 Runtime_Bad_Block 0x0013 100 100 010 Pre-fail Always - 0
187 Uncorrectable_Error_Cnt 0x0032 100 100 000 Old_age Always - 0
190 Airflow_Temperature_Cel 0x0032 068 054 000 Old_age Always - 32
195 ECC_Error_Rate 0x001a 200 200 000 Old_age Always - 0
199 CRC_Error_Count 0x003e 100 100 000 Old_age Always - 0
235 POR_Recovery_Count 0x0012 099 099 000 Old_age Always - 31
241 Total_LBAs_Written 0x0032 099 099 000 Old_age Always - 9 825 497 040

hdparm -I /dev/sda | more

/dev/sda:

ATA device, with non-removable media
Model Number: Samsung SSD 850 EVO 500GB
Serial Number: S21HNXAGA26914E
Firmware Revision: EMT01B6Q
Transport: Serial, ATA8-AST, SATA 1.0a, SATA II Extensions, SATA Rev 2.5, SATA Rev 2.6, SATA Rev 3.0
_________________
Without diversity there can be no evolution:)