| View previous topic :: View next topic |
| Author |
Message |
gengreen
Apprentice
Joined: 23 Dec 2017
Posts: 150
|
 Posted: Wed Dec 04, 2019 12:48 am Post subject: Seagate : Seconds brand new hard drive doing it |
 |
|
Hello,
I unfortunately getting some serious trouble with a seagate hard drive :
Model :
Seagate BarraCuda ST500DM009 500GB 7200 RPM 32MB Cache SATA 6.0Gb/s 3.5" Hard Drive Bare Drive
I tried again today to install Gentoo on it, at the stage3 steps, when tar xpvf stage3*... the hard drive fail with those log :
http://dpaste.com/2T9BHNF
I tried some generic distribution (Ubuntu / Fedora), I never could finish the install... I never experienced any problem with another brand like kingston/samsung.
This is the second brand new hard drive, exact model doing this. Did I miss something ? Someone known wth those log mean ?
Thanks |
|
| Back to top |
|
 |
sdauth
l33t
Joined: 19 Sep 2018
Posts: 687
Location: Ásgarðr
|
| Posted: Wed Dec 04, 2019 1:43 am Post subject: |
|
|
| Have you tried to replace the SATA cable ? And / Or connecting the drive to an other SATA port to see if it shows the same issue ? |
|
| Back to top |
|
|
gengreen
Apprentice
Joined: 23 Dec 2017
Posts: 150
|
| Posted: Wed Dec 04, 2019 2:27 am Post subject: |
|
|
yes :(
My sata cable aren't old and work fine with my ssd kingston and my HM250HI samsung (2010 model still alive  ) |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 23193
|
| Posted: Wed Dec 04, 2019 3:10 am Post subject: |
|
|
| What does the drive's long self test report? |
|
| Back to top |
|
|
sdauth
l33t
Joined: 19 Sep 2018
Posts: 687
Location: Ásgarðr
|
| Posted: Wed Dec 04, 2019 3:18 am Post subject: |
|
|
Not easy to "debug" then.. One time I solved a similar issue by switching the SATA cable (Never knew these could fail..)
Have you tried to run a badblocks test on it ? Or before that a SMART test ? A short smart test should be enough to spot hard failures (Uncorrectable sectors etc..) |
|
| Back to top |
|
|
gengreen
Apprentice
Joined: 23 Dec 2017
Posts: 150
|
| Posted: Wed Dec 04, 2019 3:38 am Post subject: |
|
|
I never did a test / report for a hard drive... smartmontools is the tools to go I suppose, will read the doc.
Something else, I was looking into the boot log :
http://dpaste.com/2WCPMQS
I found some topics on Google with the similar error, as the hard drive is brand new, I could definitively have missed something...
Thanks |
|
| Back to top |
|
|
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Wed Dec 04, 2019 6:51 am Post subject: |
|
|
| Maybe not the SATA, but the power cable/supply at fault? Seems a bit strange for two new drives to not work, but it could also be a bad manufacturing batch of them. |
|
| Back to top |
|
|
gengreen
Apprentice
Joined: 23 Dec 2017
Posts: 150
|
| Posted: Sun Dec 08, 2019 3:16 am Post subject: |
|
|
smartctl report :
http://dpaste.com/3AH9N60
Ant.P I tried with differents power supply / sata cable (including brand new cables), still the same...
I don't known much about Seagate and the quality of their product but I also own a external usb hard drive Seagate SRD00F, it work badly (slow, sometime fail to read / write...), so I'm wondering too if the problem is not a manufacturing defect...
Edit : I just saw that the hard drive is still under warranty, will try to get a replacement. |
|
| Back to top |
|
|
C5ace
Guru
Joined: 23 Dec 2013
Posts: 489
Location: Brisbane, Australia
|
| Posted: Sun Dec 08, 2019 10:12 am Post subject: |
|
|
In my experience HITACHI (now HGST) are the most reliable. I run 4 of them as RAID 5 array 24/7 since 2008. Then Western Digital Black. More than 90,000 hours without problems.
_________________
Observation after 30 years working with computers:
All software has known and unknown bugs and vulnerabilities. Especially software written in complex, unstable and object oriented languages such as perl, python, C++, C#, Rust and the likes. |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 55015
Location: 56N 3W
|
| Posted: Sun Dec 08, 2019 10:55 am Post subject: |
|
|
gengreen,
| Code: | ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 8 |
That drive has 8 blocks that is know it cannot read. The data there is lost.
Its scrap. There is no use far a drive that cannot read its own writing.
| Code: | ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
5 Reallocated_Sector_Ct 0x0033 080 080 010 Pre-fail Always - 25792 |
The drive is supposed to detect failing sectors and relocate the data before they become unreadable.
Its done that with 25,792 sectors.
As its a new drive
| Code: | ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
4 Start_Stop_Count 0x0032 100 100 020 Old_age Always - 52
9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 197 |
its probably suffered transit damage.
Return it.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
gengreen
Apprentice
Joined: 23 Dec 2017
Posts: 150
|
| Posted: Fri Dec 13, 2019 5:15 am Post subject: |
|
|
| NeddySeagoon wrote: | gengreen,
| Code: | ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 8 |
That drive has 8 blocks that is know it cannot read. The data there is lost.
Its scrap. There is no use far a drive that cannot read its own writing.
| Code: | ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
5 Reallocated_Sector_Ct 0x0033 080 080 010 Pre-fail Always - 25792 |
The drive is supposed to detect failing sectors and relocate the data before they become unreadable.
Its done that with 25,792 sectors.
As its a new drive
| Code: | ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
4 Start_Stop_Count 0x0032 100 100 020 Old_age Always - 52
9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 197 |
its probably suffered transit damage.
Return it. |
Nice thanks for those details! As the hard drive is under warranty until 2020 and your confirmation that the drive is indeed defective, I will just push the replacement process.
In a general way, what hard disk brand do you recommand (the one that you would buy) ? |
|
| Back to top |
|
|
sdauth
l33t
Joined: 19 Sep 2018
Posts: 687
Location: Ásgarðr
|
| Posted: Fri Dec 13, 2019 6:26 am Post subject: |
|
|
| gengreen wrote: | | In a general way, what hard disk brand do you recommand (the one that you would buy) ? |
Of all drives I owned from various brands, only Segate ones failed badly. In my case it was 3.5" 1.5TB.
Anyway, YMMV as always..
I recommend Western Digital or even HGST though.
Most of the time and as NeddySeagoon said it, errors like bad sectors happen because of transit damage (poorly packaged..) so one thing you should always do when you buy a new hard drive online is to run a badblocks and a smart test when you receive it. Do no put any data on it before being sure it is 100% clean.
| Code: | badblocks -wsv /dev/sda #Warning, this is a destructive test, do not issue this on a drive with data you care about on it.
smartctl --test=long /dev/sda |
After that, if there is no errors, you can consider it is safe for operation. |
|
| Back to top |
|
|
gengreen
Apprentice
Joined: 23 Dec 2017
Posts: 150
|
| Posted: Fri Dec 13, 2019 6:52 am Post subject: |
|
|
| sdauth wrote: | | gengreen wrote: | | In a general way, what hard disk brand do you recommand (the one that you would buy) ? |
Of all drives I owned from various brands, only Segate ones failed badly. In my case it was 3.5" 1.5TB.
Anyway, YMMV as always..
I recommend Western Digital or even HGST though. |
Speaking about the wolf :
Disk /dev/sdb: 931.5 GiB, 1000204886016 bytes, 1953525168 sectors
Disk model: HGST HTS721010A9
I have on too, can't remember when I got it, but it's been a long while... Good to known the good brand since benchmark website are providing false information
https://hdd.userbenchmark.com/SpeedTest/4115/HGST-HTS721010A9
"Top Hdd" : https://hdd.userbenchmark.com/ (top 15 HDD seagate..._)
| Quote: |
Most of the time and as NeddySeagoon said it, errors like bad sectors happen because of transit damage (poorly packaged..) so one thing you should always do when you buy a new hard drive online is to run a badblocks and a smart test when you receive it. Do no put any data on it before being sure it is 100% clean.
| Code: | badblocks -wsv /dev/sda #Warning, this is a destructive test, do not issue this on a drive with data you care about on it.
smartctl --test=long /dev/sda |
After that, if there is no errors, you can consider it is safe for operation. |
I added those 2 command into my personal bible, thanks |
|
| Back to top |
|
|
AJM
Apprentice
Joined: 25 Sep 2002
Posts: 195
Location: Aberdeen, Scotland
|
| Posted: Fri Dec 13, 2019 8:37 pm Post subject: |
|
|
| gengreen wrote: | Nice thanks for those details! As the hard drive is under warranty until 2020 and your confirmation that the drive is indeed defective, I will just push the replacement process.
In a general way, what hard disk brand do you recommand (the one that you would buy) ? |
Be careful with warranty replacement drives - they are often not brand new good drives, but "recertified" drives which are even less trustworthy than their brand new ones.
I have spent the past decade and a half replacing failed hard drives and Seagate are by far the worst in my experience - and that includes their "Enterprise" drives too. That said, all hard drives are a lottery and I'm very happy to see them die off; solid state drives also fail but not nearly so frequently in my experience. |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 55015
Location: 56N 3W
|
| Posted: Fri Dec 13, 2019 8:54 pm Post subject: |
|
|
gengreen,
I like HGST.
Western Digital upset me. I had two drives is a raid5 set fail with no warnings. They failed 15 min apart, so the raid set went down.
I have also lost one HGST drive from a raid5 set. It gave advance warning and I was able to replace it before it failed totally.
The other drives in that raid set are over 10 years old now.
As AJM says, warranty replacements are usually not new drives. They are refurbished warranty returns.
When you return your drive it will join the pool in process.
-- edit --
badblocks has not been useful on HDD since they reached 4GB in size.
All it usually does is trigger the bad block relocation mechanism.
Still, if you check the smart data before and after the test and you have lots of reallocated sectors, you have a problem. That does net need badblocks. A whole surface write will do that.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
alexander-n8hgeg5e
n00b
Joined: 02 Nov 2019
Posts: 58
|
| Posted: Fri Dec 13, 2019 9:48 pm Post subject: |
|
|
Hard drives fail usually more frequently early in their life and when they get old.
I'm not a drive expert but, because of the reloacated sectors
i would think i'ts broken and not cable related.
I had one that failed not after 2hours but a few days. |
|
| Back to top |
|
|
Tony0945
Watchman
Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA
|
| Posted: Fri Dec 13, 2019 10:03 pm Post subject: |
|
|
I never buy anything but WD Black. Only had one drive start to fail and that was at 4.5 years. They replaced it with a refurbished drive that they warranted for six months (the remainder of the original drive's warranty).
There is a "transit" check that smartmoncontrol can do too.
Some swear at WD, and praise Seagate. I can only relate my experience.
The WD Black on this drive is nearly ten years old, running almost continuously. Smartmon shows zero errors.
I'm considering switching to an SSD, solely for speed. |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 23193
|
| Posted: Sat Dec 14, 2019 1:36 am Post subject: |
|
|
| gengreen wrote: | | I have on too, can't remember when I got it, but it's been a long while... |
Many drives can report their total power-on hours via smartctl. This isn't exactly when you got it, but for drives in continuous service, it's probably pretty close. Even if it is wildly off (for example, if you only run the drive 4-6 hours a day), power-on hours may be more interesting than knowing how long you have owned the drive. |
|
| Back to top |
|
|
alexander-n8hgeg5e
n00b
Joined: 02 Nov 2019
Posts: 58
|
| Posted: Sat Dec 14, 2019 5:18 am Post subject: |
|
|
My disks are old, and they do not care about this, at least until now.
Just put more of them in the device.
Mix the models good up.
And make raid1's to combine them down.
Then put btrfs raid1 over that.
Btrfs can fix bitflips, if the drives fail to find them.
I have my disks somewhat frankenstein... One disk died young.
One is to small so i have lvm-combined the wd-blue and the wd-black.
So actually 3 x redundancy.
The other 2 are hitachi ultrastar and one is wd-green
I have discovered something,
smartctl -s aam,128 i think makes them quiet.
This should probably help for longer life.
Device Model : WD5000AAKS-00A7B2
Power_On_Hours : 30998
Device Model : WD7502AAEX-00Y9A0
Power_On_Hours : 12318
Device Model : HUA721010KLA330
Power_On_Hours : 41329
Device Model : WD10EADS-00M2B0
Power_On_Hours : 15960
I know, they draw power, they are slow, and make noise.
But my data is on a real disk, and they do not die if i write
to the same sector a few times. |
|
| Back to top |
|
|
Goverp
Advocate
Joined: 07 Mar 2007
Posts: 2216
|
| Posted: Sat Dec 14, 2019 10:25 am Post subject: |
|
|
FWIW, my desktop box has 4 WD Blue 320 GB drives in a RAID 5 array; 1 of them is 13 years old (the original non-RAID boot drive), the others about 7 years old. Only one recorded error in SMART, no reallocated sectors, no problems. The box is used every day, usually powered off overnight. It survived power cuts and the like.
_________________
Greybeard |
|
| Back to top |
|
|
steve_v
Guru
Joined: 20 Jun 2004
Posts: 416
Location: New Zealand
|
| Posted: Sat Dec 14, 2019 11:31 am Post subject: |
|
|
| Goverp wrote: | | FWIW, my desktop box has 4 WD Blue 320 GB drives in a RAID 5 array |
FWIW, I learned the hard way why RAID 5 sucks, when I failed to check out backblaze before buying a batch of 8 ST2000DL003 drives (hint, same platform as the ST1500DL003 in the table). Recovering ~9TB of data sucks quite hard.
Personally, with a 4-drive setup I'd go RAID 10 all the way, rebuild times for RAID5 are a disturbingly long period to be without redundancy. At 320GB it's probably not too scary, but it will only get worse with bigger drives.
My 8-drive array is RAIDZ 6 right now, all seagate skyhawks (they were cheap), no problems at 3 years in. Upgraded from 3TB WD (fecked if I can recall the model) drives that were bad to the point I trashed them all rather than giving the live ones away as I usually do.
Some drives are good, some are not. I have found that manufacturer makes far less difference than model.
Even HGST with their glowing reliability record used to be something else - something that was big, blue, and apparently no moon. Now they're owned by WD anyway.
_________________
Once is happenstance. Twice is coincidence. Three times is enemy action. Four times is Official GNOME Policy. |
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 55015
Location: 56N 3W
|
| Posted: Sat Dec 14, 2019 11:44 am Post subject: |
|
|
alexander-n8hgeg5e,
| Quote: | | smartctl -s aam,128 i think makes them quiet. |
It makes them quiet by slowing the head movements. That ruins the seek times.
If seek times don't matter, go for it.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
krinn
Watchman
Joined: 02 May 2003
Posts: 7471
|
| Posted: Sat Dec 14, 2019 1:06 pm Post subject: |
|
|
it doesn't matter the manufacturer or the drive model this much, normally a higher quality product have higher quality components and should gave higher reliability
still even having a ferrari doesn't prevent you from having a fail car.
your best clue for reliability is given by the manufacturer
if you look at what WD say, all black, red or blue have high reliability ; but if you look at WD number of years they warrant their drives, difference appears |
|
| Back to top |
|
|
alexander-n8hgeg5e
n00b
Joined: 02 Nov 2019
Posts: 58
|
| Posted: Sat Dec 14, 2019 5:22 pm Post subject: |
|
|
| NeddySeagoon wrote: | | If seek times don't matter |
Some of these species, are crazy and it sounds like they
are trying to vibrate their heads of... Especially the wd-black one,
sounds like some angry beast. It just sounded so unhealty
so i let them chill.
I compensate it by adjusting the io scheduler.
I believe in the kernel, if everything goes right,
anyway all data is in the cache, and in the case it comes to
writing, there will be a smooth movement and the head
draws calmly 100Mb/s down for half a second
at least that's what i dream of...
It would be cool to build some ssd killer hdd
array...
Some time ago i bought 4 cheap,used 2.5" harddisks from
ebay. I connected them with the cheapest sata2usb2.0 housings
to a laptop. It was 4x redundancy. On top I had btrfs.
What happened was, I lost a complete virtual machine image.
Actually only one file. I think the usb power was way to unstable
for the 4 drives. Definitively not what i would do if it was serious.
So but i was somewhat confused, i thougt the data handling would be more robust. |
|
| Back to top |
|
|
steve_v
Guru
Joined: 20 Jun 2004
Posts: 416
Location: New Zealand
|
| Posted: Sat Dec 14, 2019 6:27 pm Post subject: |
|
|
| alexander-n8hgeg5e wrote: | Some of these species, are crazy and it sounds like they
are trying to vibrate their heads of... Especially the wd-black one,
sounds like some angry beast. |
You've never owned a Quantum Fireball (or Atlas for that matter), have you. 
Speaking of old, reliable and stupidly noisy drives... my 80MB Quantum ProDrive ELS (the first "real" HDD I ever owned) is still going strong. As is a 120MB Seagate of about the same age.
_________________
Once is happenstance. Twice is coincidence. Three times is enemy action. Four times is Official GNOME Policy. |
|
| Back to top |
|
|
|
Kernel & Hardware
