I have a very annoying problem: I have a 8TB Seagate SATA drive (ST8000AS0002) which randomly dies.
So it throws a I/O error and then I can not access the device.
I tried everything: Using IDE mode, using AHCI mode, other SATA wire, other SATA port, other mainboard, other CPU
other SATA controller (CPU intern (AMD Athlon 860k) or ASMEDIA chip).
But nothing helps.
The first 8TB I had had a lot of dead blocks, but this one I checked after I received it.
Here is the dmesg report: http://pastie.org/10678615
If I umount the device (I have to force it) and try to remount (it is encrypted so I use “cryptsetup open”), cryptsetup fails and dmesg reports:
[Sa Jan 9 02:33:54 2016] sd 1:0:0:0: [sdb] tag#22 FAILED Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
[Sa Jan 9 02:33:54 2016] sd 1:0:0:0: [sdb] tag#22 CDB: Read(16) 88 00 00 00 00 00 00 00 00 00 00 00 00 02 00 00
[Sa Jan 9 02:33:54 2016] blk_update_request: I/O error, dev sdb, sector 0
The only solution is to reboot the whole system… Does anybody has an idea?
It is only that device (my other Seagate and WD HDDs do not have this problem), I had this problem also with Tumbleweed.
PS: I think it appears only if I write. But sometimes I can write about 100GiB and sometimes only 10GiB and then it dies (and sometimes it does not even dies).
How is it partitioned? DOS or GPT. DOS does not support such a large drive. What file system? When it dies what do you do to revive?
Have you run smartctl?
If it has important data on it, and you have not been able to back it up, a process which might be ill advised as it can make things worse.
I’d be inclined to make a backup image of the entire hdd with ddrescue ionice
If you need the exact process let me know
A quick thing to try that costs nothing - try blowing dust out of the SATA cable(s), both ends, and the connectors they plug into. This probably isn’t the cure, but it’s worth a try. I once thought my hard drive was failing, trying to copy files from DVD backs failed. Turned out there was dust in the DVD drive cable, I replaced the DVD drive before I figured that out lol!
Over the past few years a major colo (can’t remember the company at the moment) has been publishing its hard drive replacement statistics, first all rotating drives and recently some SDDs.
The bottom line for several of their published stats is that Seagate has had a far worse record of reliability with a substantially larger number of drives replaced inside of a year and within the first few years compared to every other manufacturer (<70% reliability IIRC).
The shining leader in reliability for quite awhile now has been HGST (Hitachi) with an overall 97%+ reliablity (IIRC).
The street prices for drives reflect these ratings, go on eBay or Amazon (or similar) and you’ll see the Seagate drives discounted while the HGST drives often selling near original cost even 3 years later.
Test your drives with something like smartmon(freeware in your OSS) or a highly capable commercial product like hdsentinel (which is a commercial product).
If you can get a good enough deal and you’re only using the drive for non-critical data, go ahead and use Seagate but be aware that full backups on the drives become even more critical than if you were using other drives. Because of potential problems, I would never encrypt on a Seagate drive which would complicate analysis and recovery.
If you can afford the premium prices, choose HGST/Hitachi drives.
Toshiba <might> be worth a flyer because it bought some Hitachi factories awhile back which likely included Hitachi technology at that time.
WD is a middle of the road selection. WD and Hitachi merged, but for legal reasons kept everything separate for a number of years until <very> recently so in most cases up until now you are <not> getting Hitachi reliability.
What about the disk health informed by S.M.A.R.T.?