SSD reliability

[Aaron Rainbolt]

On Mon, Jul 18, 2022 at 4:29 PM MR ZenWiz <mrzenwiz at gmail.com> wrote:
>
> I've been having a nightmare on my main desktop since last night.
>
> I replaced my /home hard disk drive with a Patriot P210 2Tb SSD back
> in April of this year (might have posted about it here - don't
> recall).
>
> Yesterday I started seeing the missing picture image in xscreensaver,
> which is usual;ly a bad sign, but I figured I could wait until today.
> When I tried to restart xscreensaver, it could not access its
> .xscreensaver file in my home directory - permission denied. This has
> never happened before.
>
> Today, Chrome disappeared and would not restart. I'm getting a lot of
> "System problem" reports - no specific reason why that I could see.
>
> There are a lot of glaring red errors in the /var/log/syslog and
> kern.log files, but I don't understand them. They look like this
> (small excerpt):
>
> Jul 18 00:18:30 marbase kernel: [397814.149027] EXT4-fs warning
> (device sdb1): ext4_end_bio:344: I/O error 10 writing to inode 7761645
> starting block 22757393)
> Jul 18 00:18:30 marbase kernel: [397814.149028] blk_update_request:
> I/O error, dev sdb, sector 247533632 op 0x0:(READ) flags 0x3000
> phys_seg 1 prio class 0
> Jul 18 00:18:30 marbase kernel: [397814.149029] blk_update_request:
> I/O error, dev sdb, sector 1755056128 op 0x3:(DISCARD) flags 0x800
> phys_seg 1 prio class 0
> Jul 18 00:18:30 marbase kernel: [397814.149035] EXT4-fs warning
> (device sdb1): ext4_end_bio:344: I/O error 10 writing to inode 7761647
> starting block 22757394)
> Jul 18 00:18:30 marbase kernel: [397814.149038] EXT4-fs warning
> (device sdb1): ext4_end_bio:344: I/O error 10 writing to inode 7761648
> starting block 22757395)
> Jul 18 00:18:30 marbase kernel: [397814.149040] EXT4-fs warning
> (device sdb1): ext4_end_bio:344: I/O error 10 writing to inode 7761651
> starting block 22757396)
> Jul 18 00:18:30 marbase kernel: [397814.149046] EXT4-fs error (device
> sdb1): __ext4_find_entry:1612: inode #7754143: comm ThreadPoolForeg:
> reading directory lblock 0
> Jul 18 00:18:30 marbase kernel: [397814.149047] sd 1:0:0:0: [sdb]
> tag#19 FAILED Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
> cmd_age=0s
>
> I finally took the /home drive out of /etc/fstab so I could diskscan
> it, and rebooted (there's a /home on the boot drive, of course -
> obsolete but it's functional).
>
> Diskscan has been running (?) for over four hours now, with around 6-8
> ETA, and it's found a whole lot of regions that need to be rewritten.
>
> Smartctl doesn't show any alarms.
>
> Is it likely that I got a bad SSD? My boot disk with the root
> filesystem on it is a five year or more old SSD - no problem. My
> laptop has a 1/2Tb SSD in it - also no problems.
>
> Are 2Tb SSDs that unreliable?  I sure hope not, but I'm probably going
> to have to find another one because I can't trust this one anymore.
> Once it's fixed (I hope), I'll limp along with it until I can replace
> it.
>
> Opinions, pelase?
>
> Many thanks!
>
> Mark

Generally, the larger an SSD is, the *more* reliable it is. And I
would have expected Patriot to be a good brand. However, sometimes
drives are just defective, and from what you're describing, it sounds
quite likely that your drive has failed. I'd replace it ASAP.