[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openbsd-misc
Subject:    Re: softraid0 errors after 6.8 upgrade
From:       Leo Unglaub <leo () unglaub ! at>
Date:       2020-11-23 20:36:49
Message-ID: b845f520-0bb5-b3ae-05f8-3c099c8f28e4 () unglaub ! at
[Download RAW message or body]

Hey,
thank you very much for taking the time to respond. I am going to check 
the disk itself and if that checks out okay i am going to do the large 
file trick. If that does not help i am going to do a complete reinstall.

Thanks you all for you help
Greetings
Leo

Am 22.11.2020 um 18:52 schrieb Nick Holland:
> On 2020-11-22 06:04, Leo Unglaub wrote:
>> Hi,
>> i upgraded my desktop to the latest 6.8 release. I uses sysupgrade to do
>> the upgrade and everything worked fine. But now i noticed in my dmesg
>> the following error messages:
>>
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 475440376
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 475440376
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 473833936
>>> softraid0: sd6: i/o error 5 @ CRYPTO block 477298832
> 
> sure sounds to me like your disk has issues.  Doesn't look related
> to the upgrade process.
> 
>> This only happens when i want to read certain files in /home. I checked
>> with fsck but it reports the partition to be fine. Has this something
>> todo with the upgrade? I did not find anything in the changelog.
> 
> sounds even more like your disk isn't good.
> 
> fsck checks file *system* integrity.  It does NOT check every sector
> on the disk for suitability to store data.  The percentage of the
> disk that fsck reads in the process of doing its job is very small.
> 
> If you want to test your backing disk, I'd do a:
>      # dd if=/dev/rsdXc of=/dev/null bs=1m
> where "X" is your physical disk.  If you want to see it's progress
> while running:
>      # pkill -info dd
> will tell you how much has been read so far.
> 
> IF that comes up bad, you MAY be able to "fix" your problem by
> deleting the files that are bad, then write a very large file to
> the entire partition where those damaged files were -- the disk
> will typically read-after-write verify that the data landed on the
> disk properly, and if it finds a bad spot, it will lock it out and
> put the failed write on a good spot (after you fill the disk,
> delete the "filler" file, of course).  But be aware, your disk may
> not not healthy -- yes, bad spots and reallocated space is a normal
> thing for disks, but new bad spots, not so much.
> 
> Nick.
> 

[prev in list] [next in list] [prev in thread] [next in thread]