drive errors

[add212]

timeline

2 weeks ago decided to move from a small chasis with enough room for 12 drives to a rack with 4x3cages

1 week ago finally got everything in and tested

4 days ago started noticing drive errors on drive 9 and 10 and parity.  shit must be a bad cable since i swapped out that cable with a new one from monoprice 8087-4 sata.  switched it out

 

3 days ago errors are now double what they were SOB..ok wasnt a cable.  lets do a parity check before a shutdown this time. fixed a few thousand left much more.

 

and now today finally have time off work to check into it.

 

find below my long smarts for all three drives, sadly forgot to get a copy of syslog so while your reading this gonna start it up again and get a quick log as soon as the errors start showing up.

 

Nov 29 04:57:59 Tower55 kernel: md: disk0 read error, sector=

Nov 29 04:57:59 Tower55 kernel: md: disk9 read error, sector=

Nov 29 04:57:59 Tower55 kernel: md: disk10 read error, sector=

and theres tons of them for those three logs are too big working on that now

 

*updated with syslog

*update2 just remembered when i upgraded another unraid box i added the ram i had replaced to this one , removed now and back to another parity check

smartSDB.txt

smartSDC.txt

smartSDE.txt

syslog-2014-12-03_1.zip

[add212]

~checked cables..done

~removed one hdd and replaced it with a rebuild to see if by chance it was a near miss... nothing

 

now on to reiserfsck telling me to rebuild sb.  Now there are two drives with several thousand errors disk 9 and 10.  By the wiki I NEED to know how to answer the new few questions it will be asking.

 

I also read somewhere that reiserfsck can not help with the errors on the parity?  If a parity check has already happened would it have changed the parity to contain errors it now thinks are correct?

 

thanks so how to rebuild-sb?

[add212]

as it happens the 3 drives with errors are located on an m1015 I just picked up from ebay.  Besides the obvious errors any tips on either firmware or settings it should have.  I think ive looked at most of the posts but I could have missed something important?

 

thanks

[dgaschk]

See check disk file systems in my sig.

[add212]

as per the directions i need to do a rebuild tree.

below is a copy of the reiserfs check any hints or tips are appreciated before i attempt a rebuild tree

 

thanks for your time and patience

rebuild-tree.txt

[dgaschk]

Run with rebuild tree.

[add212]

rebuilt and now doing a parity check for my sanity.

 

How to access the lost and found folder and what to do with it?

 

thanks

[itimpi]

rebuilt and now doing a parity check for my sanity.

 

How to access the lost and found folder and what to do with it?

By default the lost+found folder will have permissions that stop it being accessed over the network.  This can be fixed by running a command of the form:

newperms /mnt/disk?

where ? corresponds to the disk in question. 

 

As to what to do with the contents, that is determined as to whether you can make sense of the files in it.  If you are lucky the filename information remained intact so you can see what they are.  If not they will have been assigned cryptic names, and you can only determine what they are by inspection of the contents.  Whether sorting out such cryptic entries is worth the effort involved depends on how important you think the items are to you.