I have one parity drive, three storage drives, and one cache drive.  The main status page lists errors only on the parity drive.  None of the other drives are listed with errors.  Does this mean I need to replace the parity drive (or could the errors actually be caused by a different drive)?

 

Thank you.

I have one parity drive, three storage drives, and one cache drive.  The main status page lists errors only on the parity drive.  None of the other drives are listed with errors.  Does this mean I need to replace the parity drive (or could the errors actually be caused by a different drive)?

 

Thank you.

Those are read errors.  If they were write errors the drive would have been disabled.  Typically, they indicate un-readable sectors on the disk, or bad cabling to the disk, depending on the specific error encountered.

 

Only way to learn the health of the drive is to get a SMART report for analysis.  Look for re-allocated sectors, or sectors pending re-allocation.  Post the entire smart report here foe guidance.

 

smartctl -a /dev/sdX

where sdX =  the three letter device associated with the drive.

 

The if it is an unreadable sector, the sector may already have been re-allocated, but you will not know it unless you get the smart report.  You can also take a look in the syslog.  Evidence of the actual error will be there unless it has aged out by syslog rotation, or you've rebooted.

 

Joe L.

Hmmm...I thought I previously posted my syslog.  Perhaps I got interrupted.  I tried to attach my syslog, but the file is too large. One of the errors is shown below:

 

Jun  5 03:40:43 Data kernel: md: disk0 read error

Jun  5 03:40:43 Data kernel: handle_stripe read error: 1693473984/0, count: 1

 

The smartctl command has the following output (partially typed in, as I don't know how to copy the text from telnet):

 

>> Terminate command early due to bad response to IEC mode page

A mandatory SMART command failed:  exiting.  To continue, add one or more '-T permissive' options.

 

Also, I just realized the temperature for the parity drive is "0" on the web interface. 

 

I take it this drive is toast?

Hmmm...I thought I previously posted my syslog.  Perhaps I got interrupted.  I tried to attach my syslog, but the file is too large. One of the errors is shown below:

 

Jun  5 03:40:43 Data kernel: md: disk0 read error

Jun  5 03:40:43 Data kernel: handle_stripe read error: 1693473984/0, count: 1

 

The smartctl command has the following output (partially typed in, as I don't know how to copy the text from telnet):

 

>> Terminate command early due to bad response to IEC mode page

A mandatory SMART command failed:  exiting.  To continue, add one or more '-T permissive' options.

 

Also, I just realized the temperature for the parity drive is "0" on the web interface. 

 

I take it this drive is toast?

zip the syslog.  They compress well

You were right about that!  Here's the file.  Many thanks for helping.

syslog.zip

You were right about that!  Here's the file.  Many thanks for helping.

OK, the disk has suffered many read errors, and many write errors.

 

It could have failed, or it could have a loose cable, or seated poorly in a drive tray.

Right now, since it has stopped responding, you should probably stop the array, power down, and re-seat the cables.

 

Then, power up and see if the disk responds to a smart report request.

 

smartctl -a /dev/sdc

 

Joe L.

Thank you for the help.  I would hazard a guess it's failed since I've been using this system for a few months, and I never go in the room it's in (it's in a utility room on a shelf near my network connections).  So, it's unlikely the cables came loose. 

 

Can I keep using the array, or should I shutdown the array, too?  I have a drive on order, but it'll take me two days to get the drive, a day to test and zero it, etc.  It'll be at least three days, and I'm going on vacation at the end of the week.  So, unless I can get all this done and the array back up, I can just shut down the array and serve the house from my backup drive.