Thanks for the quick reply - especially since I have people arriving soon!  

 

So my plan is to continue to use my server as is tonight to watch movies and listen to music.  Great! 

That will work.

If I'm going to rebuild my array, could I use a brand new 1TB that I already have?  Or is that not advised until all is working correctly for a while on the old drive?

If you strongly suspect it is the drive itself, sure... you can rebuild onto the new disk.  If you do put a new disk in the array you CANNOT use the trust-my-parity procedure.   You would only end up with bad parity for everything and a completely empty disk at that.  You MUSt use the set of steps I described where you stop the array, un-assign the old and re-assign the new drive and then press "Start" (do NOT press "restore" in this situation... it will throw away your parity and make recovery impossible if the old disk is really unreadable)

I kind of understand the trust my array procedure but all that bold red type at the top makes me wonder if I'm better off just rebuilding.  I'm trying to understand it but -- I had a parity check end this morning with 0 errors.  I'm sure I was writing to the disks before, during, and after the disk was disabled.  I don't know if it was that exact disk though because I use user shares.  Realistically, I don't care if I loose any data between my last parity check and the time this disk was disabled.  To make it even more confusing, I use a cache drive.  I suppose this means nothing is ever written the array disks until 3:40am when cron starts up mover.

The warnings are there because you must think your existing drives that were last used to build parity are ALL still working and it was a losoe cable or equipment change that allows you to bring the disk back online and trust parity is still as it was.    As already said, if there is a real problem it will just be taken off line once more.  If it was a lose or bad cable, you will be able to get back running quickly.

I'll probably just try to rebuild on that new 1TB.  Please let me know if that is a bad idea.

That will work.  If you have a spare controller port, consider running it through one or more preclear_disk.sh processes.  It will prove it is a good working disk before you add it to your array.

Also, thanks JoeL for

Once unRAID marks the disk as "red" it will not restore it unless you take the following actions.

Stop the array

un-assign the disk marked as red

reboot the server

if the array starts on its own when rebooted, stop it once more

re-assign the disk

start the server once more.

 

If you go through those steps it will think a new disk was installed and proceed to rebuild it.  That will take a number of hours, more if you are serving movies too.

putting your response in such an easy to follow format.  That is very COOL!

 

You are welcome.

 

Joe L.

The disk is probably completely fine, but it had a weird problem (typical lately!) that I don't know anything about.  It kept reporting and pausing for a Dispar error flag, which according to the Libata error messages wiki, means there was an "Incorrect disparity detected", if that helps anyone!  I *think* it may be communications related, but it is not like any other cable error I have seen.  Or perhaps the drive firmware crashed.  The kernel had an issue at one point, with an execution trace that seems to point to ata related driver crash, but afterward, the syslog continues with the same error sequences, no better and no worse.  The drive finally stopped responding correctly, and was disabled.  That usually means the data on the drive is fine, completely intact.  And it looks like in your last syslog that you got the array back up and running fine, all green.  You don't have anything to fear running the Trust My Array procedure, if needed.

How long has it been since you did a thorough 12 hour memory test?  I suggest you find a time to test it again.  Also check for heat issues on the motherboard and addon cards.

RobJ -- thanks for the details on the Dispar error flag.  I really have to learn how to make sense out of the syslog.  Still looks like hieroglyphics to me.  Seems like your process is to look up the questionable entries on wiki or google.  I'll try to figure some of that out from now on.

 

I'm going to do the trust parity process tonight and then run some parity checks over the next few weeks or months to make sure all is good before I try to replace for the larger drive.

 

I've never run a memory test.  I could do it easily.  I think your talking about choosing the memtest boot option when unRAID first boots, correct?  I should just start that and let it run for 12 hours or so and look for any error messages?  I'll do so and report back sometime.  Thanks.

 

Interesting comment on the heat.  I have 3 Chenboro 3x5 (so 15 external hot swap bays).  Unfortunately, with a cache drive, swap drive, and parity drive, all my slots are full.  I very recently mounted a 1TB on a couple of 3.5 INCH hard drive mounting brackets (http://www.pccables.com/00503.htm and then velcro'd the brackets on my PSU which puts the drive right below my exhaust fan in the back of the case.  The way it is mounted means that the exhaust fan pulls air across the top of the drive.  While I had my case open, I decided to try to improve airflow.  I taped up all vents and anywhere air could get in or out.  Then I reversed the other case fan that is on top of the case to be intake instead of exhaust.  Each of the three Chenboro's have a fan that draws air through the front.  What I was trying to get was all fans (3 Chenborors, 1 on top of case) as intake and only one exhaust fan out the back.  Maybe this is not improving things but making them worse or unstable.  Is there an easy way to measure the temp on a AB9-Pro mb?

 

Thanks for the ideas.  I'll have to try to figure out (and tune) the airflow.  I know there's a pretty long thread in here I'll have to find again.  Appreciate all the help!!

I've tried to collect links to a few of the discussions on fans and air flow, in the UnRAID Topical Index, Fans section.  Not sure if the one you are thinking of is there, as I know I'm missing some, but there is some very good discussion there.

 

If you have never run one, then I really recommend running the Memory test from the unRAID boot screen.