Well I had my first drive failure yesterday - I would blast WD only to not jinx myself with the other 5 WD I have in the box right now - Touch Wood. OK on to the issues!!!

 

I noticed yesterday as I was renaming some files remotely on disk 3, windows my network stopped responding. I rebooted - same problem. went to IE and tower and noticed a red light on D6 and the machine was rebuilding data at 99.6% I waited 2 or more hours and it was still a that same place and then I could no longer access the server from IE (PC) so I shutdown the server rebooted let it parity check for 8 hours on a 500Gb drive. Then is ran fine all night green light on D6. Then I was renaming files on a another drive this afternoon and the same problem my network stopped responding and I went into IE browsing the server and this time the red light on D6 is back and the array is running, not taking any chances I hit stop on the array and the icon saying wait came up and that's all it would do. I tried to refresh would not find the server via it's name or the IP. I could telnet it once and no more.

 

So all that to say. I have shut down the server and removed the bad drive which I will replace in the morning but here is my concern, since I had to hard power off the machine without a proper shutdown sequence I know booting up that it is going to want to run a parity-check. Am I safe to install the new drive as is right now and boot up assign the drive let it format and parity check and rebuild data. I also wondered about doing the swap-disable method now that 1TB drives are pretty cost effective.

 

Any thoughts? I attached a .doc of my log at least what I could retrieve.

 

Thanks,

 

Dave

 

 

Help Please!!! Not sur ehow to continue. Here is the path. I shut down the server removed the failed disk. 2) removed the parity disk and installed it into the failed disks slot then I installed the new Larger drive into the old parity disk's slot and powered up the server. It will now not let me start the array - it is saying to many missing discs. I has read the manual over and over again. I thought that I was using the swap disable method as descibed in the manual.

 

Can I use the restore function to do what I need to do. which in the manual states that a swap-disable will upon power up copy the parity information to the new drive and then reconstruct the data on the old parity drive in the failed disk slot.

 

Am I not understanding something correctly here.

 

Help - I need this thing up and running.

 

Thanks,

 

Dave

Do NOT use "Restore" unless you want to erase all traces of what used to be on your old failed disk.

 

Restore is not well named.  It will "Store a new configuration based on the currently assigned (and working) drives" and then generate new parity based on only those drives.

 

Send an e-mail to Tom requesting help.  It is looking more and more like the parity swap feature does not work.

 

In the interim, put the parity drive back the way it was and if you have room, copy all data off the bad drive to others in your array.

 

Joe L.

Joe,

 

I have shut the machine down waiting for Tom's reply at this point. could I not put the parity drive back and the bad drive boot up the server. then shut it down and replace the bad drive with a new one, reconstruct the data and then shut it down and replace the parity drive.  I guess I will just have to wait and see what Tom says.

 

Regards,

 

Dave

Yes you can do as you described.

 

Then... If you have enough free space, you can copy everything from the bad drive to others, then shut down, un-assign it, use the "Restore" button (misnamed, in my opinion) to store a new config based on the currently assigned drives, and then re-calculate parity without the defective drive being in the array at all.

I don't think that the array's "desire" to run parity check due to the hard shutdown is the same as the array marking parity as invalid.

 

I would think that if you pulled the bad drive at this point and brought up the array, that unRAID would operate in its "missing drive" mode, and allow virtual access to the missing disk.  It should not try to run the parity check on an array that is short one drive. 

 

Then when you add the new drive and tell it to reconstruct, it should reconstruct onto the new disk.

 

After the rebuild is done, only then should the array be in a state where it could run a parity check - but at this point it would be unnecessary. UnRAID should be up and happy.

 

I am not saying this is the way it WILL work, just that this is the way I would EXPECT it to work from reading the documentation.

 

I agree with Joe L, reach out to Tom and confirm before taking the next step.  You only get to do it once.  Post back with what you learn!  This sounds like something everyone on this forum needs to know how to do with confidence.

 

- Brian

In the way I wanted to do it in the first post (swap-disable). Tom seems to think I stumbled onto a bug and he is working on a fix now. I am without my server currently and I am trying to get my XBMC up and running which involves renaming all my TV show files which are on the server.

 

I really do not want to go and get another drive I bought a wd 750 to replace the parity drive with yesterday and I did an advance warranty with WD whereas I upgraded to another 750 for 140$ in lieu of sending my 400Gb back they just void the serial number on the 400 when you do this. I would have to follow this routine without a bug fix for the swap disable.

 

re-install the parity drive to slot 1

reinstall the bad disk to slot 6

boot and start the array

stop the array and shut down

install a new 500 Gb in slot 6 to replace the bad drive because currently my parity is 500GB more $$$

boot up and start the array let it reconstruct the data on disk 6 waiting................................

stop the array shutdown

install the new 750Gb to replace the parity drive

boot up start the array rebuild parity on disk 1 waiting........................................

stop the array power down

replace the other 320gb 400gb drive I have with the old 500gb parity drive

boot start the array rebuild data on the larger drive and make more free space available waiting.............................

when completed USE server

 

anybody have any thoughts on the Samsung spinpoint drives I found 750gb drives in Ottawa, Canada for 147$.

 

Regards,

 

DVS

Wow - that's quite a task list!

 

Tom is awesome to support us through these problems.  You couldn't ask for more responsive support when catastrophe occurs.

 

I rebuilt parity on my 5.3Tb / 13 disk array yesteray.  Only one disk had an error - a new 750G WD drive.  The vast majority were Maxtor and Seagate drives - no problem with any of them.  In the future I will steer clear of WD!

 

I would recommend you buy a Seagate drive.  Fry had them on sale for $149 (see the good deals section of the forum).  Not sure if they still have them.  Do a search on the fry website for "Seagate" - I've found if you follow through the menus that often items are missing.  These are the drives with the 16M cache, not the new 32M cache.  I bought one as my parity drive and it works well.  People are reporting problems with the 32M ones anyway.

 

Good luck!

ALL Fixed with GREEN lights.

 

See this post:

 

http://lime-technology.com/forum/index.php?topic=1479.0

 

Regards,

 

DVS