[Solved] - Shares gone - possibly screwed up bigtime!

[ccruzen]

So here's the story,

I got up this morning and my server was unresponsive, could not access the gui and when I telneted (sp?) in with PuTTY and logged into the console I could not get it to shutdown.  I did a hard reset and when I turned it back on the array wasn't started but said configuration valid.  So, probably my first mistake was that I just started the array.  Things looked good so I went to work.  I logged in from work and saw that sabnzb+ hadn't started, checked unMenu and it showed uninstalled, in a rush I re-installed, started it up, it looked to be working but then I checked Sickbeard and it was running but none of my shows/config was there.  After a few minutes, the server became unresponsive again, no gui, no unMenu, telnet not working.  Got home, hard reset the thing again, again, it comes up configuration valid so I start it again (I know, I'm an idiot).  Now I just realized that my shares are no longer there at all, the gui is working fine but my shares are no longer there at all, I can't access the disk shares through smb or MC.  Now I'm worried, so I check the disk1 settings, it's showing MBR: unaligned, there are a few other disks showing that as well.  I'm stumped and also a little worried.  I've attached the current syslog if that helps.  Server stats are:

Case: Norco 4224

PSU: Corsair TX750

MB: BIOSTAR TH55 HD

CPU: Intel Core i3 550

RAM: 4GB Kingston

 

12 disks

using a cache drive.

 

New Info:

Parity check was in progress, it just stopped and red balled Disk 6.

Syslog is from after the drive redballed.

 

Any help very appreciated.

 

Edit:

Sorry, unraid Beta 12a

syslog2.zip

[lionelhutz]

There is a script unraid_partition_disk.sh that can be used to fix the MBR to aligned. I believe you run it with a -A. Search for it and you should find more info. I'm assuming you had the drives 4k-aligned so that's why you think them being unaligned is a problem.

 

Looking at the syslog, it seems the SAS card acts up a bunch before the drive failures so the drive might be OK. You should pull a SMART report and test it on another port to see if it can be read.

 

Peter

[ccruzen]

Would that help return my files?

Just want to make sure not to do any more damage if I can help it.

[Joe L.]

Would that help return my files?

Just want to make sure not to do any more damage if I can help it.

It will not do any damage.   The unraid_partition_disk.sh command might just set the partitioning in the MBR to point to the correct starting sector, which will then let the disk mount correctly, and then you can get to the files.

 

Worst you can do is point the MBR sector start to sector 63 when the file-system actually starts on sector 64, or to sector 64 when it actually starts on sector 63.  If so, the drive will not mount, and although the files are not affected, they cannot be reached.   to repair, you just run the command again, with (or without) the "-A" option to fix it correctly.

 

There is a bug in unRAID 4.7 where it is overwriting the partitioning in the MBR, and in some cases it ends up pointing a partition to the wrong starting sector, and the disk shows as unformatted since it cannot be mounted.

 

Joe L.

[ccruzen]

Thanks for the suggestions, on a whim, I got it to do a clean shutdown and restarted the server.  Everything is working as normal, the only difference is Drive 6 is still red-balled.  I'm going to try to run a smart report and then try it in a different slot.  Thanks again.

[Joe L.]

Thanks for the suggestions, on a whim, I got it to do a clean shutdown and restarted the server.  Everything is working as normal, the only difference is Drive 6 is still red-balled.  I'm going to try to run a smart report and then try it in a different slot.  Thanks again.

It is "red" because a write to it failed.  A different slot might fix an intermittent connection, but it will not magically get it to not be "red"

 

See the wiki on how to proceed to get a working drive back into the array once you fix a bad connection to it.

 

(Hint, you must un-assign it, start the array with it un-assigned, re-assign it, and then re-construct it.)

[ccruzen]

Thanks for the suggestions, on a whim, I got it to do a clean shutdown and restarted the server.  Everything is working as normal, the only difference is Drive 6 is still red-balled.  I'm going to try to run a smart report and then try it in a different slot.  Thanks again.

It is "red" because a write to it failed.  A different slot might fix an intermittent connection, but it will not magically get it to not be "red"

 

See the wiki on how to proceed to get a working drive back into the array once you fix a bad connection to it.

 

(Hint, you must un-assign it, start the array with it un-assigned, re-assign it, and then re-construct it.)

I understand, thanks for the help.