Running 5 beta 12.  I had a drive crap out a few days ago.  I got a new drive, ran a triple preclear on it (no errors), added it to the array, and started the rebuild process.  Old drive was 2TB, new one is 3TB.  The last time I checked up on it, it had been running for about 7 hours and was at ~60% done.  Just checked again and the webUI is no longer responsive.  I can still get in through the terminal.  Is there any way to either re-initialize the webUI or check on the progress of the rebuild through the command line?  I don't dare reboot until the rebuild is done - if it's even still rebuilding.

I'd wait until there is no disk activity indicated by the LEDs -- then press the power button momentarily (this should cause a graceful shutdown).

 

When you get the system rebooted, upgrade to v5.0 final => you're running a pretty old Beta version, and there's no reason to even attempt to troubleshoot issues with a version that old.

 

Unfortunately, my drive cages don't have activity lights.  Any other way to tell?

If there are no activity lights, all you can do is wait a long-enough time that you're confident that IF it finished okay, it had time to complete.    Given 60% in 7 hours; and factoring in that the inner cylinders are much slower than the outer ones, so the last 40% will take appreciably longer;  I'd guess that you needed about an additional 6-7 hours to finish.

 

To be completely "paranoid" about enough time, I'd wait a full 24 hours -- then just shut the system down via the power switch [hopefully a momentary press will invoke a clean shutdown;  if not, you'll need to hold the power switch for a few seconds to force the system off].    Then reboot the server, and see if the new drive is "green" (i.e. the rebuild finished okay).    If not, you'll need to redo the rebuild.    If so, then I'd upgrade to v5.0

 

Running 5 beta 12.  I had a drive crap out a few days ago.  I got a new drive, ran a triple preclear on it (no errors), added it to the array, and started the rebuild process.  Old drive was 2TB, new one is 3TB.  The last time I checked up on it, it had been running for about 7 hours and was at ~60% done.  Just checked again and the webUI is no longer responsive.  I can still get in through the terminal.  Is there any way to either re-initialize the webUI or check on the progress of the rebuild through the command line?  I don't dare reboot until the rebuild is done - if it's even still rebuilding.

 

from the command line type:

/root/mdcmd status

 

somewhere in its output is the current status and progress.

I believe I found the pertinent section of the mdcmd output.  Unfortunately, it doesn't seem to be doing anything.  Here's what I had at 8:30 this morning:

 

mdResync=2930266532
mdResyncCorr=1
mdResyncPos=1891383836
mdResyncDt=66053
mdResyncDb=1689656

 

And here's what I had at 7:30 this evening:

 

mdResync=2930266532
mdResyncCorr=1
mdResyncPos=1891383836
mdResyncDt=107229
mdResyncDb=1689656

 

I'm guessing mdResync is the total parity size (3TB is what I have) and Pos is the current position (1.8TB sounds about where it crapped out).  If this is correct, the Pos hasn't moved in nearly 12 hours.  The only thing incrementing is Dt, and I'm not sure what that is.  If it was working, it should definitely be done by now.  I've attached the full output taken just now, in case that helps.

 

Any suggestions?

status.txt

Taken this morning:

 

mdResync=2930266532
mdResyncCorr=1
mdResyncPos=1891383836
mdResyncDt=157913
mdResyncDb=1689656

 

After more than two days, I think it's safe to assume the parity sync is never going to complete.  What is the safest way to shut it down?

I think you're interpreting that correctly => i.e. it's "stuck" and isn't going to finish.

 

I'd press the power switch momentarily, which MAY allow a clean shutdown -- but this may also be "hung".  If it's not off after about ten minutes, then just hold the power button until it power off (~ 5 seconds).

 

When you reboot the system, it will likely restart the parity sync.  Since at this point you don't have valid parity anyway; I'd stop that;  upgrade to v5.0; and THEN do a parity sync.

 

Momentary press did nothing, so I did a hard power-down.  I'm going to give it another chance to do a rebuild as-is before I try upgrading.  The upgrade instructions advise to only do an upgrade with a healthy array, and mine's not healthy at the moment.  If it craps out again, then I'll take the chance with an upgrade to final.

Absolutely right.  For some reason I was thinking you were upgrading for 2TB to 3TB parity -- so were in a position where you didn't have valid parity.

 

But that's NOT the case => so you definitely want to get the drive rebuilt before you do the upgrade.

Think about installing unMENU before the upcoming rebuild. The reason is if webGUI (emhttp) fails again you have another method to connect through a GUI. Also you get (I think) the graceful powerdown script.

I have unmenu installed now.  When the default UI crashed, I tried unmenu, but that wasn't working either.

Even if UnMenu responded, it wouldn't have mattered, since the rebuild was "stuck".

 

Rebuild completed this time.  All looks good.  Running a parity check.  If no problems crop up, I'll upgrade to final.  Thanks for the help, guys.

WebUI/unmenu crapped out again during parity check.  Since the array is as healthy as it's going to get at this point, I upgraded to 5.0 final.  Running parity again with fingers crossed.

Attach a syslog.

It hasn't failed since moving to final, though that was only a couple hours ago.  If it locks up again, I will.

Exactly what I already mentioned.

This stock GUI simply sucks!

 

Not using plugins is NOT a viable option because stock unRAID is

also missing important features.