My machine is stuck at "...Sync filesystems..."

Steps taken to get me here:

1) disk #2 failed

2) shutdown the system, removed failed disk, installed to new disks

3) brought system back up (degraded mode)

4) started preclear on one disk

5) started preclear on second disk (10 hrs later)

6) finished preclear successful - second disk zeroing 88%

7) issued stop server command (so I could start rebuilding disk 2)

8 ) stop proceeding: "Stopping Docker...stopping libvirt...Stop HVAHI...Stop NFS...Stop SMB...Spinning up all drives...Sync filesystems..."

 

It's been like this for about 10 minutes now.  Please advise!

 

thanks

Everend

If it really won't shut down your only option is to force it - hold the power button in until it powers off or unplug the power.

 

Edit: I assume you've tried telnetting in?

 

Personally, I would not have wasted time pre-clearing disks while the array is degraded - I would get it rebuilding as soon as possible, onto a virgin disk if that's all I had to hand.

 

When you power it up again grab the diagnostics and see if you've suffered a failure of something else in the meantime.

Well I can access the system CLI.  I have a monitor and keyboard connected to it.  I can run commands there, I just don't know what to run.  I suspect I need to run a kill pre-clear command or something to let the stop proceed.  I don't think I'm at the button-hard-reset stage yet.

 

I did the preclear based on advise I read on this forum. I saw several people suggest it.  I minimized any other use of the system while degraded.

Oh it just stopped... took about 40 minutes but it stopped.

The preclear is still running.  It just finished zeroing and that's when it let the system finish stopping.    Now that I have the UI back, I've stopped the post-read phase of preclear on the second disk so I can start the disk 2 rebuild now. 

 

Once the system is 100% I'll restart the preclear.  I expect I can skip the initial pre-read, right?

 

thanks

So the pre-clear script was what prevented the server from shutting down? From the way you described it I thought it had locked up. I'm glad you managed to shut it down cleanly but I would ask you to think about what you consider is the reason for pre-clearing disks.

 

As I see it, the primary reason is to set the contents of the disk to all zeros (plus a special signature in the MBR that identifies a clear disk to unRAID) so that a new disk can be added to the array without having to recalculate parity.

 

Because a pre-clear cycle first reads the entire disk, then writes the entire disk and then reads the entire disk again it is seen as a useful test of a new disk and is quite successful at weeding out early failures, but that's its secondary purpose.

 

When you replace a failed disk the entire contents of that disk are calculated from the other disks in the array and written to the replacement, so it doesn't matter what the original contents of the replacement disk are. In other words, the replacement disk does not need to be pre-cleared.

 

If you happen to have a spare disk sitting on a shelf you might want to run a pre-clear cycle or two on it in order to test it but if you had to dash out to the store and buy a new one it's a matter of debate as to whether it's better to set the rebuild going as soon as you get home or whether you want to leave the array unprotected while you test the disk.

 

I keep spare, tested disks in case of failures and because I'm always adding or building but if I was to find myself in the situation of needing a replacement but only having virgin disks I would not waste time pre-clearing one of them. I would run the SMART short self-test (2 minutes, typically) and then get the array rebuilding. Once it had finished and I was happy I'd run the SMART long self-test (as much as 10 hours or more) and keep a closer than usual eye on the SMART status of the disk for a while.

 

You may totally disagree with me and not want an untested disk anywhere near your array and that's perfectly fine, as long as you've reached that decision yourself and are not just blindly following what other people say.

 

  Thanks for the full explanation.  The main reason I wanted to do the test first since I didn't have a lot of confidence on the new drive from Amazon.  It got good reviews but it was very similar model as the drive that failed.  I had ordered two new drives only a few hours before this one failed.

  Thanks for the full explanation.  The main reason I wanted to do the test first since I didn't have a lot of confidence on the new drive from Amazon.  It got good reviews but it was very similar model as the drive that failed.  I had ordered two new drives only a few hours before this one failed.

 

Not 100% sure but think the preclear script will start from the beginning...