Ok, I'm a unix veteran and I'm quite comfy with the command line.

 

I'm puzzled why the proper use of shutdown results in a parity check with each startup of the array. Is it necessary to stop the array before shutdown to prevent the parity check. If so; then it really needs to be added to the init routines so that it happens automagically. I can do this myself on my machine; but is that the correct procedure?

 

Cheers,

 

 

Hi John,

 

When you click 'Stop' to stop the array it will stop the samba server, flush the block cache, unmount all the data disks, patch the smb.conf file (to remove the disk shares), restart samba, then stop the driver which clears a 'dirty' bit in the 'config/super.dat' file (indicating clean stop).  You can then use a shell command to power off, or whatever.

 

The 'shutdown' shell command can theoretically do all this - just not a big priority to add the necessary code at the present.

 

When the array is 'Started' it checks the 'clean stop' bit in the 'config/super.dat' file - if not set, will initiate parity-sync.

Thanks Tom

 

I can do the work I need from here.

 

Best,

 

 

I've recently had a case where the network of the unRAID box stopped working (for whatever reason). So I tried to shut the box down cleanly by attaching a keyboard and, logging in as root and typing "shutdown 0". But it didn't work. The unRAID box stopped some processes, but stayed in Linux. This generally doesn't seem to work properly. Also pressing the power button on the unRAID case doesn't seem to work correctly for me, either. Pressing the "shutdown" box in the web interface works fine, though.

 

Well, in the end I had to do a hard reset. unRAID consequently started a parity sync at next boot.

 

I'm posting this because I think it would be a good thing if there was a way to properly shut the unRAID server down when network is lost. As far as I understand, when the network is lost I've currently no other choice than to shut the server down in a way that destroys the parity protection. Isn't this a noteworthy hole in the unRAID protection?

 

Btw, what happens if there's a power loss and a harddisk fails because of the power loss? Due to the dirty bit in the super.dat file the parity is considered to be invalid, right? Can the failed harddisk even be restored in this situation?