My backup server, which was previously running 6b6, is not functioning with 6.0.1.

 

After unsuccessfully manually upgrading (similar result to what is described below), I created a clean memory stick and configured an array of a single disk (no parity, 1 data disk).

 

After assigning the disk to array slot 1, with no parity, the GUI appeared to hang.

 

I was able to click on another tab, and then click back on the main tab, and the disk was assigned to the slot with the characteristic blue icon.

 

I then tried to start the array (with a single disk assigned to slot 1) and the GUI immediately returned but the array did not start. The drive remained blue. The button to start the array was still available. Repeated attempts resulted in the same - the array will not start. The array status shows "Stopped. Initial Configuration.".

 

There is nothing peculiar in the log.

 

This server has 2 Areca ARC (RAID) controllers, which were working fine with 6b6. (I am running a different ARC controller in my main array without difficulty. I have no reason to think that the controllers are an issue except they are not a common component.)

 

Let me know if I can help diagnose. I am attaching my diagnostics file.

tower-diagnostics-20150706-1412.zip

My backup server, which was previously running 6b6, is not functioning with 6.0.1.

 

After unsuccessfully manually upgrading (similar result to what is described below), I created a clean memory stick and configured an array of a single disk (no parity, 1 data disk).

 

After assigning the disk to array slot 1, with no parity, the GUI appeared to hang.

 

I was able to click on another tab, and then click back on the main tab, and the disk was assigned to the slot with the characteristic blue icon.

 

I then tried to start the array (with a single disk assigned to slot 1) and the GUI immediately returned but the array did not start. The drive remained blue. The button to start the array was still available. Repeated attempts resulted in the same - the array will not start. The array status shows "Stopped. Initial Configuration.".

 

There is nothing peculiar in the log.

 

This server has 2 Areca ARC (RAID) controllers, which were working fine with 6b6. (I am running a different ARC controller in my main array without difficulty. I have no reason to think that the controllers are an issue except they are not a common component.)

 

Let me know if I can help diagnose. I am attaching my diagnostics file.

 

The issue appears to be related to the super.dat file on the flash device.  How old is the USB flash stick you're using for the boot device?  Here's the troubling line from the syslog:

 

Jul  6 14:12:27 Tower kernel: read_file: error 2 opening /boot/config/super.dat
Jul  6 14:12:27 Tower kernel: md: could not read superblock from /boot/config/super.dat

 

It's in there multiple times.  The other thing that's suspicious of your flash device is the fact that the last line reported in the log was the importing of the flash device.

 

I would try a few things.  First remove the flash device and plug into a Windows PC and run checkdisk.  Then reinsert to your server and boot up again, trying to do the same thing you did previously.  Let me know if that changes anything for you.

Those 2 lines are normal for a brand new system, or someone starting over with a clean install.  He doesn't have a super.dat file, so it's a clean install, starting fresh.  The line following that indicates it is initializing a new super.dat, and at that point it seems to completely restart, reason unknown.  Good idea recommending a chkdsk.

 

Only 3 possibilities I can think of are flash drive corruption (should be fixed by chkdsk), or a bug in unRAID here, or something non-standard in the environment that's not being handled correctly.  The only things I see that are non-standard are the multiple Areca cards, and the go file.  In the go file, udev rules are changed, an areca.awk file is run, and unknown packages are installed.  Can you tell us more about those 3 items?

Those 2 lines are normal for a brand new system, or someone starting over with a clean install.  He doesn't have a super.dat file, so it's a clean install, starting fresh.  The line following that indicates it is initializing a new super.dat, and at that point it seems to completely restart, reason unknown.  Good idea recommending a chkdsk.

 

As noted, the flash was freshly formatted and prepared. I was having very similar issues when I tried to upgrade the stick from 6b6 to 6.0.1, and decided to just reformat and begin afresh to rule out some errant remnant from an old version or beta from causing a problem. I am remote and not able to remove the flash, but can all but guarantee that the flash does not have corruption. I am able to write to the flash all day long from Samba and from an ssh session without any trouble.

 

Only 3 possibilities I can think of are flash drive corruption (should be fixed by chkdsk), or a bug in unRAID here, or something non-standard in the environment that's not being handled correctly.  The only things I see that are non-standard are the multiple Areca cards, and the go file.  In the go file, udev rules are changed, an areca.awk file is run, and unknown packages are installed.  Can you tell us more about those 3 items?

 

1. I answered above. The flash is freshly formatted and can be written without issue.

 

2. I am running an Areca controller in my main array, and it runs fine. The 2 controllers in this backup server have been running for months under 6.0 beta 6 without complaint.

 

3. areca.awk is an awk program I wrote that is able to relate the Linux configuration (/dev/sd?) to the Areca configuration (/dev/sg? and Areca slot numbers). It creates a small data file that myMain uses to run smart reports on Areca attached disks. It does absolutely nothing with unRAID. I run this exact program on my main array at boot.

 

The udev rules are more complex. See THIS POST from bubbaQ. These few pieces of magic result in the Areca drives models and serial numbers being properly reported to unRAID. Again, this is something I run on my main array, and has been run on this server under 6b6. The fact that the drives are properly named in unRAID is evidence that these instructions worked properly, and I highly doubt that this has anything to do with the problem.

 

The go file has instructions to load unmenu style packages - but there are no such packages on this system. Not even screen. My main server has this command as well, and it loads a number of packages without issue. Again, this flash is about as vanilla as it can be.

 

My thoughts.

 

I try not to overly influence anyone with my thoughts on the problem - but here they are.

 

1. The array does have quite a large number of devices. There is the flash (sda), and then 26 hard drives (sdb-sdaa). The array is only being configured with a single disk but with recent licensing changes to cover attached devices vs array devices, I wonder if the sheer number of devices might somehow be interfering. This is my current best theory.

 

2. The smart folder in the diagnostics file does not have smart reports on all of the drives. This is a bit curious, but think it is an issue with the diagnostics routine. Pulling real smart reports from the Areca card is tricky but doable, and I am quite sure that the normal GUI does not have that logic included. But I believe you can pull a smart report in the traditional way. What it returns is not a normal smart report - but it is something similar. Anyway, this is a mystery and exploring it may point the way to a root cause.

Thanks, Brian.  I'm about out of ideas.  Sometimes, errors only go to the console, not the log.  You might Telnet/PuTTY in and watch it through that start-array/restart cycle, see if anything unusual pops up.

I will see if anything shows on the server console, and investigate removing a drive or two and see if that will allow the array to start.

Nothing on the console.

 

I removed one drive and array started. There is a problem if physical drives in the server go above 25, regardless of number of drives assigned to the array.

 

Array will not start with more than 25 hard drives.

Nothing on the console.

 

I removed one drive and array started. There is a problem if physical drives in the server go above 25, regardless of number of drives assigned to the array.

 

Array will not start with more than 25 hard drives.

i am sure I remember a post stating that the limitation to 25 devices being in the machine had been removed for the Pro license, although there was still a limit of 25 devices being used by unRAID..  However I guess there could still be lingering issues if such large configurations have not had much testing.

I remember the same discussion ... and the same outcome (i.e. no limits imposed by Pro).

 

Even more interesting is that this was working fine with Beta 6 ["... which was previously running 6b6, is not functioning with 6.0.1 "]    Brian => were there any configuration differences between what was working in Beta 6 and the final version?

 

Backup server is pretty vanilla. But with 6b6 I loaded a few small packages like screen that are not yet installed in 6.0.1.

 

I didn't mention but GUI is still struggling with the drive count. Seems to hang when assigning a slot. Have to click on another tab and then back on the main tab to get it to register. Probably took a full minute to assign one disk and get the gui responsive again.

Backup server is pretty vanilla. But with 6b6 I loaded a few small packages like screen that are not yet installed in 6.0.1.

 

Actually I meant hardware config differences ... i.e. same number of connected devices; same controllers; etc.    In particular, did it work fine with all of the drives you HAD connected before you moved to 6.0.1 and had to remove a drive?

 

Actually I meant hardware config differences ... i.e. same number of connected devices; same controllers; etc.    In particular, did it work fine with all of the drives you HAD connected before you moved to 6.0.1 and had to remove a drive?

 

I took a running 6b6 system (non-parity protected backup array), upgraded the flash to 6.0.1, booted, and tried to start the array. It would not start.

 

So I rebuilt the memory stick to 6.0.1 vanilla, booted, defined a one drive array, and it would not start.

 

So I went into one of the RAID card configurations and removed a volume (I didn't actually physically remove a drive). Rebooted, defined my 1 drive array, and it started.

Interesting.  Clearly the code that was added r.e. connected devices isn't working as planned, since LT said there wouldn't be any restrictions on Pro.  [i'm pretty sure that was all added well after Beta 6.]

 

 

 

 

Interesting.  Clearly the code that was added r.e. connected devices isn't working as planned, since LT said there wouldn't be any restrictions on Pro.  [i'm pretty sure that was all added well after Beta 6.]

 

It's not starting because attached devices (not counting flash) > 25.  Not supposed to happen with Pro, will fix in next release, sorry about that.

It's not starting because attached devices (not counting flash) > 25.  Not supposed to happen with Pro, will fix in next release, sorry about that.

 

Thanks Tom! If you need help testing this fix out, let me know.