I have a 14 disk array. 1 drive for parity. I changed my parity drive to a 4TB drive a couple of days ago and that went fine. I rebuilt the parity and then did a parity check. Zero errors. All working good. I went to swap out disk#8 which was a 1TB for a 4TB drive but this is where things go wrong. The GUI doesn't come up.  So I log into the system via telnet and copy my SYSLOG out and view it, but for the life of me cannot find out what the heck is wrong. I see 8 mounted disks, so I'm missing 5 mounts. I see ALL the drives in the SYSLOG, but no GUI or all drive mounts. I've attached the syslog for assistance on this one. I usually can figure this out but I'm at a loss here. All the drives are in drive cages, so I re-seated all the drives multiple times all with the same result.

 

Also, when I type in REBOOT in the console the machine will not restart. I have to manually force it down.

Also, how do I try and RUN the native unraid GUI within the console so I can visually see the error.

Is it: emhttp &  ?

 

 

 

**Update**

 

So I took out the 4TB drive and put back the 1TB drive. The system came up with the GUI but not the array started. When I try to start the array it tells me this "the replacement disk must be as big or bigger then the original". Well I would have started a parity build at that point if I could but now I'm stuck in between drives not able to use either one of them. I can't access ANY of the mounts because of this. I'm really at a standstill on this one.

 

**Update**

I was able to run the array WITHOUT disk #8. Unprotected so I can access my data. So I went back and installed the 4TB drive and ran the array without it and it came up fine. Unraid does see the drive so I stopped the array, put the new drive back into disk #8 and then started the array. Lots of HD lights came on flashing all over for a minute then just one solid HD light on disk #8 (new drive) and the GUI still says "mounting disks". Since this is a larger drive, does it take this long to mount? I've included the SYSLOG from this session and I'm going to just leave it like this in case it just needs time, but from all experience it doesn't take more then a couple minutes to mount drives, unless unraid is doing some checks or scan or something. This log is named SYSLOG2.TXT and you'll see it trying to mount /mnt/disk8 but that's all that happens. It has been already 10 minutes. Sorry, had to zip the log, such a small size of upload size here.

 

Before I ever tried to use the new drive I tested it in a Windows server machine. Just to make sure the drive was ok before installing it in unraid. If I'm unable to use this drive, how would I go back to the original 1TB drive or a 2TB drive? Unraid won't let me use that size now because this drive is a 4TB. Even though it has never been mounted or worked, somehow unraid is forcing me to stay with this 4TB size now. Is there a way I can revert back to a 2TB or 1TB size? Even if I try to do it in the GUI, it tells me I cannot.

 

syslog.txt

syslog2.zip

From what I read, you didn't 'preclear' the new 4TB data drive.  If that is correct, then your system will sit in that state for several hours while unRAID writes zeros to every one of the 4TB.

 

If you had used Joe's preclear script, the disk would have been filled with zeros while you still had access to the array - then, when you swapped the new drive in to a data slot, it would have only taken a few minutes to 'format'.

From what I read, you didn't 'preclear' the new 4TB data drive.  If that is correct, then your system will sit in that state for several hours while unRAID writes zeros to every one of the 4TB.

In this case, that does not apply. Swapping a new drive into a currently occupied slot triggers a rebuild, where the data from all the other disks including parity is used to write the data from the removed drive onto the new replacement. Preclear and clearing is only used when introducing a disk into a whole new slot, with no existing data to rebuild. Theoretically adding a disk then does not put any data at risk during the process, where replacing a disk with a larger one leaves you unprotected from a second disk failure during the rebuild process. Disk replacement leverages the failure recovery mode of unraid to clone the contents to the new drive.

 

It is still a very good idea to run a couple preclear cycles on a new drive you are planning to use as a replacement, but not as an unraid prerequisite, only as a stress test to weed out early failure hard drives. Drive failure during a rebuild is a messy process to recover from, because unraid immediately records the new drive into the existing slot. If the OP had a saved copy of his super.dat file from before the upgrade attempt, he could just put the old drive back in along with the super.dat file that corresponded with that configuration.

 

Backing up the entire contents of your flash drive and saving a syslog before attempting any upgrades or drive changes is cheap insurance against this type of situation.

After waking up I'm looking at the system doing a rebuild on drive #8. All it ended up being was a timing issue. It just took a very long time for the drive to mount. I've added/replaced/rebuilt this server a couple times over and never remember having to wait such a long time for a new drive to mount. According to the log, it took about an hour for the drive to mount. Very strange indeed. I'm seeing errors in the log now which I'll either post in another thread or wait until the rebuild of drive #8 is over. Not sure if they are critical or not. Hard to tell.