Hi all,

so I have been copying my movie collection onto my newly configured unRAID box. I decided to upgrade the cpu, this went fine and all booted up with no issues. I then noticed that when copying files to the unRAID server everything grinds to a halt. I notice one of the drives is showing a lot of errors (20k for running a little over an hour). So I am thinking that the server falling over maybe due to this drive creating so many errors.

 

My next step is to replace the drive which is producing the errors but when I shut down the system and reboot with the new drive installed the array will not restart due to the missing drive. It will not allow me to do anything else.

 

In the Wiki it simply says to shut down the server, replace the drive and in one place it said I should see a blue dot beside the new drive. None of this is happening.

 

Any ideas on:

1. The server slowing to a crawl and becoming unreachable.

 

2. How to replace the drive? Following the Wiki instructions do not seem to be working.

 

Help!!!

So having slept on this overnight, I checked the server before heading to work and it is still up and no errors showing for any drives.

 

Being that this is a brand new setup, the parity has not completed, would this process be the issue for the machine being so slow? I have copied close to 1TB of data to the server, should I allow the parity to complete or does that have no impact on the server usage?

 

My other thought overnight is that the PSU might not be up to the task, I have an Antex VP450F 450w psu with an AMD A10 -6800 cpu, 8 gig of ram and 3 hard drives (2 Seagate 3TB and a WD Green 4TB parity disk) . The MB is an MSI FM2-A75MA-E35.

 

I just bought a pro key for unRAID and so I really want to get this working.

 

Help!!!

If you want data protection, better let the parity calculation complete first, it might take a while but it is worth the wait.

 

Well the system is brand new and so I could easily just wipe everything and move forward if I could confirm what the issue is with my setup.

 

I have been searching online for compatibility issues with the cpu upgrade and the main board and all I am finding is that the original version of the board needed a bios update to work correctly with the cpu, so I will check this when I get home to make sure the board's bios is the correct version for the new cpu (although I believe both parts are Richland and before the cpu upgrade everything worked fine. I did clear the CMOS when I upgraded the cpu to avoid any issues with old settings)

 

 

Well the system is brand new and so I could easily just wipe everything and move forward if I could confirm what the issue is with my setup.

 

I have been searching online for compatibility issues with the cpu upgrade and the main board and all I am finding is that the original version of the board needed a bios update to work correctly with the cpu, so I will check this when I get home to make sure the board's bios is the correct version for the new cpu (although I believe both parts are Richland and before the cpu upgrade everything worked fine. I did clear the CMOS when I upgraded the cpu to avoid any issues with old settings)

 

Your CPU can't be the bottleneck, it is more than capable. Though, you are right that some motherboards need a BIOS upgrade to support the Richland version.

 

Check your sata & power cabling, it is likely more in that direction.

Ok then I think I will pick up some new sata cables and possibly a PSU in case I need to swap that too

So looking through the forums and checking on my hardware it seems my PSU is a 2 Rail design, could this cause the type of issue I am seeing? I have all the hard drives plugged into the same cable coming from the psu, the second cable has nothing attached to it. Would a low output not just cause the system to crash or is this just a Windows OS symptom and would be handled differently in a Linux distro like unRAID?

I picked up a Corsair TX650M, so that should eliminate any psu incompatibilities plus I picked up 3 new sata cables to replace the older ones I had (I already had 2 brand new ones which I will leave in the system).

 

I also got my license so I should be able to add a cache drive and an additional 4tb drive I picked up over the weekend.

In the Wiki it simply says to shut down the server, replace the drive and in one place it said I should see a blue dot beside the new drive. None of this is happening.

The wiki assumes your server already has a valid parity drive in place. Until you complete the parity calculations, you can just set a new configuration and assign the drives however you wish. Once parity is valid, adding a new drive requires it to be cleared before it can be added to the protected drives.

 

Unless you are in a real hurry to get your server up and running, I strongly suggest you search for and use the preclear script written by Joe L. to confirm your new drives are error free. It will erase all the data on any drive you run it on, so you'll have to redo populating your array, and once again wait for parity to build. However, it's well worth it to ensure that your drives don't suffer from infant mortality.

So I searched and found the info on pre clear. I will most likely run it once I get the server running properly. My thought at that point would be to simply format the drives I have, run pre clear and then create a new array and copy my data onto the new fresh setup.

Everything is currently on other media so there is nothing mission critical that I can lose.

There is no need to format the drives first. Simply preclear them, then unRaid will format them when they are added to the array.

I thought I read in the thread on preclear that it will not allow it to run on a drive assigned to an array?

I thought I read in the thread on preclear that it will not allow it to run on a drive assigned to an array?

Correct. Formatting has nothing to do with whether it's assigned to the array. Set a new configuration, and don't assign any drives. You can then preclear your drives, no need to format them first.

So when adding a drive in version 5 it gives the option to clear the drive. Is this the same as pre clear or something different?

If the drive is not already precleared, unRaid will preclear and then format the drive.  The array is unavailable during this time.

 

I'm not familiar though with the option to clear that you mention, but then I've never added a non-precleared drive to the array.

When I added the drive to the array it prompted me to clear the drive (the drive is brand new 4tb). It is currently at 11% of the clear process after approximately an hour of processing and the array is not available while it is processing.

The drive was not pre-cleared so unRAID is clearing the drive. The system is unavailable during clearing.

Is this the same as pre-clearing the drive as indicated earlier in this thread?

Is this the same as pre-clearing the drive as indicated earlier in this thread?

Same end result, but not the same function. Preclear checks the smart status of the drive, exercises it and confirms that it reads back correctly what is written, then checks the smart status of the drive again and notes any changes. This is all done with the drive not assigned to the array, so you still have access to the other array drives while it works. After a drive is successfully precleared, you can then add it to the array and it will not have to go through the lengthy clearing process that keeps you from accessing the rest of the array.

 

The stock clearing stage only writes zeros to the drive, it doesn't confirm that the drive is reading back correctly. It assumes the drive is good.

Perfect, exactly what I needed to know.

One last question, if I stopped the system pre clear, took the drive out of the raid and then ran the manual pre check which checks the drive, would this be ok or would stopping the system pre clear at 45% make the drive unusable?

One last question, if I stopped the system pre clear, took the drive out of the raid and then ran the manual pre check which checks the drive, would this be ok or would stopping the system pre clear at 45% make the drive unusable?

If I understand you correctly, yes, you can stop the stock unraid clearing process, unassign the drive, then run Joe L.'s preclear script on it. I'm not sure what the unraid interface will show at that point, worst case scenario is that you will need to set a new config, and reassign all the drives and rebuild parity once again. I'm assuming you really haven't started moving data for real at this point, if you have, you will have to start over again if you preclear all your drives. (Which you really should, just for the sake of gaining confidence in the drives.)

Well I left the pre clear run over night and it completed and unRAID was ready to format the drive as I was leaving home. Does the JoeL pre clear script take as long to run as the unRAID pre clear process? I would assume so, but the upside would be that the array could be still running due to the drive running the pre clear script would not be part of the array.

Does the JoeL pre clear script take as long to run as the unRAID pre clear process? I would assume so, but the upside would be that the array could be still running due to the drive running the pre clear script would not be part of the array.

It actually takes a good bit longer, because the script actually checks the drive fully. But, as you said, the array is fully available, so it is a good tradeoff.

The pre-clear script takes roughly 4 times as long as a clear, as it does a pre-read, then the clear, then a post-read.  The pre-read takes ~ the same time as the clear; the post-read takes about twice as long.  But as noted above, the advantage of using JoeL's script is that the array is fully available during the process ... and the clearing operation is fully validated, which is not the case when the array does it natively.