version: 5.0-rc5

 

I pulled a drive (after emptying) and performed a "new config" via the gui, reordering the drives and correctly setting up the parity.  A full parity was completed, as well as a parity check.  No issues with parity check or accessing any of my data.

 

I also went into the shares (which where auto recreated when the array came online) and updated the disks since I wanted to change the order.

 

Issue is, it appears to be ignoring that data and writing to the wrong drive.

 

For example, I have a share for my machine backup that uses M$ synctoy to backup my user directory.  I have it assigned to disk 1, excluding disk 2,3.  It's currently writing to both disk 1 AND disk 3.  As well, when browsing to \\unraid\sharename, I see the data from both disk 1 and 3, further proving it's using both drives for an unknown reason.

 

Any ideas?  Seen this before?

most people use "disk1,disk3" and not "1,3"

 

I have no idea if the numbers alone will work.  I never used include/exclude on disks.

 

I also have no idea when the values are read.  Many times you need to stop and re-start the array for them to become effective.  (the new values are only perhaps only read when the array is first starting)

Let me try that.  Up until I did the new array config, I had four drives in and data was going to the correct location.  I have also stopped and started the array, however have not rebooted (running a preclear on a drive that will likely choak hard given the major decrease in speed from test 1 to test 3).

Let me try that.  Up until I did the new array config, I had four drives in and data was going to the correct location.  I have also stopped and started the array, however have not rebooted (running a preclear on a drive that will likely choak hard given the major decrease in speed from test 1 to test 3).

See this example in the wiki from Tom's release notes;

http://lime-technology.com/wiki/index.php/Release_Notes#Changes_from_4.2-beta3_to_4.2-beta4

Let me try that.  Up until I did the new array config, I had four drives in and data was going to the correct location.  I have also stopped and started the array, however have not rebooted (running a preclear on a drive that will likely choak hard given the major decrease in speed from test 1 to test 3).

See this example in the wiki from Tom's release notes;

http://lime-technology.com/wiki/index.php/Release_Notes#Changes_from_4.2-beta3_to_4.2-beta4

 

Ok, using the full "disk1" syntax worked, no array restart or reboot required.  Thank you.

 

However, two things strike me as odd:

1) No error check or notice when invalid text is entered?  Very bad idea IMO.

2) Somehow with 5 shares and 4 disks with different settings (2 shares were 1 disk, 1 spanned 3, 1 spanned 2, 1 spanned 1) it seemed to work, but not after the array was rebuilt and drive order changed?

 

Not doubting the full "diskx" should be used if that's in the release notes, but odd...

Ok, using the full "disk1" syntax worked, no array restart or reboot required.  Thank you.

/quote]Great.quote]

However, two things strike me as odd:

1) No error check or notice when invalid text is entered?  Very bad idea IMO.

I agree.  The expected format should be enforced/described.

Not doubting the full "diskx" should be used if that's in the release notes, but odd...

true.

If the share exists on an excluded drive, directory listings will still include the files on that drive. It's my understanding that writing to the share will exclude the specified excluded drives with the exception that if you have a split level and the split directory is on the excluded drive, it will still write to the excluded drive.