Lost power to my unraid box and it is wanting it to rebuild parity.  I'm just trying to get about 4TB of data up there - so instead of waiting for it to rebuild I figured out I could just unassign the parity drive and (hopefully) the copies would go faster.  Unfortunately - the user shares don't seem to be working.  Is this due to losing the parity - or is it something else?

 

Edit:  Downgrading from 4.5-beta6 to 4.4.2 seemed fixed this. 

Lost power to my unraid box and it is wanting it to rebuild parity.

When an unRAID server does not shut down properly, such as after a power outage, it is very normal and very good that it wants to finish all writes to the drives (replay transactions) and run a parity check, to ensure there is no damage to the file systems and parity information.  I would like to respectfully suggest that what you saw was not the start of a parity build, but a parity check.  Now you are right that a parity operation would greatly slow down any copying while it is in progress, so you could if you want, abort the parity check after a few percent, do the copying, then run a full parity check later.  I don't recommend that though, because it is better to be sure that parity is 100% correct, if a drive were to fail during your copying.

 

When a parity check is started on boot, because of a bad shutdown, it starts *before* the drives and shares are fully mounted, so the mounting of the drives and the transaction replay on most of them will take much longer than usual.  The creation of the User Shares will start much later therefore, and take a lot longer before they too are available.  So I suspect that what you experienced was just the long delays in the drives and shares becoming fully available, due to the very good and very important steps being taken by the kernel and unRAID modules, to restore full data integrity.

 

May I respectfully say that, as far as I know, there is no known reason why v4.4.2 would be better than v4.5-beta6, and in fact, I believe most users would say the opposite or that they are essentially equivalent.  I think what has happened is that the first boot after the power outage was VERY slow due to the reasons given above, which finished replaying incomplete transactions and parity checked the first few percent of the drives, but had not finished setting up the shares at the moment when you checked them.  The subsequent boot with any version would then work fine.  But you really need to do a full parity check still.

 

If you were to unassign the parity drive and make changes to any data drive, you WOULD have to rebuild parity later, but yes the copying would be much faster.

RobJ - thank you for your reply.  You may very well be correct.  I am really trying to get about 5TB of data moved to the lime box for a redundant backup - I've been trying to for a few weeks now.  At this point - I'm just ready to get the data over there then do a parity build.  If a drive dies in the meantime, it isn't going to be a big deal I'll just have to stick another one in and re-sync that part of the backup.

 

As soon as my data is all moved over I will rebuild the parity.  When I'm finished I'll confirm your diagnostic that this was not a bug but just a matter of me not waiting long enough.  It *seems* like the 4.4.2 loads quite a bit faster - so you're probably right, it just hadn't built the shares yet.