June 27, 201412 yr Normally, Im VERY happy with the speeds of my unRAID server. Parity checks and drive rebuilds consistently run 110MB/s+. However, when using MC or mv to move files around, I get less than 25MB/s. I would understand if these were a bunch of small files, but right now, I'm moving 800MB+ MKVs. I get the same speed if I use my Windows workstation to move the files. There is NO speed advantage to doing it on the server itself. Seems rather strange to me. Any ideas?
June 27, 201412 yr remember that its writing/updating parity at the same time, so will be slowed down by this 'double write' operation - even when doing cross disk moves/copies via the server console itself. could look into using a cache drive if better transfer speed is required - but all it effectively does is defer the parity protected copy
June 27, 201412 yr Author remember that its writing/updating parity at the same time, so will be slowed down by this 'double write' operation - even when doing cross disk moves/copies via the server console itself. could look into using a cache drive if better transfer speed is required - but all it effectively does is defer the parity protected copy Yeah, I forgot about re-calculating parity. I have a cache drive, but its only a 180GB SSD. It would take too long to do the double-copy with cache. When I upgraded to v6, I forgot to set my split-levels. Things are a mess and Im trying to clean it up.
June 27, 201412 yr ah sugar - and if the data is already on the array, just in the wrong place, the cache drive wont really help you.. if you just do the moves using /mnt/disk* paths (not using /mnt/user/* paths), the moves should be quicker (potentially instant) if the moves are within the same disk.. which, probably isnt the case for a split level problem another option, depending on how desperate you are, is disable parity for the duration of the moves, and recalculate a new one once done - but Im not sure I'd really recommend that
June 27, 201412 yr Author ah sugar - and if the data is already on the array, just in the wrong place, the cache drive wont really help you.. if you just do the moves using /mnt/disk* paths (not using /mnt/user/* paths), the moves should be quicker (potentially instant) if the moves are within the same disk.. which, probably isnt the case for a split level problem another option, depending on how desperate you are, is disable parity for the duration of the moves, and recalculate a new one once done - but Im not sure I'd really recommend that Yeah...I'm not disabling parity! I am using /diskX paths. I'm in the middle of pre-clearing four 4TB drives, then replacing the remaining 2TB drives in my array. That process is going to take at least a week or two I figure. I guess I'm just going to have to be patient.
June 28, 201412 yr I just went through a file reorganization on 3 servers myself. I did use a cache drive. I copied from /mnt/user/<share> to /mnt/cache/<share> so that all files were continuously available and could do it with one mc copy operation to the cache and then one to put it back to a disk. Rather than /mnt/disk2/<share> & /mnt/disk4/<share> & /mnt/disk12/<share> all to /mnt/disk8/<share> which would take three or more copy operations if the directories were split up on 3+ drives like I had mine. I did it this way so ensure I could always access the files and to move the files with little attention by me. Took me several months to move about 100+TBs of data but with a 2+TB cache drives I could queue up a 2-6 hour copy operations to the cache drive and 6-12 hour copy to move it back to the array and let it move while I was asleep or at work.
June 28, 201412 yr Author I just went through a file reorganization on 3 servers myself. I did use a cache drive. I copied from /mnt/user/<share> to /mnt/cache/<share> so that all files were continuously available and could do it with one mc copy operation to the cache and then one to put it back to a disk. Rather than /mnt/disk2/<share> & /mnt/disk4/<share> & /mnt/disk12/<share> all to /mnt/disk8/<share> which would take three or more copy operations if the directories were split up on 3+ drives like I had mine. I did it this way so ensure I could always access the files and to move the files with little attention by me. Took me several months to move about 100+TBs of data but with a 2+TB cache drives I could queue up a 2-6 hour copy operations to the cache drive and 6-12 hour copy to move it back to the array and let it move while I was asleep or at work. Thanks! Maybe I should replace my cache drive temporarily. I have an extra 4TB I could use. Something to think about. I wish I could have TWO cache drives! I really like running Plex on an SSD.
June 28, 201412 yr Thanks! Maybe I should replace my cache drive temporarily. I have an extra 4TB I could use. Something to think about. I wish I could have TWO cache drives! I really like running Plex on an SSD. Mount your SSD outside of the array and use a larger drive as a temporary cache drive. Instructions for doing so were posted by jumperalex here http://lime-technology.com/forum/index.php?topic=30496
June 28, 201412 yr Author Thanks! Maybe I should replace my cache drive temporarily. I have an extra 4TB I could use. Something to think about. I wish I could have TWO cache drives! I really like running Plex on an SSD. Mount your SSD outside of the array and use a larger drive as a temporary cache drive. Instructions for doing so were posted by jumperalex here http://lime-technology.com/forum/index.php?topic=30496 Genius! Thank you!
June 29, 201412 yr Author BobPhoenix, Just wanted to say "Thank You" again! I got a 2TB cache drive installed and mounted my existing cache SSD as a separate disk and all is working perfectly.
Archived
This topic is now archived and is closed to further replies.