After realizing that I had my server on a 10/100 switch, I swapped in a gigabit switch & everything is pleasingly faster!

 

As a test, I copied a file (~3GB) from the server to my Win10 machine and saw speeds between 68-85MB/s. When I copied the file back (to a different destination), I saw speeds peak in the 60MB/s range, but mostly it was in the 30-40MB/s range, with a dip down to 9MB/s.

 

Obviously, the network hardware is the same going both directions, is the difference in speed due to the difference in hard drives at each end? I was writing to my cache drive on the server, which is a 250GB 2.5" reiserfs formatted Samsung Spinpoint with 8MB cache at 5400RPM, while the drive in my Win10 box is a 1TB 3.5" NTFS formatted Hitachi 7K1000.C with 32MB Cache at7200 RPM.

 

The attached image is of the write from Win10 to unRAID.

 

The more I think about it, the more the answer becomes obviously "yes", but confirmation is nice. Also, since I've got my docker.img on this drive, will I see any significant docker performance improvement if I go to a faster drive?

See saw transfer is a "caracteristic" of reiserfs disks, it should be steady with xfs, as for the speed, that is an older and slower disk, you would get better speeds with a more recent disk or better yet, a SSD.

I was figuring the general write speed was a disk issue. I didn't know about the see saw speed aspect of reiserfs.

 

Thanks for the confirmation and the information.

Sorry for resurrecting the dead, 

 

I have a similar issue it seems when i write to my array from the array (copy from one share to another)..... this yo yo issue happens on either a cache write or direct to the array.

 

 

 

I do the copy via a root share, like this guide below.

 

When writing to the array from another machine to the cache or array it keeps a steady performance. Is this a normal read write situation?

 

 

 

 

 

bump,