> 1) are you sure your traffic is traversing the high bandwidth link? > Always good to check ...
Yup, only have one on each side, one switch connecting them together. (ex network engineer, always check the network first, we can't be trusted) :) > 2) how many files are you xfering? Are these generally large files or > many small files, or a distribution with a long tail towards small > files? small, 100-200 files, definitely not a MDT issue > 3) wire speed xfers are generally the exception unless you are doing > large sequential single files. There are tricks you can do to enable > this, but they are often complex. You can use the array of > writers/readers, and leverage parallelism, but you risk invoking > congestion/pause throttling on your switch. agreed, if i use IOR or something alike i can drive lustre to 900MB/sec. i figured i wouldn't get rsync to push that hard, but i was hoping for better then 100MB/sec i'm doing this locally on the box, so i'd hoped ssh and it's compression/encryption factors wouldn't come into play, but i guess they do some how. i've been messing about with the '-e ssh -T -c <cipher> -o Compression=no -x' option to see if i could tease out something relevant, but that hasn't turned up anything i'd be happy if i could even break 120MB/sec at this point. i wonder if there's something odd with the rsync verions, the client side is redhat 6, i'll see if i can pull in the source and recompile. maybe there's some enhancement i'm missing _______________________________________________ Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit https://beowulf.org/cgi-bin/mailman/listinfo/beowulf