Menu

Transfer GigaFiles

2016-03-05
2016-03-08
  • Alfonso Huerta

    Alfonso Huerta - 2016-03-05

    Hello, I've been trying to set up a unicast transfer between 2 locations with an average 70 ms ping response up to 160 Mbits. The issue I have is a lot of NAKs and the client sending notifications back to the sever a lot of time after the server has stopped sending sections. I assume this may have to do with timing but I have no clue where to start debuggin this. If I use the congestion control, transfer speed is way too low. Running with the exact same parameters with a file 400 MB size runs fine, while a 34 GB file just keeps doing serveral passes until it aborts because of timeout.

    Thanks for the help

     
  • Dennis Bush

    Dennis Bush - 2016-03-05

    With the large file size, it's possible that disk I/O could be a limiting factor. You could try increasing the lower limit of the round trip time to give more time to write to disk and respond to requests by using the -r option to the server, or you could try adjusting the cache size using the -c option on the client.

     
  • Alfonso Huerta

    Alfonso Huerta - 2016-03-08

    Increasing the lower limit helped to get a more stable connection. But Im getting many NAKs accross the transfer and when it gets to the second pass it runs very fast and breaks with a transfer error "Transfer aborted by 0xnnnnnnnnn: Transfer timed out". Do I need to increase it more? I raised lower limit from 0.05 to .2 , would you suggest to go way up ?

     
  • Dennis Bush

    Dennis Bush - 2016-03-08

    You could try bringing up the minimum RTT more, or (given the number of NAKs) you could also try reducing the rate.

     

Log in to post a comment.

Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.