Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a robocopy job that creates a problem every time it runs. It saturates the bandwidth in such a way that no other traffic can get through. I read about the /IPG flag but am going in circles trying to find a number. I found calculators that will give me the amount of time it will take to move different size chunks, but I don't care about that. All I want is to allow other traffic across the network at the same time as robocopy. My network is 1Gbe. Anybody know of a usable number off-hand? I'm set at /IPG:10 right now and that's not working. Perhaps someone might know of a better alternative to robocopy?


share|improve this question
up vote 2 down vote accepted

I don't know usable numbers off the top of my head, but 10 sounds way to small to me. Since network speeds tend to increase by 10x each generation, that is how I would grow my values of IPG to find a working number. So try 100, 1000, 10000, etc., until you find something that works.

share|improve this answer
Interesting, thank you. Isn't IPG rated in milliseconds? Would and IPG:1000 cause a 1 second gap between packets? Or am I misunderstanding? – eth0 Nov 26 '12 at 18:52
You could be right. My point is to make sure you weren't starting at 1 and got to 10, and then gave up. If you try 100 now and it's too slow, then go for something between 10 and 100. Also, who cares what the actual delay is as long as it completes your copy in an acceptable amount of time and doesn't kill your network? – longneck Nov 26 '12 at 18:55
I find the IPG mainly functions like a short break that lets something else establish itself on the bandwidth share. Because of this using IPG to obtain say 70% bandwidth usage is a waste of your time. I find that 50 is a better starting point than 10, as 10 is a bit to short for another connection to establish itself well on my connections. I'm guessing this is due to the handshake at the start of the connection. I'm guessing that double the average ping time between servers is a reasonable guess for minimum IPG. Typical ping times here over internet links are about 20~30ms – BeowulfNode42 Jun 23 '15 at 5:37

Please note that the IPG feature in Robocopy is incompatible with the MT (multithreading) option. If you specify /MT, the /IPG switch will have no effect no matter what number you use.

If you're not using /MT, then just keep increasing the IPG until it no longer takes all the bandwidth.

Since you asked for alternatives, you could limit outbound transfer rates via GPO. (Probably not the best option.)

You could also use DFS Replication instead, which has scheduling and bandwidth usage features.

share|improve this answer
Thank you. We're not using the /MT option. DFS looks interesting, I'll start researching that. Thanks. – eth0 Nov 26 '12 at 19:05

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.