scp and rsync stall when transferring from Ubuntu hardy heron 8.04

[Vermind]

steve c, yours seems to be a different issue. Mine occurred only on transfers originating from the Ubuntu host, and only when SSL was used with the transfers. tla versioning system commits (large ones), rsync transfers, and scp transfers were stalled. If I put files in my HTTP share folder, they could be downloaded by others without problems. If I transferred files using netcat, they could be transferred as well. Transfers using Skype and msn also succeeded. There never was a problem with update downloading.

[steve c]

Perhaps so. I've never seen scp stall before so I assumed it was the same as your issue.

[Ubuntu Warrior]

At the risk of sounding very pessimistic here, the fact that rsync doesn't work properly on Ubuntu is a MAJOR issue. We currently use rsync to back up our Zimbra files hosted on an Ubuntu 6.10 server but it seems to arbitrarily leave out some files. Permissions are correct and doing a tar on the files works fine.

We tried to use scp but this also had problems in reliably transferring files across systems.

Suppose I don't need to highlight that the reliable transfer of business information is critical to an enterprise solution.

At this stage we have been forced back onto a Sun Solaris box until we can sort this issue.

Any ideas?

[vhex]

I had similar transfer problems with my wifi. Try adding this at the end of your /etc/sysclt.conf and then reload it via sysclt -p

Thank you SO MUCH!

[Vermind]

Hello,
It appears stalls are occurring again when I transfer to the University Linux server. When I first tried the suggested net modifications, transfers there did not stall. Something is wrong...

[steve c]

Well, you can chalk my problems up to my switch dying. Apparently the few different ports I tried in my tests had all failed while the others were working so it seemed like not a switch problem.

When different symptoms appeared on my other machines (not running ubuntu), I noticed that setting my speeds from 1000baseT to 100baseT caused a significant speed up, but it was still slow. Wireshark showed a large number of duplicate tcp packets being sent. I pulled the switch out of the network and replaced it with a slower router and now my network issues have disappeared.

Thanks for all of the help and my apologies for wasting your time.

[MartijnNL]

One thing: if you're using Firestarter (firewall), that might cause the problem.

There's something about that problem here:

https://help.ubuntu.com/community/Fi...%20connections

The solution in that article didn't solve the problem for me. Uninstalling firestarter did. Just switching firestarter off is not enough. I'll go try ufw.

[detroit/zero]

I had similar transfer problems with my wifi. Try adding this at the end of your /etc/sysclt.conf and then reload it via sysclt -p

net.core.rmem_default = 524288
net.core.rmem_max = 524288
net.core.wmem_default = 524288
net.core.wmem_max = 524288
net.ipv4.tcp_wmem = 4096 87380 524288
net.ipv4.tcp_rmem = 4096 87380 524288
net.ipv4.tcp_mem = 524288 524288 524288
net.ipv4.tcp_rfc1337 = 1
net.ipv4.ip_no_pmtu_disc = 0
net.ipv4.tcp_sack = 1
net.ipv4.tcp_fack = 1
net.ipv4.tcp_window_scaling = 1
net.ipv4.tcp_timestamps = 1
net.ipv4.tcp_ecn = 0
net.ipv4.route.flush = 1

Worked for me.. Thanks!

I have two computers on a home network and am constantly moving files between them. I upgraded one to Jaunty not so long ago, and have been having problems with scp transfers ever since. I also noticed that ssh sessions from one to the other would periodically freeze.

It's a little early to be certain, but I just added those lines to the end of sysctl on both systems and now everything seems to be going OK.

Thanks again!

[Vermind]

For me, this was fixed in Jaunty.

[caue.rego]

for me, the MTU thing from 1500 to 1490 fixed the issue in Jaunty.