Hi,
I'm trying to use rsync (2.6.8) through STunnel (4.16) to synchronize two PC over the internet (they both have the same version of stunnel and rsync).
On the server, I have STunnel running with this configuration: cert = stunnel.pem
; Some performance tunings socket = l:TCP_NODELAY=0 socket = r:TCP_NODELAY=0
; Some debugging stuff useful for troubleshooting debug = 7 output = stunnel.log
[rsync] accept = 273 connect = 873
On the client I have this configuration: cert = stunnel.pem
; Some performance tunings socket = l:TCP_NODELAY=0 socket = r:TCP_NODELAY=0
; Some debugging stuff useful for troubleshooting debug = 7 output = stunnel.log
client = yes
[rsync] accept = 873 connect = UrlOfMyServer:273
Then I connect with rsync using rsync -a -h --progress --stats --partial Account@localhost::RemotePath LocalPath
I think the connection is not really stable. RSync sometimes hang. The log file on the client side show 2006.09.11 21:33:11 LOG7[2008:1468]: SSL_read returned WANT_READ: retrying at 30Hz (that's right, it means 30 times a second). My log file nearly gets bigger faster than my downloaded file ;)
Is it expected? What does this line mean? I turned TCP_NODELAY to 0 thinking that it my give me synchronous sockets and that it would remove that error, but it didn't...
Any help appreciated, Thanks
Vincent Lascaux wrote:
I'm trying to use rsync (2.6.8) through STunnel (4.16) to synchronize two PC over the internet (they both have the same version of stunnel and rsync).
I may have omitted the most important part: both client and server are windows machines.
It's a well-known bug in cygwin library (used by rsync on Windows).
https://bugzilla.samba.org/show_bug.cgi?id=2208
Best regards, Mike
It's a well-known bug in cygwin library (used by rsync on Windows).
Thanks. I worked around the issue setting a timeout of 5 minutes on every state (idle, busy, close, connect...) and doing rsync in a loop. While it's not a perfect solution, it works well enough for me.