Paul Slootman <[EMAIL PROTECTED]> writes:
> The current 2.6.4 version does indeed have extra code to prevent
> timeouts happening, however that should only take effect when both sides
> of the transfer has at least version 2.6.4 and when the --timeout option
> is used...
> Of course, there have been many bug fixes as well, which may have solved
> the original problem; I can't find any clear reference about this in the
> NEWS file however.
Alas, no joy. Same problem with rsync 2.6.4-2 on both ends:
lothlorien:~> rsync -av -e 'rsh -x' [EMAIL PROTECTED]:Mail/ ./Mail/
This rsh session is encrypting input/output data transmissions.
receiving file list ... done
created directory ./Mail
./
active
archive/
archive/debian/
archive/debian/.marks
archive/debian/.overview
archive/debian/1
archive/debian/2
archive/flame/
archive/flame/.marks
archive/flame/.overview
archive/flame/1
archive/flame/10
and a hang at that point. This is using Kerberos rsh, but I've duplicated
the same thing with ssh in the past and doubt that's changed.
On the server end (windlord):
windlord:~> ps -ef | grep rsync
eagle 25262 25261 0 15:00 ? 00:00:00 tcsh -c rsync --server --sender
-vlogDtpr . Mail/
eagle 25263 25262 8 15:00 ? 00:00:01 rsync --server --sender
-vlogDtpr . Mail/
eagle 25280 11229 0 15:00 pts/15 00:00:00 grep rsync
windlord:~> strace -p 25263
Process 25263 attached - interrupt to quit
select(2, NULL, [1], NULL, {34, 890000}) = 0 (Timeout)
select(2, NULL, [1], NULL, {60, 0}) = 0 (Timeout)
select(2, NULL, [1], NULL, {60, 0}
on the client end (lothlorien):
lothlorien:~> ps -ef | grep rsync
root 414 408 0 Feb08 ? 00:00:00 supervise rsyncd
eagle 29326 663 2 14:59 pts/0 00:00:00 rsync -av -e rsh -x [EMAIL
PROTECTED]
eagle 29327 29326 0 14:59 pts/0 00:00:00 rsh -x windlord -l eagle rsync -
eagle 29329 29327 0 14:59 pts/0 00:00:00 rsh -x windlord -l eagle rsync -
eagle 29334 29326 0 15:00 pts/0 00:00:00 rsync -av -e rsh -x [EMAIL
PROTECTED]
lothlorien:~> strace -p 29334
Process 29334 attached - interrupt to quit
select(6, [5], [], NULL, {15, 370000}) = 0 (Timeout)
select(6, [5], [], NULL, {60, 0}) = 0 (Timeout)
select(6, [5], [], NULL, {60, 0}
so as you can see, the network protocol appears to have deadlocked and
both sides are waiting for the other. This is the same symptoms I've been
seeing all along, I'm afraid.
--
Russ Allbery ([EMAIL PROTECTED]) <http://www.eyrie.org/~eagle/>
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
