What to do when backups are failing, and the log says "Socket timed out after waiting for more than 1,800 seconds to read/write"
What happens here is that the socket doesn't get closed properly when R1soft is finished doing its job. The socket will stay in a Close_Wait state indefinitely.
Restart R1soft. Most of the time this will not work, and the server will need to be restarted. Firstly try a graceful restart with /etc/init/.d/cdp whatever, or /usr/sbin/r1soft/etc depending on version for where the startup stuff is. You should try a "status" first to see what it's doing. Typically, a restart operation doesn't work as it doesn't wait long enough for the sockets to close and just tries to start up again. If issuing a stop also doens't work, pull out the PID and kill it. Use a ps and grep for cdp. You can try to kill/ kill -9 the PID and if that works, sweet, start r1soft. If the PID refuses to go down, you'll have to get permission to restart the machine. Backups will not function until this is resolved.