Hi All,
strange thing happened the other day on our network. We have a couple of linux servers (2.6.17, x86) with backup exec agent installed.
After running fine for about a month, both servers still show the exported directories, but on connect from the BE server they throw a message "...is out of network connections.." or something of a sort.
On closer look on the Linux side, it appears that after backup agent starts, there is an established session with the BE server that looks like this (11.5.5.4 is the BE server, .65 is the linux box with agent):
tcp 0 0 11.5.5.65:8364 11.5.5.4:6101 TIME_WAIT
tcp 0 0 11.5.5.65:8193 11.5.5.4:6101 TIME_WAIT
After about 30 seconds, the ports used on the client change, and the port number increases:
tcp 0 0 11.5.5.65:8196 11.5.5.4:6101 TIME_WAIT
tcp 0 0 11.5.5.65:8195 11.5.5.4:6101 TIME_WAIT
-- some time passed---
tcp 0 0 11.5.5.65:8197 11.5.5.4:6101 TIME_WAIT
tcp 0 0 11.5.5.65:8196 11.5.5.4:6101 TIME_WAIT
And so on. Eventually, the local port number reaches 65535 and the agent stops responding, so it looks like the agent doesnt "roll over" after reaching max port number for tcp.
Has anyone seen this before, is there some sort of a fix for this?
Here's the agent config file on the Linux server:
name linux
export /etc as etc write_protect
export /usr/local/zope as zope write_protect
tell backup
tell_interval 30
exclude_dir /dev
exclude_dir /proc
Thank you!
max
