Shell Programming and Scripting

There's probably a better way to do what I'm doing and if so, let me know an alternative.

I have a script that runs rsh for every entry in a txt file to go get the file size off the remote server. I then compare the remote size with a local size.

My script worked fine until recently when I was comparing a large list (600+). After about 400 files, the rsh command returned nothing.

My repeated command would be similar to:

Code:

rsh server1 dir /u2/web/a.jsp 2> /dev/null |awk '{print " ",$5," ",$9}'|tr -s " "

/u2/web/a.jsp would be one of the 600 files I'm getting the size of.

I'm able to continue making it work by putting in a 2 second sleep every 10th file.

Any thoughts as to why and/or possibly any alternatives?

Thank you for your time.

Moderator's Comments:

Use code tags please, see PM.

First off, UNIX doesn't have dir, it's probably just an alias to ls -l.

Why not list more than one file at once, to reduce the number of rsh calls?

You could also use a tool like rsync to compare and synchronize remote folders with each other.

The system might run out of sockets.
Place a sleep 1 between the rsh calls!
Bundle commands to one host

Code:

rsh host '
command1
command2
'

--
Just seen your confirmation.
A closed TCP connection stays a while in TIMEWAIT (can be seen with netstat) in the hope of a re-open. But a rsh process cannot re-open the connections of the previous rsh commands. If rsh are fired too rapidly the system runs out of sockets.