Hi. I venture to post here, as my question relates to Perl on linux server.

I'm operating a high-cost application on a shared server (web hotel), where I have no access (nor will they make such customisations ) to memory/cpu time ulimit specifications. Each process is maximally taking hours to process (and I would hate to impose input limitations), which is untenable for the shared server host.

The application is as follows: web form sends perl script three files, which are sent to another server for processing, and while the processing takes place, shared server is *not doing anything* but keeping up the ssh connection (should not take that much cpu necessarily should it?), and finally result files are copied back.

Therefore, (1) I would need to be able to limit use of the one processor from a perl cgi script. (How) can that be done? Typically 70% of cpu power is used through the process, so alternatively (2) I would like it to use a different limit for the ssh-tunnel file transfers than for the 'idle time' (i.e. high-low-high cpu usage curve). Since that might be even better, can that be done (again from a script, not necessarily perl)?

thanks, Harri S