Hi,
We have at least one user who runs a large (300-400MB) program on our
small 2100 server that has 196MB of ram and 1.2GB of swap area. If this program
runs alone or with few other but smaller programs, the system behaves
well and only o few kB/s of paging happens. This is the case by night, but when
people start to work, paging goes up to 2-3MB/s and nobody works, the solution
adopted sofar being to kill the big one.
I would like to know if there is any program that can be used to suspend
a process and checkpoint it (that was the term used in the past) to the disk and
later on reactivate it where it was suspended so we can make a better use of our
resources? I have heard that this exists on large system (Crays and some
mainframe) but I have seen nothing in DU documentation. We run OSF1 3.2b.
Thanks in advance. I'll summarize if I get answers.
Philippe
pgouffon_at_if.usp.br
Instituto de Fisica,
Universidade de Sao Paulo.
Received on Thu Mar 21 1996 - 17:56:25 NZST