Sure enough (didn't notice): parallel: Warning: ssh to zzz@servername only allows for 10 simultaneous logins. You may raise this by changing /etc/ssh/sshd_config:MaxStartup on zzz@servername Using only 9 connections to avoid race conditions.
Thanks Ole and Martin. However, after changing to eg, MaxStartup 50:30:60 and running seq 100 | parallel -j+0 --eta --sshlogin 32/servername echo, 31 jobs run. j +1 makes this 32, but why dose j+0 only give 31? On Wed, Oct 24, 2012 at 2:36 PM, Ole Tange <[email protected]> wrote: > On Tue, Oct 23, 2012 at 10:02 PM, Grant Mc Auley <[email protected]> > wrote: > > I have access to a remote server with 4 cpu x 8 cores. However, when I > try > > eg, parallel -j+0 --sshlogin 32/servername, only the default 9 jobs are > run > > at a time. No matter what combinations of -j +/- x and --sshlogin > > z/servername I try, I do not get above 9 jobs. > > > > Sorry if I missed, it but is this a hard upper limit? If not, how can > > increase the number of cores used on the server? > > Does GNU Parallel print anything when you run: > > seq 100 | parallel -j+0 --sshlogin 32/servername echo > > /Ole >
