2 – InfiniPath Cluster Administration
Performance and Management Tips
IB6054601-00 D 2-17
Q
0zwxSL7GP1nEyFk9wAxCrXv3xPKxQaezQKs+KL95FouJvJ4qrSxxHdd1NYNR0D
avEBVQgCaspgWvWQ8cL
0aUQmTbggLrtD9zETVU5PCgRlQL6I3Y5sCCHuO7/UvTH9nneCg==
Change the file to mode 600 when finished editing.
4. On each node, the system file
/etc/ssh/sshd_config must be edited, so that
the following four lines uncommented (no
# at the start of the line) and are set
to
yes. Each of these lines is normally present, but commented out and set to
no by default.
RhostsAuthentication yes
RhostsRSAAuthentication yes
HostbasedAuthentication yes
PAMAuthenticationViaKbdInt yes
5. After creating or editing these three files in steps 2, 3 and 4, sshd must be
restarted on each system. If you are already logged in via
ssh (or any other
user is logged in via
ssh), their sessions or programs will be terminated, so do
this only on idle nodes. Tell
sshd to use the new configuration files by typing
(as root):
# killall -HUP sshd
NOTE: This will terminate all ssh sessions into that system. Run from the
console, or have a way to log into the console in case of any problem.
At this point, any user should be able to login to the
ip-fe front end system, and
then use
ssh to login to any InfiniPath node without being prompted for a password
or pass phrase.
2.9.1
Process Limitation with ssh
MPI jobs that use more than 8 processes per node may encounter an SSH throttling
mechanism that limits the amount of concurrent per-node connections to 10. If you
need to use more processes, you or your system administrator should increase the
value of ’MaxStartups’ in your sshd configurations. See appendix C.8.8 for an
example of an error message associated with this limitation.
2.10
Performance and Management Tips
The following section gives some suggestions for improving performance and
simplifying management of the cluster.
2.10.1
Remove Unneeded Services
An important step that the cluster administrator can take to enhance application
performance is to minimize the set of system services running on the compute