25.06.2015 Views

Administering Platform LSF - SAS

Administering Platform LSF - SAS

Administering Platform LSF - SAS

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Chapter 8<br />

Understanding Resources<br />

Automatic Detection of Hardware Reconfiguration<br />

Supported platforms<br />

Dynamic changes in ncpus<br />

Some UNIX operating systems support dynamic hardware reconfiguration—<br />

that is, the attaching or detaching of system boards in a live system without<br />

having to reboot the host.<br />

<strong>LSF</strong> is able to recognize changes in ncpus, maxmem, maxswp, maxtmp in the<br />

following platforms:<br />

◆ Sun Solaris 2.5+<br />

◆ HP-UX 10.10+<br />

◆ Compaq Alpha 5.0+<br />

◆ IBM AIX 4.0+<br />

◆ SGI IRIX 6.2+<br />

<strong>LSF</strong> is able to automatically detect a change in the number of processors in<br />

systems that support dynamic hardware reconfiguration.<br />

The local LIM checks if there is a change in the number of processors at an<br />

internal interval of 2 minutes. If it detects a change in the number of<br />

processors, the local LIM also checks maxmem, maxswp, maxtmp. The local LIM<br />

then sends this new information to the master LIM.<br />

Dynamic changes in maxmem, maxswp, maxtmp<br />

If you dynamically change maxmem, maxswp, or maxtmp without changing the<br />

number of processors, you need to restart the local LIM with the command<br />

lsadmin limrestart so that it can recognize the changes.<br />

If you dynamically change the number of processors and any of maxmem,<br />

maxswp, or maxtmp, the change will be automatically recognized by <strong>LSF</strong>. When<br />

it detects a change in the number of processors, the local LIM also checks<br />

maxmem, maxswp, maxtmp.<br />

Viewing dynamic hardware changes<br />

lsxxx Commands<br />

bxxx Commands<br />

<strong>Platform</strong><br />

MultiCluster<br />

There may be a 2 minute delay before the changes are recognized by lsxxx<br />

commands (for example, before lshosts displays the changes).<br />

There may be at most a 2 + 10 minute delay before the changes are recognized<br />

by bxxx commands (for example, before bhosts -l displays the changes).<br />

This is because mbatchd contacts the master LIM at an internal interval of 10<br />

minutes.<br />

Configuration changes from a local cluster are communicated from the master<br />

LIM to the remote cluster at an interval of 2 * CACHE_INTERVAL. The<br />

parameter CACHE_INTERVAL is configured in lsf.cluster.cluster_name<br />

and is by default 60 seconds.<br />

This means that for changes to be recognized in a remote cluster there is a<br />

maximum delay of 2 minutes + 2*CACHE_INTERVAL.<br />

<strong>Administering</strong> <strong>Platform</strong> <strong>LSF</strong> 149

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!