Platform LSF Administrator's Primer Version 6.2

Common LSF Problems
Platform LSF Administrator’s Primer
56
the PRODUCTS line in LSF_CONFDIR/lsf.cluster.cluster_name must
contain:
PRODUCTS=LSF_Base LSF_Manager
Modify the PRODUCTS line to fix the error.
See Chapter 4, “Working with LSF Licenses” for information about working with a
permanent LSF license.
◆ lsf.conf is not in the location specified by LSF_ENVDIR.
Check that LSF_LICENSE_FILE parameter in
lsf.conf is correct.
LSF_LICENSE_FILE can be either:
❖ The full path name to the license file
For example:
LSF_LICENSE_FILE=/usr/share/lsf/lsf_62/conf/license.dat
❖ For a permanent license, the name of the license server host and the port
number used by
lmgrd in the form port_nu mber@host_name.
For example, if your license file contains the line:
SERVER hosta 68044d20 1700
LSF_LICENSE_FILE would be:
LSF_LICENSE_FILE=1700@hosta
Permanent license problems:
◆ Run lmhostid on your license server host, and check that the host ID matches the
first line of your license file.
The
lmhostid command is in LSF_SERVERDIR. For example:
/usr/share/lsf/lsf_62/6.2/sparc-sol2/etc/lmhostid
◆ Check the path to the LSF vendor daemon lsf_ld in the license file. The path
must be specified in the DAEMON line of the license file. For example:
DAEMON lsf_ld /usr/share/lsf/lsf_62/6.2/sparc-sol2/etc/lsf_ld
◆ If there are multiple license management daemons running on your machine, or the
default license port 1700 is being used by another process, edit the license file and
change the port number as described in “Setting the FLEXlm server TCP port in
license.dat” on page 44.
◆ Use lmstat to check if you have enough licenses for your cluster. Each CPU on
LSF server hosts (not each server host itself) in the cluster uses one LSF server
license. Each LSF client host (not each CPU) uses one LSF client license.
◆ After making any changes to your permanent license, stop lmgrd first, then stop
lsf_ld before restarting lmgrd; otherwise, lmgrd will start another lsf_ld
right away and
lsf_ld will not restart correctly.
3 LIM unavailable
Sometimes lim is up, but the lsload command displays one of the following error
messages:
◆ lsid: ls_getclustername() failed: LIM is down; try later
◆ LSF daemon (LIM) not responding ... still trying
◆ ls_limcontrol: Communication time out