Chapter 4. Verify the Installation

After completing the installation, you can run the confidence script to verify that the installation process has completed successfully. All of the administration scripts are in the /usr/local/bin directory.

What confidence Verifies

The confidence script checks the following:

  • That the user can use the rsh(1) command among all nodes (in any combination) within the cluster

  • That the batch system is correctly configured and functioning

  • That a sample test job completes and is correct

Using the confidence Script

You must run the confidence script on the head node.

To run this script, perform the following steps:

  1. Log in to the head node as the application user (preferably, not as root).

  2. Execute the confidence script.

  3. Make sure that /usr/local/bin is in your path. For example,

    (For bash shell) PATH=$PATH:/usr/local/bin; export PATH
    (For csh shell)setenv PATH "${PATH}:/usr/local/bin"

The following example shows output for a successful test on a two-node cluster:

[sgigenomecluster@windy test]$ confidence
PBS processes currently running.
LSF processes currently running.
CT-BLAST scheduler selected: lsf
Running schedulers: PBS [yes] LSF [yes]
Testing that user has rsh access to and from all hosts.
Shipping a script to each host that will rsh to all other hosts.
PASS  All hosts can rsh to every other host.
INFO  hostnames match SGIGenomeCluster node names.
Executing LSF tests
PASS 1 node(s) configured and available to LSF.

Testing that LSF can execute a simple script ...
Submitting test job to default LSF queue. may take some time
Checking that output matches cluster node list ...
PASS  LSF bsub output contains node windy output.
PASS  LSF bsub output contains node gusty output.
Basic LSF test complete.
Executing ctblast test, this will take some time ~5 minutes
PASS CT-BLAST successful.
confidence: No errors detected.  Success!!!

To verify the packages and configuration information of a particular node in the cluster, execute the clusterinfo command, as follows:

$ /usr/local/bin/clusterinfo

Output is as follows:

###### SGI GenomeCluster info on gusty #################
hosts uname -a info: 
Linux gusty 2.2.16-4SGI_39smp #1 SMP Fri Sep 29 00:32:45 EDT 2000 i686 unknown

SGI GenomeCluster RPMS 
--------------------------------
data-1.0-1
ncbitoolkit-6.1-1
package ctblast is not installed

SGI GenomeCluster files
--------------------------------
/etc/nodes file does not exist.
/etc/ctblast/ctblast.config file exists :
    statuspath=/etc/ctblast/qstat
dbpath=/cache/database
dbhost=sgibio07:/data
submit=/usr/local/bin/qsub
status=qstat
sortingmethod=cousein
killjob=/usr/local/bin/qdel
group=ctblast
tmppath=/tmp
scheduler=pbs
blastall=/usr/local/bin/blastall
End of file
/etc/ctblast/ctblast.parameters file exists :
    # format is as follows
# parameter name,which configuration file it is allowed in, the value of the parameter, 
where to place the configuration file, a comment about the parameter
blastall,all,/usr/local/bin/blastall,all,"The absolute path to the blastall binary"
group,all,ctblast,submission_node,"The queue to use for job submission"
sortingmethod,all,cousein,submission_node,"The load balancing method to use"
dbhost,all,sgibio08:/data,all,"The name of the machine:/path acting as a 
central repository for the databases"
dbpath,all,/cache/db,all,"The path to the databases on the on the compute node"
tmppath,all,/tmp,all,"The location of a temporary directory to use for job 
construction/submission"
computenodetmp,all,/cache/run/,all,"the location of the remote temporary run 
directory"
scheduler,all,pbs,submission_node,"The name of the job scheduler to use"
submit,all,/usr/local/bin/qsub,submission_node,"The program used to enqueue a job"
status,all,qstat,submission_node,"The method used to query the run status"
statuspath,all,/etc/ctblast/,submission_node,"The location of the status inquiry method"
getstatus,all,/usr/local/bin/qstat -f ,submission_node,"The command used for status inquiry"
killjob,all,/usr/local/bin/qdel,submission_node,"How to kill errant jobs"
mkdir,all,/bin/mkdir,all,"the mkdir command"
rcp,all,/usr/bin/rcp,all,"the rcp command"
pwd,all,/bin/pwd,all,"the pwd command"
hostn,all,/bin/hostname,all,"the hostname command"
date,all,/bin/date,all,"the date command"
echo,all,/bin/echo,all,"the echo command"
rm,all,/bin/rm,all,"the rm command"
End of file
~/.ctblastrc file does not exist.
.ctblastrc file exists :
    dbhost=windy:/data
status=bjobs
rcp=/usr/bin/rcp
sortingmethod=cousein
killjob=/usr/local/lsf/bin/bkill
pwd=/bin/pwd
hostn=/bin/hostname
group=ctblast
blastall=/usr/local/bin/blastall
statuspath=/etc/ctblast/
dbpath=/cache/db
submit=/usr/local/lsf/bin/bsub
computenodetmp=/cache/run/
mkdir=/bin/mkdir
tmppath=/data/RUNS
date=/bin/date
rm=/bin/rm
echo=/bin/echo
scheduler=lsf
shell=/bin/bash
End of file
CTBLASTALL user is using lsf scheduler

Checking configuration files ...
No errors found.

LSF 4.0.1, Jun  2 2000
Copyright 1992-2000 Platform Computing Corporation

My cluster name is SGIGenomeCluster
My master name is windy

Checking configuration files ...

There are warning errors.

Do you want to see detailed messages? [y/n] y
Checking configuration files ...
Jan  4 12:17:27 2001 19444 6 4.0.1 minit: Trying to call LIM to get cluster name ...
Jan  4 12:17:27 2001 19444 6 4.0.1 Batch is enabled
Jan  4 12:17:27 2001 19444 6 4.0.1 JobScheduler is disabled
Jan  4 12:17:27 2001 19444 6 4.0.1 autoAdjustInit: Auto-adjustment is disabled
Jan  4 12:17:27 2001 19444 4 4.0.1 do_Groups: File 
/usr/local/lsf/mnt/conf/lsbatch/configdir/lsb.hosts at line 54: Empty hostgroup section
Jan  4 12:17:27 2001 19444 3 4.0.1 main: Local host is not master 
Jan  4 12:17:27 2001 19444 6 4.0.1 Checking Done
---------------------------------------------------------
No fatal errors found.
Warning: Some configuration parameters may be incorrect.
         They are either ignored or replaced by default values.

SGI GenomeCluster processes
--------------------------------
root      1573     1  0  2000 ?        00:53:09 /usr/local/lsf/etc/lim
root      1575     1  0  2000 ?        00:00:02 /usr/local/lsf/etc/res
root      1577     1  0  2000 ?        00:00:24 /usr/local/lsf/etc/sbatchd
root      1578  1573  0  2000 ?        00:00:34 /usr/local/lsf/etc/pim
root       898     1  0  2000 ?        00:00:00 /usr/share/pcp/bin/pmcd
root      1042     1  0  2000 ?        00:00:00 pmlogger -P -c config.default 20
[biopak@gusty biopak]$ 

To obtain help for confidence , use the -h option with the confidence command. If you want debug information, use the -v option.