<HTML> <HEAD> <TITLE>Re: NRPE: Unable to read output</TITLE> </HEAD> <BODY> I guess I should include the plugin script: #!/usr/bin/perl -w #check for item count in ListingCatlogue URL use POSIX; use strict; use File::Basename; use Getopt::Long; use vars qw( $opt_critical $opt_warning $opt_username $opt_password $opt_instance $opt_help $opt_usage $opt_version ); sub print_usage(); sub print_help(); # # Version and Author info my $author = "Matthew Litwin"; my $datemod = "September 2, 2009"; my $version = "0.1.0"; # 0.1.0 - initial cannibalization from check_lcs_update.pl my $progname = basename($0); my %ERRORS = ('UNKNOWN' => '-1', 'OK' => '0', 'WARNING' => '1', 'CRITICAL' => '2'); my $sqlplus = '/opt/oracle/product/10gRac/bin/sqlplus'; my $sqlfile = '/usr/local/nagios/libexec/check_scrub_backlog.sql'; Getopt::Long::Configure('bundling'); GetOptions ( "c=s" => \$opt_critical, "critical=s" => \$opt_critical, "w=s" => \$opt_warning, "warning=s" => \$opt_warning, "u=s" => \$opt_username, "username=s" => \$opt_username, "p=s" => \$opt_password, "password=s" => \$opt_password, "i=s" => \$opt_instance, "instance=s" => \$opt_instance, "h" => \$opt_help, "help" => \$opt_help, "usage" => \$opt_usage, "V" => \$opt_version,"version" => \$opt_version ) || die "Try `$progname --help' for more information.\n"; sub print_usage() { print "Usage: $progname -w WARNING -c CRITICAL -u username -p password -i instance\n"; print " $progname --help\n"; print " $progname --version\n"; } sub print_help() { print "$progname - check item count in ListingCatlogue URL\n"; print "Options are:\n"; print " -c, --critical\n"; print " -w, --warning\n"; print " -u, --username\n"; print " -p, --password\n"; print " -i, --instance\n"; print " -h, --help display this help and exit\n"; print " --usage display a short usage instruction\n"; print " -V, --version output version information and exit\n"; } if ($opt_help) { print_help(); exit $ERRORS{'UNKNOWN'}; } if ($opt_usage || !($opt_critical && $opt_warning && $opt_username && $opt_password && $opt_instance && $opt_instance)) { print_usage(); exit $ERRORS{'UNKNOWN'}; } if ($opt_version) { print "$progname $version\n"; print "$author, $datemod\n"; exit $ERRORS{'UNKNOWN'}; } if (!-x $sqlplus) { print "sqlplus not found or not executable at: $sqlplus\n"; exit $ERRORS{'UNKNOWN'}; } if (!-r $sqlfile) { print "sql command file not found or not readable at: $sqlfile\n"; exit $ERRORS{'UNKNOWN'}; } # This SQL request returns a row count. my $scrubcount=`$sqlplus -S $opt_username/$opt_password\@$opt_instance \@$sqlfile`; if ($? == 0) { chomp($scrubcount); $scrubcount =~ s/^\s+//; if ($scrubcount =~ m/^\d+$/) { print "Scrub Backlog Count: $scrubcount\n"; my $state; $state="OK"; if ($scrubcount >= $opt_warning) {$state="WARNING";} if ($scrubcount >= $opt_critical) {$state="CRITICAL";} exit $ERRORS{$state}; } else { print "Output not a numeric value\n"; exit $ERRORS{'UNKNOWN'}; } } else { print "sqlplus error: $?\n"; exit $ERRORS{'UNKNOWN'}; } On 9/3/09 12:48 PM, "Matthew Litwin" <<a href="mlitwin@stubhub.com">mlitwin@stubhub.com</a>> wrote: <BLOCKQUOTE>I wrote a perl plugin that seems to work fine when I run it locally on the remote host as the nagios user, however when I try to execute it via NRPE I get the old familiar nebulous “NRPE: Unable to read output”. I have debugging on for NRPE logging and it doesn’t tell much more. (Note this is in reverse order is it is out of splunk. IPs are obfuscated) Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 903583 daemon.debug] Connection from X.X.X.X closed. host=sjvp00dbs001.XX.com Options sourcetype=syslog Options source=udp:514 Options 589/3/09 7:27:01.000 PM Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 869297 daemon.debug] Return Code: 1, Output: NRPE: Unable to read output host=sjvp00dbs001.XX.com Options sourcetype=syslog Options source=udp:514 Options 599/3/09 7:27:01.000 PM Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 757686 daemon.debug] Command completed with return code 1 and output: host=sjvp00dbs001.XX.com Options sourcetype=syslog Options source=udp:514 Options 609/3/09 7:27:01.000 PM Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 462736 daemon.debug] Running command: /usr/loca/nagios/libexec/check_scrub_backlog.pl -w 1000 -c 2000 -u XXXX -p XXXX -i XXXX Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 881351 daemon.debug] Host is asking for command 'check_scrub_backlog' to be run... host=sjvp00dbs001.XX.com Options sourcetype=syslog Options source=udp:514 Options 629/3/09 7:27:01.000 PM Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 385967 daemon.debug] Host address is in allowed_hosts host=sjvp00dbs001.XX.com Options sourcetype=syslog Options source=udp:514 Options 639/3/09 7:27:01.000 PM Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 654915 daemon.debug] Connection from X.X.X.X port 34549 host=sjvp00dbs001.XX.com Options sourcetype=syslog Options source=udp:514 Options 649/3/09 7:27:01.000 PM Sep 3 19:27:01 sjvp00dbs001.XX.com Sep 3 19:27:01 nrpe[12434]: [ID 879649 daemon.debug] Handling the connection... The command that is trying to run looks corrected- arguments are being sent correctly. I have tested the command using the same user that NRPE is running as, nagios. I did build in output for error messages in case there were problems like issues with ENV, but obviously they are not making it to NRPE. I confirmed my output text has EOLs. What is especially weird is that the return code makes it back correctly but the output is just “NRPE: Unable to read output” even as logged by nrpe. The output should look like this, which is what it gets locally when ran: Scrub Backlog Count: 1083 Nothing unusually with that, right? Any help appreciated, thanks. Does anyone have any other suggestions on how I might debug this? Since NRPE is working for all my other plugins </BLOCKQUOTE> </BODY> </HTML>