Hi! This is a problem that I've been having for quite some time. Searching the archives turns up a few other people reporting the same problem, but I can't find a solution. I use check_by_ssh to run the check_disk plugin on a few remote systems. When I run this from the command line, it always does the right thing. It returns the correct string and return code (checked with 'echo $?'). It works flawlessly. /usr/local/groundwork/nagios/libexec/check_by_ssh -H 10.194.145.142 -t 60 -l "nagios" -C "libexec/check_disk -w 400 -c 200 -p /" DISK OK - free space: / 14049 MB (82%);| /=3099MB;16747;16947;0;17147 -bash-3.2$ echo $? 0 -bash-3.2$ However, when this is set up to run through Nagios, it usually reports the check in the UNKNOWN state...even though it still gives the correct "OK..." string. [1234518396] SERVICE NOTIFICATION: fredriks;absaus0180;ssh_disk_root;UNKNOWN;service-notify-by-email;DISK OK - free space: / 1 4049 MB (82%): Occasionally, Nagios will change it to OK or WARNING or whatever is correct at the time, but for a good 90% of the time, it just sits there in UNKNOWN. It doesn't seem to matter if the target machine is Linux or Solaris, nor can I find any pattern to when and where it happens. Again, it *always* works perfectly when run from the command line. Worth to notice, i run this service check on simular systems with same arguments, and no problems what so ever. I've played with giving check_by_ssh a hostname instead of an IP address. I've tried it both with '-l nagios' and without (it's always run from nagios account). I've also tried the -f flag (the purpose of which I don't really understand). Nothing seems to help. Has anyone else seen this problem and/or found the cause of it? Thanks! /Fred