The number is just the exit code from the script. Normally you'd have an "exit 0" at the end of the script when everything is OK.

These are custom client-side scripts you're running, so what the different status numbers mean I cannot tell.

Debugging - you can run the scripts by hand, and then get the exit code from the "$?" environment right after the script finishes.

Regards, Henrik

Hi Henrik,

Nice to hear from you. I hope you're not still on vacation and answering this email. If you are, my apologies, this can wait, but I'm going to send the responses.

The exit status seems to be coming from Hobbit, line 528 in common/hobbitlaunch.c The emcpp is a custom script that checks EMC's PowerPath software on a given host. The temp is the temperature.sh (V4.0) script from deadcat. The emcpp script returns 0, temp script returns 1. The log data is from clientlaunch.log. My standard client package includes these tests, as most of the machines here do run PowerPath and DellOM. The interesting thing about temp is my BBOSTYPE variable has an extra carriage return in it. [hobbit at hershey ext]$ echo "$BBOSTYPE" linux

[hobbit at hershey ext]$ echo "\"$BBOSTYPE\"" "linux "

[hobbit at hershey ext]$ ../bin/bbcmd bash 2007-08-14 17:59:54 Using default environment file /apps/hobbit/client/etc/hobbitclient.cfg [hobbit at hershey ext]$ env SED=/bin/sed GREP=/bin/grep BBDISP=0.0.0.0 WCC=/usr/bin/wc BBDISPLAYS=128.221.22.29 HOSTNAME=hershey UPTIME=/usr/bin/uptime TAIL=/usr/bin/tail SHELL=/bin/bash TERM=xterm AWK=/bin/awk HISTSIZE=1000 WHO=/usr/bin/who BBCLIENTLOGS=/apps/hobbit/client/logs SSH_CLIENT=128.221.22.29 48509 22 DF=/bin/df -Pk BBOSTYPE=linux

HOBBITCLIENTHOME=/apps/hobbit/client SSH_TTY=/dev/pts/0 CUT=/bin/cut USER=hobbit LS_COLORS= LD_LIBRARY_PATH=:/usr/local/freetds/lib SORT=/bin/sort HEAD=/usr/bin/head SYBASE=/usr/local/freetds TOP=/usr/bin/top RM=/bin/rm DATE=/bin/date EXPR=/usr/bin/expr BBPORT=1984 PATH=/usr/kerberos/bin:/bin:/usr/bin:/sbin:/usr/sbin:/etc MAIL=/var/spool/mail/hobbit MV=/bin/mv EGREP=/bin/egrep PWD=/apps/hobbit/client/ext INPUTRC=/etc/inputrc PS=ps ax BB=/apps/hobbit/client/bin/bb LANG=en_US.UTF-8 MACHINE=hershey CP=/bin/cp LN=/bin/ln SSH_ASKPASS=/usr/libexec/openssh/gnome-ssh-askpass MAXLINE=32768 HOME=/apps/hobbit SHLVL=2 TR=/usr/bin/tr ID=/usr/bin/id BBTMP=/apps/hobbit/client/tmp MACHINEDOTS=hershey LOGNAME=hobbit SSH_CONNECTION=128.221.22.29 48509 10.250.132.12 22 WC=/usr/bin/wc -l LS=/bin/ls LESSOPEN=|/usr/bin/lesspipe.sh %s FIND=/usr/bin/find BBHOME=/apps/hobbit/client DFCMD=/bin/df -Pk CONFIGCLASS=linux

CAT=/bin/cat TOUCH=/bin/touch G_BROKEN_FILENAMES=1 UNIQ=/usr/bin/uniq _=/bin/env [hobbit at hershey ext]$

Thanks for your assistance -Grs- Gregory R Shea EMC Corporation

No, it's from the script. The code you're referring to picks up the exit code from the child process that runs the script; or - if the process was terminated by a signal - then it picks up the signal number. hobbitlaunch uses the wait3() routine (line 512) to get the exit status code, and then the WIFEXITED/WEXITSTATUS macros (line 517/518) for a process that terminates normally and the WIFSIGNALED/WTERMSIG macros (line 527/528) for processes terminated by a signal. All of this comes from the child process exit code.

Linenumbers from Hobbit 4.2.0.

...

The emcpp script returns 0, temp script returns 1. The log data is from clientlaunch.log.

How did you check that? Remember to run the scripts as the Hobbit user, since access rights may be different for your account and the Hobbit account.

$BBHOME/bin/bbcmd bash ./emcpp ; echo $?

...

My standard client package includes these tests, as most of the machines here do run PowerPath and DellOM. The interesting thing about temp is my BBOSTYPE variable has an extra carriage return in it.

A bit unusual - BBOSTYPE normally comes from the "uname -s" output - but it's got nothing to do with the log messages you see.

Response in other email

...

The temp is the temperature.sh (V4.0) script from deadcat. The emcpp script returns 0, temp script returns 1. The interesting thing about temp is my BBOSTYPE variable has an extra carriage return in it. [hobbit at hershey ext]$ echo "$BBOSTYPE" linux

It seems version 4.1 of the temperature.sh script (cannot find 4.0) chokes on this, reporting

sh-3.2$ ./temperature.sh Unknown OS linux

sh-3.2$ echo $? 1

Exactly, temperature.sh checks the $BBOSTYPE variable for known OS types case "$BBOSTYPE" in solaris) get_temp_solaris ;; osf) get_temp_osf ;; linux|redhat) get_temp_omreport ;; *) echo "Unknown OS $BBOSTYPE" exit 1 esac

It should match on the linux|redhat) line but doesn't

-Grs-

...

...

sh-3.2$ ./temperature.sh Unknown OS linux

sh-3.2$ echo $? 1

Exactly, temperature.sh checks the $BBOSTYPE variable for known OS types It should match on the linux|redhat) line but doesn't

No, because "linux<newline>" doesn't match "linux". You need to figure out where that extra newline comes from; the Hobbit client "runclient.sh" script picks it up from the "uname -s" output, and that should not (normally) include a newline. So either uname is broken on these boxes, or BBOSTYPE is being set somewhere else.

In your example above, it shows the problem. The temperature.sh script gets the $BBOSTYPE from the environment, not set anywhere. Also from your example, the exit code from the script is 1. That's the same exit code I get when I run the script manually, but it exits because $BBOSTYPE don't match. Look near the bottom of the temperature.sh script. I have to assume that you're running Linux, so why does temperature.sh, exit 1 when you run it?

Whether or not I have convinced you of the above issue, how does the script terminate with a status of 5 or 7, when it returns 1?

Thanks -Grs-