Simple shell script to log system status

[p0000h]

Sorry in advance for this but I am totally new to Unix scripting and would appreciate some help. Working mainly on Network routers and switches, I am having trouble putting together a little script to run periodically outputting the disk, cpu and memory usage of a unix host. The host is a Nokia firewall but it uses a cut down BSD derived unix OS.

I would appreciate if anyone can post how I would actually get something like this to run say with cron to have 3 or so commands eg.. /usr/bin/uptime, /usr/bin/vmstat 1 4, /bin/dk - k every minute and log to a local file or a syslog server.

I know I can send all logs to a syslog server no problems but to get thes 3 specific commands to run would allow me to troubleshoot a recurring problem on the host. The limitations of the job pervent me from using snmp.

Regards,

Greig (user: p0000h)

 

[pmcmicha]

With your examples your already half way there...

Try this:

#!/usr/bin/ksh -p

# This script will gather server info.
# Setup Variables here. (Add more if needed)
LOG="LOG_DIR_HERE/LOG_FILE_NAME_HERE"

# Setup functions here. (Add more if needed)
MTIME() { # This captures the time using military std.
print "`/usr/bin/date \"+%T\"`"
}

# Start log file.
exec 1>> ${LOG} # Capture stdout
exec 2>> ${LOG} # Capture stderr

# Now run command for system info.
MTIME
/usr/bin/uptime

MTIME
/usr/bin/vmstat 1 4

MTIME
/bin/dk -k

#EOS


Next setup cron to run this script every minute.

 

[p0000h]

Thank you very much for the reply, it will provide me with a starting point for trying to understand the mechanics of the script.

Regards,

Greig (user: p0000h)