Use Nagios to Check Your Zypper

If you use Nagios to monitor your system and run openSUSE on a remote server the bash script presented here will check for online updates and is designed to be run by Nagios so that the result will appear on the Nagios service-detail page.

The script is pretty unsophisticated as it just parses the output from the zypper command. A more sophsiticated solution might interact directly with libzypp, the library that provides zypper with its functionality. Of course, that's not possible using bash. Based on a quick scan of the libzypp documentation it appears the only current option for doing that is C++.

A Nagios plugin script outputs information in two ways:

  • It writes a short status message (one line of text) to the standard output which Nagios displays on the service-detail page.
  • It sets its exit status to indicate to Nagios the status of the service:
    • 0 if the service status is "OK"
    • 1 if the service status is "WARNING"
    • 2 if the service status is "CRITICAL"

If you run zypper from the command line you will see output similar to the following:

  $ sudo zypper list-updates
  * Reading repository 'openSUSE-10.3-Updates' cache
  * Reading repository 'Main Repository (OSS)' cache
  * Reading repository 'Main Repository (NON-OSS)' cache
  * Reading repository 'Packman Repository' cache
  * Reading installed packages [100%]

  Repository:           | Name             | Version | Category | Status
  ----------------------+------------------+---------+----------+-------
  openSUSE-10.3-Updates | dhcpcd           | 5390-0  | optional | Needed
  openSUSE-10.3-Updates | openmotif22-libs | 4540-0  | optional | Needed
  openSUSE-10.3-Updates | ruby             | 5483-0  | security | Needed

The script looks for the "----" line and then breaks the subsequent lines into fields, using the "|" as the separator character. Using this data it then builds the output text and sets the status. The status is set to 0 if there are no updates available, 1 if there are updates available, and 2 if any of the updates are security related.

Running zypper can take a while so you can run the script in test mode to get some immediate satisfaction. In test mode it contains some test zypper output which it parses:

  $ env TEST=1 sh check_zypper
  3 Updates needed:  package-1(required/Needed), package-3(required/Needed), package-4(required/Needed) (1 Optional update)
  $ echo $?
  1
The output text is what Nagios would display on the service-detail page. Nagios will set the service status to "WARNING" since the script status on exit is 1 (meaning updates are availabe).

The script follows:

#!/bin/bash

max_packages=3
TEST=${TEST:-0}
VERBOSE=${VERBOSE:-0}
nl='
'

# Read list of packages that need updating according to zypper.
OIFS=$IFS
IFS=$nl
if [[ $TEST -eq 0 ]]; then
    zypper_lines=($(/usr/bin/sudo /usr/bin/zypper list-updates 2>&1))
    stat=$?
    if [[ $stat -ne 0 ]]; then
        echo "Zypper exited with code: $stat"
        exit 1
    fi
else
    zypper_lines=(
        "test line 1"
        "test line 2"
        "------------"
        "Main Repository | package-1 | 50-0  | required | Needed"
        "Main Repository | package-2 | 48-0  | optional | Needed"
        "Test Repository | package-3 | 40-0  | required | Needed"
        "Test Repository | package-4 | 40-0  | required | Needed"
        )
fi
IFS=$OIFS

# Count the number of optional and non-optional packages.
npackages=0
noptional_packages=0
nsecurity_packages=0
in_packages=0

for ix in ${!zypper_lines[*]}
do
    line=${zypper_lines[$ix]}
    if [[ $VERBOSE -ne 0 ]]; then echo $line; fi

    if [[ $in_packages -eq 0 ]]; then
        if [[ $line =~ ^---- ]]; then in_packages=1; fi
    else
        IFS='|'
        set -- $line
        IFS=$OIFS
        if [[ $# -eq 5 ]]; then
            trepo=$(echo $1)
            tpackage=$(echo $2)
            tversion=$(echo $3)
            tcategory=$(echo $4)
            tstatus=$(echo $5)
            
            if [[ "$tcategory" == 'optional' ]]; then
                let noptional_packages++
            else
                repo[$npackages]=$trepo
                package[$npackages]=$tpackage
                version[$npackages]=$tversion
                category[$npackages]=$tcategory
                status[$npackages]=$tstatus
                let npackages++
                if [[ "$tcategory" == 'security' ]]; then let nsecurity_packages++; fi
            fi
        fi
    fi
done
        
# Output summary.
if [[ $npackages -ne 1 ]]; then s1='s'; else s1='' ; fi
if [[ $noptional_packages -ne 1 ]]; then s2='s'; else s2='' ; fi

n=0
for ix in ${!package[*]}
do
    t[$n]="${package[$ix]}(${category[$ix]}/${status[$ix]})"
    let n++
    if [[ ix -eq $max_packages ]]; then break; fi
done

if [[ ${#package[*]} -gt $max_packages ]]; then t[$n]="..."; fi

if [[ ${#t[*]} -gt 0 ]]; then
    if [[ $npackages -gt 0 ]]; then
        echo -n "$npackages Update$s1 needed: "
        first=1
        for ix in ${!t[*]}
        do
            if [[ $first -eq 1 ]]; then
                echo -n ' '
                first=0
            else
                echo -n ', '
            fi
            echo -n "${t[$ix]}"
        done
        if [[ $noptional_packages -gt 0 ]]; then
            echo -n " ($noptional_packages Optional update$s2)"
        fi
    else
        echo -n "$noptional_packages Optional update$s2"
    fi
    echo
    stat=1
    if [[ $nsecurity_packages -gt 0 ]]; then stat=2; fi
else
    echo "OK, no package updates available"
    stat=0
fi

exit $stat

There are a few things you need to do to integrate the script with Nagios. First setup sudo to allow the Nagios user to run zypper. Use visudo to add this to the sudoers file:

   nagios    ALL=(ALL) NOPASSWD:/usr/bin/zypper
Second, because running zypper takes a while you need to increase the service-run-time in the Nagios configuration file /etc/nagios/nagios.cfg:
  service_check_timeout=600
And lastly, of course, you need to setup Nagios to run the script:

define command{
        command_name    check-zypper
        command_line    /path/to/check_zypper
        }

...

define service{
        use                             generic-service
        host_name                       localhost
        service_description             Updates
        is_volatile                     0
        check_period                    24x7
        max_check_attempts              1
        normal_check_interval           600
        retry_check_interval            60
        contact_groups                  admins
        notifications_enabled           0
        notification_options            c
        notification_interval           960
        notification_period             24x7
        check_command                   check-zypper
        }
Among other things, this tells Nagios to run check_zypper every 10 hours and if it fails, to retry in an hour.

p.s. Yeah, the title is pretty "cheap humor", but then again I'm not the one who decided to call my package manager zypper.

Mitch Frazier is an embedded systems programmer at Emerson Electric Co. Mitch has been a contributor to and a friend of Linux Journal since the early 2000s.

Load Disqus comments