Introduction

aix-logo

Despite all inroads made by Linux, commercial UNIX variants such as IBM's AIX are still used in numbers, especially when the application software stack comes from the same vendor. Cross-platform migrations can be hard to introduce, even with clear benefits. Since AIX is running exclusively on IBM's line of CPU's and hardware, it is harder to get the professional support due to the limited HW availability. As a result, OS patch management for AIX is lacking and it is a cumbersome manual process. Generating AIX patch update notifications into our standard Nagios patch monitoring promises to improve the AIX patch management.
IBM, like all major OS vendors, introduced a web-based online patch update service that simplifies the process to select and obtain the latest patches. With AIX 5L Version 5.3, the new "Service Update Management Assistant" feature SUMA was released to help automate the patching.

Plugin Design with SSH

By using the 'suma' command, we can query the IBM update server for applicable patches to our particular OS installation. The settings for this update query can be adjusted to set the usual proxy server parameters to enable access from the intranet. By processing and formatting the query output for Nagios, we can create a plugin that notifies our admins about the latest IBM patches available. It also serves to inform the security and compliance team to about outstanding patches to better judge the systems risk level.

For transporting the list of patch updates back to Nagios, we have 2 different design choices, each with it's own advantages and disadvantages:

  1. Active, remote plugin command execution using key-based SSH
  2. Passive, scheduled check result transmission through SNMP traps
First, we will look at the SSH-based remote plugin execution.

Configure and test the 'suma' command on AIX

Running the 'suma' command requires root privileges, so we either need login as root, use 'su' or, if available, use 'sudo'. 

$ sudo suma -c
Password:
        FIXSERVER_PROTOCOL=http
        DOWNLOAD_PROTOCOL=ftp
        DL_TIMEOUT_SEC=180
        DL_RETRY=1
        MAX_CONCURRENT_DOWNLOADS=5
        HTTP_PROXY=
        HTTPS_PROXY=
        FTP_PROXY=
        SCREEN_VERBOSE=LVL_INFO
        NOTIFY_VERBOSE=LVL_INFO
        LOGFILE_VERBOSE=LVL_VERBOSE
        MAXLOGSIZE_MB=1
        REMOVE_CONFLICTING_UPDATES=yes
        REMOVE_DUP_BASE_LEVELS=yes
        REMOVE_SUPERSEDE=yes
        TMPDIR=/var/suma/tmp
        FIXSERVER_URL=www14.software.ibm.com/webapp/set2/fixget

$ sudo suma -c -a HTTP_PROXY=http://192.168.100.184:80/
$ sudo suma -c -a DL_TIMEOUT_SEC=10

$ sudo suma -c |grep HTTP_PROXY
        HTTP_PROXY=http://192.168.100.184:80/

Now we are able to run a 'suma' query to list the latest patches available. See also 'man suma'. Unless we are already fully patched, we will now receive a list of patch filenames containing the fileset and latest (Version.Release.TL.Fix) patch level (example below). The *.bff file extension stands for (b)ackup (f)ormat (f)ile, IBM's unique file format the patch is delivered in. 

$ sudo suma -x -a RqType=Security -a Action=Preview
****************************************
Performing preview download.
****************************************
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/OpenGL.OpenGL_X.dev.vfb.5.3.7.1.bff
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/OpenGL.OpenGL_X.rte.base.5.3.7.1.bff
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/OpenGL.OpenGL_X.rte.soft.5.3.7.1.bff
...
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/sysmgt.websm.apps.5.3.7.1.bff
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/sysmgt.websm.rte.5.3.7.2.bff
Summary:
        217 downloaded
        0 failed
        0 skipped

Download and install the check-aix-update.pl script (for SSH)

The script check-aix-update.pl was written to run suma, compare the returned list against the list of installed filesets generated with lslpp and format the result to be usable for Nagios. It is to be placed on the AIX server and can be run directly through SSH. By testing it directly on the AIX server, we can see the Nagios output. Note the OS version, proxy setting and update URL that is additionally provided in the performance data section. 

$ sudo /scripts/check-aix-update.pl
WARNING - 211 update(s) available: X11.Dt.lib Version 5.3.7.2 X11.Dt.rte Version 5.3.7.3
X11.apps.rte Version 5.3.7.1 X11.base.lib Version 5.3.7.2 X11.base.rte Version 5.3.7.5
bos.64bit Version 5.3.7.7 bos.acct Version 5.3.7.8 bos.adt.base Version 5.3.7.3
bos.adt.debug Version 5.3.7.3 bos.adt.include Version 5.3.7.7 bos.adt.insttools Version
 5.3.7.2
...
perfagent.tools Version 5.3.7.4 printers.rte Version 5.3.7.2|OS Version 5300-07-01-0748,
 Proxy http://192.168.100.184:80/, Update-URL www14.software.ibm.com/webapp/set2/fixget

Prepare the remote plugin execution through SSH on the Nagios system

Let's talk shortly about the implications of running the update check plugin through SSH. We need to either create a passwordless SSH key or code the password into the Nagios server side configuration in order to allow for the automated plugin check execution. With the need for having root rights to execute the plugin on the remote side, AIX security now starts to depend on the Nagios system. The advantage of this setup is a "active" plugin that can be executed any time through Nagios, giving direct feedback. There is a way to avoid the security implications of the SSH-based plugin execution: using a passive check setup that is triggered on the AIX server itself. More about this further down, first the SSH-based setup: 

nagios@susie112:~> ssh-keygen -t dsa -b 2048 -f .ssh/nagios-key
Generating public/private dsa key pair.
Enter passphrase (empty for no passphrase):
Enter same passphrase again:
Your identification has been saved in .ssh/nagios-key.
Your public key has been saved in .ssh/nagios-key.pub.
The key fingerprint is:
af:31:12:5a:d1:13:bb:78:56:19:4f:a1:68:89:9d:f3 nagios@susie112

copy .ssh/nagios-key.pub to user@192.168.245.129:/home/user/.ssh/authorized_keys
(chmod to 600 for rw-------)

Now we can test the remote plugin execution from the Nagios server: 

nagios@susie112:~> ssh -i .ssh/nagios-key user@aixserver3 "sudo /scripts/check-aix-update.pl"
WARNING - 210 update(s) available: X11.Dt.lib Version 5.3.7.2 X11.Dt.rte Version 5.3.7.3
X11.apps.rte Version 5.3.7.1 X11.base.lib Version 5.3.7.2 X11.base.rte Version 5.3.7.5
bos.64bit Version 5.3.7.7 bos.acct Version 5.3.7.8 bos.adt.base Version 5.3.7.3
bos.adt.debug Version 5.3.7.3 bos.adt.include Version 5.3.7.7 bos.adt.insttools Version
 5.3.7.2
...
perfagent.tools Version 5.3.7.4 printers.rte Version 5.3.7.2|OS Version 5300-07-01-0748,
 Proxy http://192.168.100.184:80/, Update-URL www14.software.ibm.com/webapp/set2/fixget

We are creating and testing the nagios plugin to let Nagios run the SSH command: 

nagios@susie112:~> vi /srv/app/nagios/libexec/check_aix_patchlevel.sh
#/bin/sh
/usr/bin/ssh -o StrictHostKeyChecking=no -i /srv/app/nagios/.ssh/nagios-key \
-l aixhost-user $1 "sudo /scripts/check-aix-update.pl"

nagios@susie112:~> /srv/app/nagios/libexec/check_aix_patchlevel.sh 192.168.245.129
WARNING - 210 update(s) available: X11.Dt.lib Version 5.3.7.2 X11.Dt.rte Version 5.3.7.3
X11.apps.rte Version 5.3.7.1 X11.base.lib Version 5.3.7.2 X11.base.rte Version 5.3.7.5
...
perfagent.tools Version 5.3.7.4 printers.rte Version 5.3.7.2|OS Version 5300-07-01-0748,
 Proxy http://192.168.100.184:80/, Update-URL www14.software.ibm.com/webapp/set2/fixget

Nagios service configuration

After creating the SSH plugin script, we add it to commands.cfg and create our AIX patch service configuration file. 

susie112:~ # vi /srv/app/nagios/etc/objects/command.cfg
# 'define check_aix_patch'
define command{
  command_name check_aix_patchlevel
  command_line $USER1$/check_aix_patchlevel.sh $HOSTADDRESS$
}

nagios@susie112:~> vi /srv/app/nagios/etc/objects/aix-patch-services.cfg
###############################################################################
# Define a servicegroup for AIX patch service checks
# All AIX patch service checks will be members of this group
###############################################################################
define servicegroup{
  servicegroup_name        aix-patch-checks  ; The name of the servicegroup
  alias                    AIX Update Checks ; Long name of the group
}
###############################################################################
# Define the generic patch check service template
###############################################################################
define service{
  name                          generic-patch-aix
  active_checks_enabled         1
  passive_checks_enabled        1
  parallelize_check             1
  obsess_over_service           1
  check_freshness               0
  notifications_enabled         1
  event_handler_enabled         1
  flap_detection_enabled        1
  failure_prediction_enabled    1
  process_perf_data             1
  retain_status_information     1
  retain_nonstatus_information  1
  is_volatile                   0
  check_period                  24x7
  max_check_attempts            3
  normal_check_interval         720             ; check only once every 12 hours
  retry_check_interval          2
  contact_groups                aix-admins, security-team
  notification_options          u,w,r		; notify unknown, warn, recovery
  notification_interval         1440            ; notify only once per day
  notification_period           24x7
  register                      0
  servicegroups                 aix-patch-checks
}
###############################################################################
# AIX Patch Update Check via SSH - AIXHOSTS 1-3
###############################################################################
define service {
  use                           generic-patch-aix
  host_name                     aixserver1, aixserver2, aixserver3
  service_description           check_aix_patch
  check_command                 check_aix_patchlevel
}

Running the AIX patch check notification through SNMP traps

By setting up a passive check through SNMP traps, we avoid the security implications of configuring SSH user access to each AIX host. Instead, we need to change the check-aix-update.pl to send its data through a SNMP trap. AIX comes with the 'snmptrap' command, so instead of sending the update information to STDOUT, we call this trap program. The new plugin is named aix_update_trapsend.pl and it is configured to run on the AIX hosts once a day through cron. On the top of the perl script are the settings for the SNMP server and trap community to be used by 'snmptrap'. On the Nagios side, I needed to update /etc/snmp/snmptrapd.conf to accept and handle the AIX patch update traps. I also needed to update to the Nagios traphandler program send_trap_data.pl with a new filter that is processing, formatting and injecting the new AIX patch trap messages into Nagios.

Plugin Design with SNMP Traps

susie112:/ # vi /etc/snmp/snmptrapd.conf
# capture cron-scheduled patch update notification from AIX:
# SNMPv2-MIB::snmpTrapOID.0 = RFC1155-SMI::enterprises.2.0.1
traphandle   RFC1155-SMI::enterprises.2.0.1  /srv/app/nagios/libexec/send_trap_data.pl

Nagios configuration with SNMP trap setup

Here the Nagios configuration for the same AIX patch check configured as SNMP trap: 

nagios@susie112:~> vi /srv/app/nagios/etc/objects/aix-patch-services.cfg
###############################################################################
# Define the aix patch trap service template
###############################################################################
define service{
  name                          generic-patch-aixtrap
  active_checks_enabled         0               ; traps are only passive checks
  passive_checks_enabled        1               ; yes, check passive
  parallelize_check             1               ; yes, please
  obsess_over_service           0               ; we don't run extra commands
  check_freshness               1               ; check if a report came in
  freshness_threshold           93600           ; 26 hour threshold, since
                                                ; check runs only once a day
  check_command                 no-patch-report ; this command is run only if
                                                ; service results are "stale"
  notifications_enabled         1               ; send notifications
  event_handler_enabled         1               ; yes, but we have none
  flap_detection_enabled        0               ; with daily updates, no need
  failure_prediction_enabled    1               ; dependency checks
  process_perf_data             0               ; don't send this to perfdata
  retain_status_information     1               ; yes, we keep it
  retain_nonstatus_information  1
  is_volatile                   1               ; enable for passive checks
  check_period                  24x7            ; always check for submissions
  max_check_attempts            1               ; one trap is enough
  normal_check_interval         1
  retry_check_interval          1
  contact_groups                aix-admins, security-team
  notification_options          w,r             ; notify warnings and recovery
  notification_interval         1440            ; notify once a day
  notification_period           24x7            ; always notify
  register                      0               ; template, don't register
  servicegroups                 aix-patch-checks
}
###############################################################################
# Receive SNMP traps for AIX update notifications from hosts withouth SSH
###############################################################################
define service {
  use                           generic-patch-aixtrap
  host_name                     aixserver1, aixserver2, aixserver3
  service_description           check_aix_patchtrap
}

Comments

AIXTHOST3@[/home/myuser]> sudo /usr/sbin/suma -x -a RqType=Security -a Action=Preview \
| grep -v 5.3.7
Password:
****************************************
Performing preview download.
****************************************
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/devices.chrp.base.ServiceRM.1.3.0.72.bff
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/invscout.rte.2.2.0.12.bff
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/perl.rte.5.8.2.71.bff
Download SUCCEEDED: /usr/sys/inst.images/installp/ppc/radius.base.rte.2.1.7.1.bff

I could not verify yet why these IBM package versions differ from the typical AIX level numbering (i.e. 5.3.TL.SP) and if it is OK to simply include them in normal patch notifications. If you happen to know more, please drop me a line - public [at] frank4dd [dot] com. I would be happy to simply trust 'suma' and avoid the additional check against 'lslpp'-reported inventory.

Credits, copyrights original scripts and links

Topics:

More Information:

Screenshots

AIX patch service group AIX patch service detail2 AIX patch service hosts AIX patch service detail AIX patch service detail AIX patch notification