|
Introduction
One piece of Mom functionality that seems to be frequently sought after is the
requirement to periodically check if Routers, Non MOM-Agent Servers or for that
matter any IP oriented devices, are up and running. Traditionally
this is achieved using a Ping operation. Mom supplies this functionality
through the use of the OpScrUtil.INet object. This is incorporated in a
script that runs as a response to a Timed Event data provider. The
scripts executes at say 2 minute intervals and reads a list of IP or FQDN
addresses to ping from a local text file. If there is no response within
the administrator specified timeout, an alert can be raised containing the
failed IP address. If the response is successful an event can be
raised. Notice that either or both responses are possible.
You can for instance de-activate the event collection on the grounds that it
creates too much data and are only interested in the failures.
Once alerts and or events are being returned to the DCAM from your ping server,
you can process them in the normal way i.e. consolidate, send emails, execute
responses, etc.
The solution comes as a Management Pack - ICMPPingList.AKM file which contains
the following:
| ICMP Ping Servers |
A Computer Group. This group has no attributes defined and so by
default does not contain any members. You can either include servers or
add your own attribute definitions. The Computer Group is associated with
the ICMP Ping Server Rules rule processing group. |
| ICMP Ping Server Rules |
A Processing Rule Group containing the rules used to generate events and
alerts |
| ICMP PingList |
A VbScript containing the ping logic |
| ICMP Ping |
A Public Views folder containing a Ping Alerts view and a
Ping Events view. |
Instructions for installation and setup
-
Decide which of your managed servers (or workstations) i.e. those that have the
MOM agent installed, will be used to perform the ping operations. You may
chose to spread the load across strategically placed servers. For
instance placing the ping functionality on a managed server in a DMZ that has
internet access allows for the development of rules to test if external
internet servers are available. Alternatively you may decide to
place the functionality on remote servers because all devices should be
accessible to those networks.
-
Create a locally available text file (by default C:\pingList.txt)
which contains a single line entry for each IP address or Fully Qualified
Domain Name that you want to ping.
-
Import the ICMP PingList Management Pack included with this
download
-
Modify the ICMP Ping Servers Computer Group definition to include
your servers.
-
Force an Agent Manager re-scan to populate the Computer Group
with your computers. You can check if this works by examining the All
Computer Groups node in the MMC and confirming that ICMP Ping Servers contains
more than zero members.
-
Modify the Event Processing Rules to suit your
requirements. For example whichever rule you chose to use you must
configure the script parameters.
| PingListFile |
Char: The full path and name for the locally stored text
file that conatins a single line entry for each valid IP or FQDN address to be
pinged. (Default is C:\PingList.txt) |
| blAlertOnFailure |
Boolean: Switches On/Off the script's instruction to
create a WARNING alert for each address in the ping list that failed to respond
for whatever reason. (Default is TRUE) |
| blEventOnSuccess |
Boolean: Switches On/Off the script's instruction to
create an INFORMATION event (ID = 4905) for each device in the ping list which
successfully responded to the ping request. Use this if you do not want
to collect data about routine success pings. (Default is TRUE) |
| RequiredResponseTime |
Integer: The maximum number of milliseconds the ping
method will wait before reporting a failed response. (Default = 1000) |
| BufferSize |
Integer: The number of bytes of data to include inside
the ping request packet. (Default = 32) |
-
Cause the rules to be updated on the targeted agents by right
mouse clicking the Rules node in the MMC and selecting 'Commit
Configuration Change'
-
Check the results in the two views provided under the Public
Views\ICMP Ping node.
TroubleShooting
The script contains no error handling because MOM does this just as
efficiently. Errors with the script such as incorrectly configured
parameters, non-existent or inaccessible pingList file and other associated Mom
failures are reported as alerts under the relevant computer. Ping
failures through non responsive addresses, non resolvable addresses, network
failures etc will appear as normal ping failure alerts in the Alerts View.
If a valid IP device is constantly being reported as unreachable
you may need to adjust the RequiredResponseTime parameter to
lengthen the timeouts. This is the same equation as running the Win2k
ping utility from the command line with the -w switch. You can experiment
with this to find the optimum setting.
A known issue is that for certain ping failures, Mom returns
the misleading error message 'Error due to lack of resources' where this is
clearly not the case. For example, it does this if the intended
target fails to respond within the requiredResponseTime parameter, instead of
simply reporting a time out error. In any case it can be ignored as the
alert is more significant than the message
To Uninstall
-
Delete the Processing Rule Group ICMP Ping Server Rules
-
Delete the Computer Group ICMP Ping Servers
-
Delete the Script ICMP PingList
-
Delete the locally stored file used for the ping list on each
server where it has been installed
-
Events and Alerts created by the ICMP PingList script will be
deleted from the database when they are old enough to be caught by the default
Mom Database Grooming tasks.
Please accept this script as is. Any feedback will be gratefully received
at Debug@Huntland.co.uk
|