In some cases the TEPS navigator will hang on the “more” nodes, I haven’t bothered trying to run down the how and why, I just want to see all child-nodes when I select a parent-node.

Here’s how you do it…

Add the following lines to the teps.jnlpt, then rebuild the teps.jnlp

<property name=”cnp.http.url.host” value=”$HOST$”/>
<property name=”cnp.navigator.branch.pagesize” value=”1000″/>
<property name=”cnp.navigator.branch.threshold” value=”500″/>

If your new to or just need to learn more about AWS.. this is a great slide presentation I stumbled across

As of ITM 6.2.1, there is a perceived capability to now rename situations within the editor.

However, the rename is actually a new attribute called ‘Full Name’ and only changes the TEPS display name for the situation; the EIF will still send the original situation name.

IBM responds this is “working as designed” — a load of crap if you ask me and a “feature” that should not have been introduced at all if it were to be so counter-intuitive.

Here’s IBM’s response:

Problem(Abstract)
Situation names that are less than 32 characters in length do not populate the “Fullname” field when running the ‘tacmd viewsit -s ‘ command.

Symptom
tacmd viewsit shows null Fullname Field

Cause
The Fullname field is only populated when the original situation name is 32 characters or greater or if a situation name is changed.

Resolving the problem
When a situation is created with a name less than 32 characters in length, the situation name is stored in the “Name” column. At this point, the “Name” column serves two purposes: a unique ID for the situation and the situation display name. The “Fullname” column is blank.

If an existing situation name is changed, the “Name” column still serves as the unique ID and will not be updated to reflect the situation name change.

The new situation name is stored in the “Fullname” column, the new name now assumes the role of situation display name. The length of the new name does not matter, the new name must be stored in the “Fullname” column, since the “Name” column containing the unique ID cannot be changed.

The only time we should expect the “Full Name” field to be populated is when the original situation name is 32 characters or greater or when a situation name has changed from it’s original name (regardless of the length of the new situation name).

Situation

An ITM6 Unix (UX) or Linux (LZ) agent detects that its remote filesystems are unavailable, a “stale” connection.

Automate an action to remount it if possible

Methodology

Two (2) ITM situations are created where

1. detects the mount point has become stale
   
2. the other is triggered by the “correlated situation” condition of #1 being tr
   

Situation #1 : Detect “Stale” remote FS

Simple enough – if “Space Available” fails collection, there is an issue

Situation #2 : Remount Situation

Formula Conditions

Use the “Situation Comparison” for a condition against the situation above

Take Action

Note: the lines are strung together on one line in the Take Action field, to make it more legible here, newlines are after the semicolons

f=”&{Linux_Disk.Mount_Point}” ;
u=`umount -f $f 2>&1 && echo $f`;
m=`mount $f 2>&1 && echo $f`;
echo -e “umount: $u\nmount:$m” | mail -s “ITM ACTION: Remount $f” junkmail@JdsMedia.net

ITM v6 log files use a hexadecimal timestamp (to save space? who the hell knows), which adds unnecessary effort when the reason you’re looking at the logs is to determine an issue in the first place. In any case… Here’s the script I wrote when I first encountered the nonsense in ITM v6 logs a few years ago:

#!/bin/perl
foreach (<STDIN>) {
if (/^[^\s\d\w]+([\w\d]*)/) {
@t=localtime(hex($1));
$time=sprintf(“%02d:%02d:%02d %02d/%02d/%04d”,
$t[2],$t[1],$t[0],$t[4]+1,$t[3],$t[5]+1900);s/^[^\s\w\d]+[\w\d]*/$time/;}print $_;}

Here’s a one-liner that Venkat.Saranathan at Gulfsoft.com cranked out, rendering my script pretty much obsolete

perl -lane ‘if ($_ =\ /^(.)([\dA-F]+)(\..*)/) { printf “%s%s%s”, $1, scalar(localtime(oct(“0x$2″))),$3; }’

In a nutshell: use the ‘eventcreate‘

I had an issue with ITM Windows Agent creating events which I was then sending out email alerts via a UNIX script.

The simple problem was backslashes in windows paths where the “Take Action” notification script was literalizing characters which might be part of the EventLog Description

For example:

"c:\Damned these\nagging\slashes\argh"

became

c:\Damned these
agging\slashesrgh

Yes, simply quoting my string would have solved most of my problem, IT DID HOWEVER lead me to a handy windows command for testing, called ‘eventcreateid‘

Example : eventcreate /ID 101 /SO FooBar /D “Test Message for ITM” /T ERROR /L Application

My PMR to IBM Tivoli support

I’d be surprised if there wasn’t already an ER for this, but if you
don’t mind, please submit one.

Do you know if anybody else has asked or criticized this major
limitation.

Its hard to believe that there isn’t something that emulate what a
simple one line script would accomplish.

In the ER, I would expect support for perl regex pattern, where patterns
can be indexed to return the matching index as a displayable value.

match process string against /opt/IBM*\s+(\S+java*\)s+server1 pattern and return pattern matching (\S+java*\)

IBM’s Response

Yes, we already have an ER opened for this case, please note down the
request number: MR0108092213

****

ITM 6 has limitations when monitoring processes.

Customer needs to monitor processes with very long names, in this case
the WebSphere NodeAgent process with 4999 characters. The unique string
“nodeagent” is at the very end of the process name and the Unix OS agent
cannot find this process: The attribute command_name is limited to 100
characters and for command_name_U to 768 Bytes.
Cust needs a possibility to monitor such long process names using the
standard ITM 6 OS agent.

****

It is already a recognized request. You can follow up on the progress of
the request with your IBM Rep. Please let me know if you have any other
questions.

Obtaining Product Codes for IBM Tivoli Monitoring (ITM v6)

Local method:

Parse the proddsc.tbl file on a UNIX/Linux system to get the list by doing this:

UNIX
awk -F\| '/^[^*#]/ {print $1,$2}' ${CANDLEHOME}/registry/proddsc.tbl | sort | uniq


Or, IBM’s site http://www-01.ibm.com/support/docview.wss?rs=2366&context=SSZ8F3&dc=DB520&dc=DB560&uid=swg21265222&loc=en_US&cs=UTF-8&lang=en&rss=ct2366tivoli

ITM v6 log files use a hexadecimal timestamp (to save space? who the hell knows), which adds unnecessary effort when the reason you’re looking at the logs is to determine an issue in the first place. In any case…

Here’s the script I wrote when I first encountered the nonsense in ITM v6 logs a few years ago:

#!/bin/perl

foreach (<STDIN>) {
 if (/^[^\s\d\w]+([\w\d]*)/) {
  @t=localtime(hex($1));
  $time=sprintf("%02d:%02d:%02d %02d/%02d/%04d",
  $t[2],$t[1],$t[0],$t[4]+1,$t[3],$t[5]+1900);
  s/^[^\s\w\d]+[\w\d]*/$time/;
 }
  print $_;
}

Here’s a one-liner that Venkat.Saranathan at Gulfsoft.com cranked out, rending my script pretty much obsolete

perl -lane 'if ($_ =\ /^(.)([\dA-F]+)(\..*)/) { printf "%s%s%s", $1, scalar(localtime(oct("0x$2"))),$3; }'