We are using ADDM 8.3.2 with May 2012 TKU. Cocerning Apache Web Server, we are getting apx 1/3 that have no version details. Apache Tomcat we are missing version detail on apx 15%. This is a pretty high percentage. Wondering if anyone has encountered this issue or might no a fix.

Many Thanks
Mike

Our online documentation (Apache HTTP Server, Apache Tomcat) describes in some detail how the patterns attempt to obtain the information.

These products are also often embedded in things which sometimes means that some of the ways the patterns attempt to obtain the version may not work.
The most common reasons outside of this are that either ADDM user hasn’t got enough permissions to run certain commands or access certain files (via the patterns), or that ADDM could not obtain enough information about the environment (e.g. paths where the processes are running from) to allow the patterns to gather this information.

Discovery Access page showing the commands executed and results will be of help in troubleshooting the former. Looking at the primary inference for the SI without version information may point you to the latter.

Thank you again Nikola. Can I assume that Adobe ColdFusion Server might also have the same issues?

Our ADDM admin running the discovery should have sufficient permissions to run all commands. I am wondering if one of the issues concerning ADDM not obtaining enough information about the environment (e.g. paths where the processes are running from) to allow the patterns to gather this information…Is this related to communications between servers. In other words, if ADDM is scanning that server and there is no communication going at that time would ADDM miss that? Would more scans likely catch it?

Mike

Apache Web Server and Apache Tomcat pattern may run a command or try to obtain a file in order to get the version information. For this they need to know the location the binary is running from (in case of Apache HTTPD)) or the location the product is installed (in case of Tomcat). This information is obtained from the trigger process command-line.
If the information is not available, e.g. the Apache web server process in the process listing on a host appears as just ‘./httpd’ instead of ‘/opt/foo/bar/httpd’, then the patterns cannot run those commands, or get access to a file because they do not know where to look for it. ADDM does not trawl the file system as this would potentially increase significantly the load on the target system, and would most certainly decrease the discovery performance.

Hello Nikola. Hope you had a great 4th!. I have been pulled in some other directions lately, just now getting back to the fun stuff. I have been looking at BMC site: Adding privileged execution to commands”. I followed the instructions to take a look at what we are doing in ADDM PRD. The web page indicates ADDM out of the box has no commands using privileged execution. It provides an example of the PRIV_LSOF command that should be there.
…
PRIV_LSOF() { sudo “$@”
}
… What I see when I open the lsof script:

temp_lc=$LC_ALL
LC_ALL=C
export LC_ALL
if [ `uname -r | cut -d. -f2` -lt 7 ]; then PRIV_LSOF lsof -l -n -P -F ptPTn -C -i 2>/dev/null
else PRIV_LSOF lsof -l -n -P -F ptPTn -i 2>/dev/null
fi
LC_ALL=$temp_lc
export LC_ALL

The command above running sudo is not here. My question is can this effect pulling the versioning detail from the UNIX servers?

It could. You can see all the commands the patterns tried to run (and their output) on a host via the Discovery Access for the host.

In order for the pattern to run a command using increased privileges it will have to have been coded to make use of PRIV_CMD which you then update (like PRIV_LSOF) for the appropriate platform in the ADDM platform scripts and enable its use in the pattern via the configuration UI for that pattern.