Article Figure 1 Figure 2 Figure 3 Figure 4
Figure 5 Figure 6 Figure 7 Figure 8 Figure 9
Figure 10 Sidebar Table 1 Table 2 Table 3
Table 4 Table 5 Table 6 sep2004.tar

Problem Profile

Managing software in the Solaris environment is easy when dealing with a single server or two. When you have more than 400 servers, however, it becomes impossible. A big gap in open enterprise Unix administration is the ability to manage/monitor software from a centralized location. Administrators need to know their environments and the state of their environments.

Salt Lake 2002 Problem

Scenario 1: System constantly moving from test environments to sporting venues (e.g., World Cup Matches prior to Olympics).

Scenario 2: Software distribution consisted of scripts and rsyncing directories.

  • Unix systems were frequently never up all at once so the push process failed.
  • Constant danger of updates being inconsistent; system has no knowledge of what is correct.
  • No tracking capabilities when syncing directories.
Solution

We needed a centralized baseline in which an agent installed on the machine locally could query the systems inventory using vendors packaging format and then compare against global inventory and determine status.