Architected Clearcase Infrastructure(s) completed with HA/DR at a time IBM Rational did not support such solutions
1st project in Glen Allen, VA: Replaced all-in-one CC server, on an end-of-life host, with registry, view, and VOB servers on 3 separate hosts. The small box hosting the registry server has a redundant fail-over server. The viewserver host is the fail-over host for the VOBserver, and vice-versa, due to budget constraints. HP ServiceGuard was the tool to switch the control of the disk arrays to the fail-over host, which would resume the DNS alias of the failed host, on top of its existing aliases. Symbolic links were used to facilitate views and VOBs move, as well as disk expansion.
2nd project in McLean, VA:
Early phase: Fixing the ineffective HA/DR solution of 1 remote redundant host and establishing a NO-failback strategy, i.e. once fail-over is done, no fail back is necessary until a failure requiring a new fail-over. This is possible as access latency is minimal over the distance of 20 miles to remote location.
Final phase: Expand an infrastructure of an all-in-one server host with its redundant host into one with 6 hosts, including 1 redundant host for registry server and 4 non-redundant fail-over hosts:
3rd project in Tarrytown, NY:
For the first project: Would have used stub Unix accounts to prevent users from accessing Clearcase servers instead of the awkward restricted landing site.