diff options
author | thrall <thrall@ae88bc3d-4319-0410-8dbf-d08b4c9d3795> | 2004-06-26 19:54:02 +0000 |
---|---|---|
committer | thrall <thrall@ae88bc3d-4319-0410-8dbf-d08b4c9d3795> | 2004-06-26 19:54:02 +0000 |
commit | bbfbd908070a0639e5218692c1e6ab1ded4a6a7d (patch) | |
tree | ff2d7b78cb9e0bde67b36e199ca775e97848238b | |
parent | a3385ebdac8dc1c1f86e201faf9e8555f8d1e177 (diff) | |
download | ATCD-bbfbd908070a0639e5218692c1e6ab1ded4a6a7d.tar.gz |
Added readme describing how to run
-rw-r--r-- | TAO/orbsvcs/examples/RtEC/Federated_Kokyu/dynamic_topology_test/README | 30 |
1 files changed, 30 insertions, 0 deletions
diff --git a/TAO/orbsvcs/examples/RtEC/Federated_Kokyu/dynamic_topology_test/README b/TAO/orbsvcs/examples/RtEC/Federated_Kokyu/dynamic_topology_test/README new file mode 100644 index 00000000000..0eeadf07ee6 --- /dev/null +++ b/TAO/orbsvcs/examples/RtEC/Federated_Kokyu/dynamic_topology_test/README @@ -0,0 +1,30 @@ +I've got the test set up like so: A single Client (Supplier) with two +modes: NORMAL and FAULT_TOLERANT (FT). Every timeout, it does work then, +when in NORMAL mode, it pushes events to only the Middle1 process; when +in FT mode, it pushes events to both Middle processes. Currently, I have +it start in NORMAL and switch to FT after the first time the task +executes. There are 2 middle processes, Middle1 and Middle2. They simply +receive events, do work, and, based on the mode of the Client (they can +tell from the event type), push events in the same mode to the Server. +The Server subscribes to events from both Middle processes. If the +events indicate NORMAL mode, the Server performs work every time it +receives an event. If the events indicate FT mode, the Server performs +work only after it receives an FT event from both the Middle1 and the +Middle2 processes. + +Thus, we have a single task with 3 subtasks: T1_1 is executed in the +Client. T1_2 is executed in Middle1 and if in FT mode in Middle2. T1_3 +is executed in the Server when Middle1 is done (in NORMAL mode) or when +both Middle1 and Middle2 are done (in FT mode). + +Here's how to run it: I have provided scripts (Client, Server, Middle1, +and Middle2) to run the processes. You need to run the Client first, +then the Middle processes (any order), then the Server. The Middle +processes run their ORB for 305 seconds and the Server runs its ORB for +310 seconds to accommodate the delay between starting each process. + +Here's were to run it on the cluster: +Server (Ron) +Middle1 (Harry) +Middle2 (Hermione) +Client (Norbert) |