We have several Linux processes implemented in various technologies, Java, C++, etc. They interact with each other by passing messages on Websphere MQ. If any process crashes, we would like it to be restarted automatically for a configured number of times.
Would it involve a change in the applications, such as periodically raising a heartbeat to indicate that the application is in good health?
Thanks,
Yash
At my previous job we had a similar problem. We developed our own solution. We implemented a watcher program in two different technologies: one in Java and one in C++ using QT. Each had a list of programs to watch. For every program that the watcher watch we had maximum time between two heartbeats, what program to run on every heartbeat and what program to run if the time between two heartbeats maxed out. Watcher in Java had an entry for the watcher in C++ and vise-versa.
