D
dbeavon
Guest
If it is crashing then you may have event logs (application log in windows) to back that up. Windows should detect that the _mproapsv has crashed and write a message to application log. At least that will keep track of how frequently this happens. But despite the crash, PASOE should recover and start a new msagent (_mproapsv) to replace the one that crashed. Hopefully your customer can attempt to repeat the PASOE request if they see an unusual error message. And you should also have protrace files that Progress can examine. Sometimes those files can be sufficient to identify the problem, if you give them to tech support (... the ABL stack and/or the native stack trace can be helpful clues). The _mproapsv shouldn't crash ... we still have some issues with leaks, and sometimes with concurrency restrictions in the process, but never any unusual crashes (or at least none that would happen in _mproapsv and not also in _progres). We've been running PASOE for about a year now. The only times I recall any surprising crashes were when my r-code was stale (just delete all the old stuff and replace it with freshly compiled code), and when DEP was enabled.
Continue reading...
Continue reading...