This content has been marked as final. Show 1 reply
Well, even if the database version is different and the operating system is different, I cannot hide the fact that this issue looks so similar to an old issue on Oracle XE10g.
Although I managed to mitigate the problem over time by adding multiple dispatchers, it never went completely away, hence, from time to time, I need to kill the linux process that goes nuts.
My personal feeling is that there is some resource (either a buffer or a counter or whatever) that constantly builds up until it blows up, causing this odd CPU spike.
Interestingly enough in the case of Oracle XE10g this doesn't mean that the database become slow or unresponsive, on the contrary, it works normally (provided you have more than one dispatcher process otherwise it stops responding), but the CPU meters report that the CPU is spinning at 100%.
If you manage to find out what is causing it, I'll be most glad to hear from you, may be there is something that can be done in order to prevent such occurrence.