What does alert "xx% of background dump area used" in oracle OEM 10g? 1

[saywatt]

I have these messages on my oem console that 96% of background dump area is used. The same thing for core dump area and user dump area. I cleaned out the directories and still get the 96%. What am I not getting here? Do I need to take some other action than clearing out those directories? Do I need to restart the agent? I cannot get these high numbers to go away. And, I have plenty of space. I saw a previous forum on this question and did not see any answer that addresses my concerns. Please advise what the 96% is and how to get it down to where it does not cross my metric threshold so it does not appear as an alert (other than raising the threshold). Thank you in advance

 

[karluk]

I guess I would not rule out the possibility that you are getting a misleading alert from OEM, but personally I have found alerts concerning directories filling up to be quite reliable. So I will make the working assumption that the alert means exactly what it says - the mount point where your background dump, core dump and user dump directories are located is 96% full. But note that this refers to the entire mount point, not just to the individual directories. Deleting log and trace files from these directories doesn't always address the underlying problem, because typically log and trace files are small and comprise only a small fraction of the space used on what's likely to be a large mount point. You need to look at the mount point as a whole and see if there are any large files you can delete or move off of the mount point. Probably some clean-up will be possible, but eventually you may be forced to add some disk to increase the size of the mount point. It all depends on how much clean-up is realistic.

 

[saywatt]

Karluk, thank you for the quick reply. I was unaware that the number referred to everything under the mount point. I learned something. But, I have 70GB for that mount point. When I do a df -k, it tells me that 68% is used. I guess this is what is confusing me now.

 

[karluk]

The 68% used reported by df -k does, of course, contradict the alert you are getting from OEM. If you are certain that this is correct, I would suggest opening a Service Request with Oracle support to address an apparent bug in OEM.