We have ganglia on IBM p55 (2x2 CPU) installed (8 lpars, aix5.3). In ganglia this machine is defined as ganglia cluster, lpars are nodes. We use power awere gmond's (3.0.7) and gmetad (3.0.7) is installed one of lpars.
It works well, but I have doubts about its usefulness on lpars, especially in CPU monitoring. Metrics like cpu_[user|wait|system|idle] and graph "CPU last our" are wrong both in shared and dedicated environment. I can understand why it's wrong in shared CPU. But I don't know why is that with dedicated CPU ? I can see different values with ganglia and nmon. I.e: LPAR1 : 2 dedicated CPU, usage around 30% of (2) CPU for 2 hours (topas). Ganglia (CPU last hour) shows : 15 % nmon (physical avg) : 25 % - which is true I think . 1. Why is that ? 2. "CPU total" shows 32 . I think it shows 8 (lpar) x 4, but this is not the real number of CPU. Can I change it ? 3. Which metrics do you monitor in shared CPU environment ? Which are reliable ? regards ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://sourceforge.net/services/buy/index.php _______________________________________________ Ganglia-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-general
