#3 I put the information in the Troubleshooting section. Do not hesitate to reformat. #2 And I created the jira https://issues.apache.org/jira/browse/HBASE-1214
About HBASE-616, using the concurrent GC should help, the only problem is that the concurrent GC performs full GC sometimes, it is therefore not a solution and I don't think it should be the normal mode of HBase. Another idea (haven't tried it yet) is to have a larger part of the heap used for Eden space (short life objects), in order to avoid mid-living object to be promoted to the old space, and collected at full gc time. Anyway, if the heap almost fits in physical memory, all gc strategies should be quick. I'll continue to monitor the heap usage, and if I find something interesting, I post. stack-3 wrote: > > Thanks also for the considered message below J-A. I echo J-D's suggestion > that the #3 note belongs on the troubleshooting page. Your suggestion > would > help with, https://issues.apache.org/jira/browse/HBASE-616, IIUC? If you > agree I can add it, no problem. Will follow up on the other items in a > while. > St.Ack > > > > -- View this message in context: http://www.nabble.com/HBase-0.19.0%3A-Xcievers---shutdown---JVM-GC-settings-tp22158436p22180687.html Sent from the HBase User mailing list archive at Nabble.com.
