Strange G1 behavior

[Kirk Pepperdine]

Hi Bernd,


> On Oct 17, 2017, at 12:12 PM, Bernd Eckenfels <ecki at zusammenkunft.net> wrote:
> 
> Hello,
> 
> First of all I wanted to mention as I understand it the application is a graph query server,

I think it’s a triple store.

> therefore freeing and creating of large graphs is the workload the GC would have to deal with. (I think Thomas described that allocation pattern - it would lead to spiky promotions if a query transaction lasts longer then a new collection cycle (or multiple))

Right, unfortunately we’ve gotten fixated on this particular log where as this is a pattern I’ve seen across many JVMs running in a number of different domains that are not SparQL. Since these are from older JVMs (8 and 7) I didn’t post them as I assumed with the rate of change in G1 that it would be difficult to pin the behavior on a pathological cases. I will try to dig out these older logs if you think it will help.
> 
> And in Addition, Thomas did you removed your graph or is Java.net <http://java.net/> refusing to serve pages now?

I’ve got all the relevant views in Censum and if one is missing I can add it in short order so I can visualize what Thomas is describing.

java.net is defunct, has been for a few months. I’ve tried to convince the OTN.. or what ever they’ve decided to rebrand themselves as that the loss of java.net is a real loss as that site was well branded and now… but that is a different issue and a different discussion and not for this list ;-)

Regards,
Kirk

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20171017/1a30bd45/attachment.htm>