Long TTSP

[yu.zhang at oracle.com]

Hi,

There is a development flag +DieOnSafepointTimeout that will create a 
core dump when the time to safepoint exceeds the threshold.

I suggest you change this flag to product and rebuild jdk. Because 
running a development build may skew the application.

Then you can debug the core dump and see what is going on.

I found some times slow block device can cause the TTSP reach seconds. 
There is a good discussion

https://groups.google.com/forum/#!topic/nosql-databases/OSBlUVp0vbw

I am interested to learn what you have found.

Thanks

Jenny


On 05/03/2017 10:01 AM, José Muñoz wrote:
> Hi,
>
> I'm analyzing the performance of a VM with heap of 160g and I see TTSP 
> of up to 2 seconds.
>
> I added some flags to get more information:
>
> -XX:+PrintSafepointStatistics
> -XX:+SafepointTimeout
> -XX:SafepointTimeoutDelay=1000
> -XX:PrintSafepointStatisticsTimeout=1000
> -XX:PrintSafepointStatisticsCount=1
>
> but I couldn't find any flag to print the stack trace of a thread that 
> is executed when it fails to reach the safepoint after the delay. I 
> would like to find a way to figure out which code is executed when a 
> thread reached the TTSP threshold ?
>
> Thanks,
> Jose
>
>
> _______________________________________________
> hotspot-gc-use mailing list
> hotspot-gc-use at openjdk.java.net
> http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/hotspot-gc-use/attachments/20170503/6eea9407/attachment.html>