Troubles with Shenandoah

Aleksey Shipilev shade at redhat.com
Mon Apr 8 14:13:29 UTC 2019


On 4/8/19 3:59 PM, Simone Bordet wrote:
>> I currently suspect that it's running out of code cache, and drops back
>> to interpreter. -XX:ReservedCodeCacheSize=2048m seems to fix it for me.
>> Can you verify it?
> 
> Just had another crash with fastdebug and
> -XX:ReservedCodeCacheSize=2048m, attached the hs_err.

#  fatal error: Safepoint sync time longer than 10000ms detected when executing Cleanup.

jdk/jdk has the code that SIGILLs the thread that failed to reach the safepoint, and thus provides
better diagnostics than this.

-Aleksey



More information about the shenandoah-dev mailing list