Hello,
It looks like the current glassfish sizing tries to allocate more RAM that the server has remaining and consequently memory swap, and connections to MySQL piled up then App crashes, followed by server reboot.
This now seems to happen frequently...we tried to make some changes to the heap size (Xmx/Xms), max-thread-pool-size and innodb_buffer_pool_size with no much luck ..
- It is a physical Linux; 504GB RAM; 7GB swap; 2 sockets; 40 CPUs
- v4.3.8.0; performance-profile=large
Current set-up (domain.xml):
-Xmx200g
-Xms200g
NewRatio=2
SurvivorRatio=8
max-thread-pool-size="360"
Cuerrent set-up (db.cnf):
innodb_buffer_pool_size=184320M
Here is what i would like to understand... Do these values look reasonable? if not, could you please suggest what needs to be adjusted?
Do we need to use -Xmn or can the young gen be managed via NewRatio & SurvivorRatio? As the NewRatio=2, my understanding is current -Xmn is half the heap size (1:2) i.e ~100g.
Awaiting some quick response...
Thanks,
Sundar
