On Sun, Mar 19, 2017 at 12:48 PM, André Warnier (tomcat) <a...@ice-sa.com> wrote: > On 17.03.2017 14:54, Christopher Schultz wrote: >> >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA256 >> >> Eric, >> >> On 3/16/17 11:01 PM, Eric Chua wrote: >>> >>> I am running tomcat 8.0.121. When I start my tomcat, it seems to >>> be eating up all the memory on my system. I have 16 GB, and it >>> keeps on going. >> >> >> What are your memory-related parameters when you launch the JVM? >> >> Note that Java *never* gives any memory back to the OS, even when the >> heap-usage goes down. This is a Java thing, not a Tomcat thing. >> >>> Then when I try to kill the process, it dies but 12 GB is still >>> being used even though everything is turn off. >> >> >> That makes no sense at all. Please provide some evidence this is >> happening. >> >>> The only way to reclaim the memory is to reboot. >> >> >> If this is true, then you have some kind of awful kernel bug. >> >>> I am running on redhat 6.5 and can't figure out what could be >>> causing this. I run the tomcat as a local user, and I know there >>> aren't any other processes running as the local user. I am running >>> a spring MVC 4/Java 8/ struts web application. I have two of them >>> with the same issue. Any help would be appreciated. When I try to >>> view all the running processes I cannot see where most of the 12 gb >>> are being used. The system came up with 2.2 gb used and after I >>> start one web application it goes to 14-15gb. >> >> >> Ok. >> >>> The funny thing is that I can kill it to reclaim the memory. Only >>> a reboot works. >> >> >> You mean you CAN'T kill in to reclaim memory, right? >> >>> I am running a VMware instance with vcenter version 6.5. This >>> does not happen with Java 7 with tomcat 7. Any help would be >>> appreciate. >> >> >> Something tells me you are reading or interpreting something >> incorrectly, here. Can you please share your raw data, and where you >> got that raw data? Something like reports from free/ps/top/sar/etc.? >> > > The OP might be looking at "memory usage" in the Vmware GUI, and confusing > "memory allocated to that Virtual Machine", with "memory usage within the OS > of that Virtual Machine". > If Vmware at some point allocated more memory to that Virtual Machine, it > may never reduce it until some other VM wouls need it (or indeed until the > OS of the VM is rebooted). > > With Vmware birtualisation, it can easily get a bit confusing when trying to > figure out "memory usage". Try figuring out what happens to Linux memory > swapping for instance. > (Or "ballooning").
Agreed. One could easily find any rogue JVMs with a "ps aux | grep java" Anyway, here's what I would do - as a superuser, run Mission Control - it will list any JVMs running. If there are any JVMs running, other than the Mission Control JVM, connect to the one with the high memory usage to investigate. If there are no other JVMs running, then there's your answer - there is no rogue JVM consuming 12-14gb RAM. --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
