Thanks Matt, I'll look for GC tuning params for java-1.8.0-openjdk. These hosts do have two processors and 8G of ram.
[kxn2@amq-d02 ~]$ grep processor /proc/cpuinfo processor : 0 processor : 1 [kxn2@amq-a02 ~]$ grep processor /proc/cpuinfo processor : 0 processor : 1 [kxn2@amq-p02 ~]$ grep processor /proc/cpuinfo processor : 0 processor : 1 --- Karl Nordström Systems Administrator Penn State IT | Application Platforms ________________________________ From: Matt Pavlovich <[email protected]> Sent: Monday, February 7, 2022 9:47 PM To: [email protected] <[email protected]> Subject: Re: GC overhead limit exceeded Hello Karl- Those error messages are usually indicative of a garbage collection (GC) tuning problem at the JVM level. On any busy system running Java, the default GC settings may not force garbage collection frequently enough in order to keep memory resources available. I suggest you look into garbage collection logging and GC tuning options for your version of the JDK. Thanks, Matt > On Feb 7, 2022, at 2:57 PM, Nordstrom, Karl <[email protected]> wrote: > > Hello activemq users, > > We are running v5.16.3 on our amq dev nodes and hit these errors today. Has > anybody else encounter these? The heap size on dev is 1G while it's 4G on > acceptance. I wonder if 5.16.3 requires more space than 5.15.12 did? > > 2022-02-07 15:22:39,325 | ERROR | JobScheduler: JMS Caught exception in > mainloop | org.apache.activemq.store.kahadb.scheduler.JobSchedulerImpl | > JobScheduler:JMS > java.lang.OutOfMemoryError: GC overhead limit exceeded > > 2022-02-07 15:22:45,761 | ERROR | Checkpoint failed | > org.apache.activemq.store.kahadb.MessageDatabase | ActiveMQ Journal > Checkpoint Worker > java.lang.OutOfMemoryError: GC overhead limit exceeded > > > --- > > Karl Nordström > > Systems Administrator > > Penn State IT | Application Platforms
