On 18/08/2015 18:16, Christopher Schultz wrote: > Musafir, > > On 8/18/15 1:08 PM, Mark Thomas wrote: >> On 18/08/2015 17:10, musafir wrote: >>> On Tue, Aug 18, 2015 at 11:57 AM, Mark Thomas <ma...@apache.org> >>> wrote: >>> >>>> On 18/08/2015 16:50, musafir wrote: >>>>> Hello, ugraded tomcat from 6.0.36(jdk 6) to tomcat7.0.57 >>>>> (jdk7).. os:centos 6.6 ...seeing high cpu spike after >>>>> upgrade.... doing top and thread dump shows only thread >>>>> that is always using cpu is: >>>> >>>> How have you matched the thread using CPU in top with the >>>> thread in the thread dump? >>>> >>>> Mark >>>> >>>> >>>>> "ContainerBackgroundProcessor[StandardEngine[Catalina]]" >>>>> daemon prio=10 tid=0x00007ff8fc599800 nid=0xc1d2e waiting on >>>>> condition [0x00007ff803ffe000] java.lang.Thread.State: >>>>> TIMED_WAITING (sleeping) at java.lang.Thread.sleep(Native >>>>> Method) at >>>> org.apache.catalina.core.ContainerBase$ContainerBackgroundProcessor. > run(ContainerBase.java:1513) >>>>> >>>> > at java.lang.Thread.run(Thread.java:745) >>>>> >>>>> disabled backgroundProcessorDelay >>>>> (backgroundProcessorDelay=-1) ..don't see any thread that is >>>>> always using cpu.. >>>>> >>>>> Q: why backgroundProcessorDelay will cause cpu spike. Is this >>>>> disabled in tomcat6 by default? Q: will disabling >>>>> backgroundProcessorDelay cause any other performance issues >>>>> with tomcat7. >>>>> >>>>> >>>>> Thanks for Help!! >>>>> >>>>> Regards >>>>> >>>> >>>> Hi Mark pid java >>>> >>>> top -H -p pid for thread id cpu eating thread >>>> >>>> kill -3 pid for thread dump >>>> >>>> and hexa nid for thread id > >> Then it looks like you did something wrong because the thread you >> identified is sleeping. It won't be using any CPU. > > Assuming that the thread has been mis-identified, could it be this? > > https://bz.apache.org/bugzilla/show_bug.cgi?id=58151 > > That bug was filed/fixed against trunk/8.0. Any chance there is a > similar bug in Tomcat 7?
It doesn't look like it. That bug was introduced when the non-blocking I/O support was added in Tomcat 8. The reads in Tomcat 7 look OK. top + lwp id + thread dump is the way to get to the bottom of this. Something isn't right in the description above. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
