Larry, On 11/17/15 10:13 AM, Cohen, Laurence wrote: > Chris, that's what I'm planning on doing. But how do I get the pid since > it will change each time I bounce the java process? Also, I'll have to > figure out how to read the thread dump.
You should be able to use "ps" and "grep" to find the process id. Don't bother reading the thread dump: just email it to yourself. -chris > On Mon, Nov 16, 2015 at 5:48 PM, Christopher Schultz < > ch...@christopherschultz.net> wrote: > >> Larry, >> >> On 11/16/15 5:08 PM, Cohen, Laurence wrote: >>> Thank you Christopher. I'm going to start with the thread dump since we >>> are using jdk and that appears very straightforward. Part of my dilemma >> is >>> that the problem is occurring on a private network where I do not have >>> access to the internet. >> >> You could have your 'wget' script trigger a jstack before bouncing >> Tomcat. That way, you don't have to have quick access to the server when >> it starts misbehaving. >> >>> Our public facing application with the same exact >>> build is not experiencing this problem, so I'm wondering if this is a >>> network issue on the private side. I'll start here though. >> >> It's nice that it's not the other way around. It's *usually* the other >> way around: dev/test is juuuuust fine but prod is hosing all over the >> place. >> >> -chris >> >>> On Mon, Nov 16, 2015 at 4:55 PM, Christopher Schultz < >>> ch...@christopherschultz.net> wrote: >>> >>>> Larry, >>>> >>>> On 11/16/15 4:42 PM, Cohen, Laurence wrote: >>>>> Are there any tools that come with Java that I can use to troubleshoot >> an >>>>> intermittent problem we are having? The problem is that several times >> a >>>>> day, our Tomcat applications will stop responding and I'll have to >>>> restart >>>>> them to get them working again. It's gotten to the point where I have >>>>> written a script which does a wget every 10 minutes against an object >> in >>>>> the DB, and if it fails, it will restart our apps. >>>> >>>> Consider using a real monitoring tool. There are some free ones >>>> available, such as Nagios and ighinga, that aren't much more complicated >>>> than your wget script, except that they have alerting, history, etc. all >>>> built around them. They also let you sample LOTS of things. >>>> >>>>> I've also done some statistics gathering and imported them into a >>>>> spreadsheet so I can see what is going on at the time the system is >>>>> crashing. All I can see is that the Tomcat connections are spiking. >>>> >>>> Spiking to a particular limit? What does your connector configuration >>>> look like? And your deployment? >>>> >>>>> We are running Tomcat 7.0.59 with two apps, Postgres 9.2 on the backend >>>>> which is not administered by us, and httpd on the front end, 2.2.15. >> The >>>>> httpd server and app server are RHEL6. >>>> >>>> Just a single Tomcat instance? That narrows things down a bit. How are >>>> you reverse-proxying? mod_jk? mod_proxy_http? >>>> >>>> What does your JNDI DataSource configuration look like? >>>> >>>> Are you able to take a thread dump when the server seizes-up? >>>> >>>> >> http://wiki.apache.org/tomcat/HowTo#How_do_I_obtain_a_thread_dump_of_my_running_webapp_.3F >>>> >>>> This will tell you what the server is doing. I suspect you'll see a >>>> bunch of threads waiting on a database connection or something like >> that. >>>> >>>> -chris >>>> >>>> --------------------------------------------------------------------- >>>> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org >>>> For additional commands, e-mail: users-h...@tomcat.apache.org >>>> >>>> >>> >>> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org >> For additional commands, e-mail: users-h...@tomcat.apache.org >> >> > > --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
