We have a couple of critical Java Web Apps running on a bunch of Apache Tomcat 7 servers, on top of Centos 6.x.
From time to time, once or twice a week, one of the Tomcat instances suddenly shuts down, with no trace, no OutOfMemory Exceptions, no other Known/Unknow exceptions are being logged. I have checked Tomcat Logs, Application specific Logs, /var/log/* logs and no evidence of an issue.
The incoming traffic/users are load balanced to several other Tomcat HTTP Servers, and by the time of the issue, the other servers are behaving OK.
The server that shuts down varies each time, so I can not blame a specific server for this failure.
Linux flavor and version are CentOS release 6.3 (Final).
Each server has 2 CPU Cores and 4 GB of RAM.
Can you please guide me on troubleshooting this error, I am running out of ideas and tools.
Thanks in advance!
Answer
Presumably you are writing your own log files for your Web-App to debug this. If you’re not finding any logs in the normal tomcat locations, then it might not be tomcat that’s having the issue and your web-app instead. If the web-app experienced an exception that was caught and handled, it could have killed tomcat without leaving a trace. Only if tomcat receive an unexpected error would it leave logs in it’s own log files; your app needs to log what your app is doing.
If you already have logging for your app, then you might need add more logging statements and increase the verbosity of those statements. You might want to look into things like log4j to help control this.
Attribution
Source : Link , Question Author : Santiago Campuzano , Answer Author : Andrew
