jstack | Preemptive jstack for production debugging | Hacking library

We are running our kafka stream application on Azure kubernetes written in java. We are new to kubernetes. To debug an issue we want to take thread dump of the running pod.

Below are the steps we are following to take the dump.

1. Building our application with below docker file.

   ...

I've Linux Mint 20.2 Cinnamon and I tried to install jenkins. Detailed steps for installation is as below:

1. Installing java - installed via apt

I have more than one java application running in separate docker containers. I am trying to collect monitoring data such as GC log, thread dump, heap dump from the java process running inside a container using tools like jstat, jstack, jmap. Is it possible to capture this information from the host(outside containers)?

I am new to the containerized world. I understand that PID namespace of the host and container is different. When I execute jstack > thread_dump.txt from the host, it shows error message: Unable to open socket file /proc/root/tmp/.java_pid: target process doesn't respond within 10500ms or HotSpot VM not loaded

Where PID is process id from the host PID namespace.

When I execute jstack inside container ( docker exec -it ) then it is able to capture thread dump.

Where PID is process id from the container PID namespace.

Any hints on how to solve this?

I am running Tomcat 9 on an Ubuntu 20.04 OS using OpenJDK 64-Bit Server VM (build 25.275-b01, mixed mode). When I try to gather diagnostics info using jstat, jstack, etc, I see PID not found. jps also cannot identify the Tomcat process id.

I have checked several posts like one, two, three, four, etc, but none of the answers given in these helped me to solve my problem!

Even though I am passing the username with which the Tomcat process is running, jstat cannot find that process: sudo -u tomcat jstat -gc 476174 5000

In case it matters:

1. I can see that the Tomcat process is started with -Djava.io.tmpdir=/tmp This folder is owned by root user but has full permissions(777) enabled.
2. When the Tomcat process is started, I can see a folder with name systemd-private-e6d8b5dc224848f8a64a3e943ac2e9c4-tomcat9.service-UH5knj (the last few chars after service- change every time the process is restarted) getting created with owner as root (probably because I start tomcat service using sudo service tomcat9 start) and this folder has permission of rwx------.

Any hints on how to solve this issue?

Thanks, Shobhana

I've been trying to find ways to obtain the Thread Dump from a Java Application on my Windows server running on jre 1.8.0_144.

Non of the monitoring utilities like jcmd jstack jconsole are available in either the bin or lib folders of the Java environment directory.

I have come across several applications online that claim to perform the same task but haven't found a reliable one yet.

Appreciate it anyone has a recommendation or a suggestion. (Changing the JRE version, unfortunately, has been ruled out as an option)

By going to the Amazon EMR summary page, I can only see command for how to connect to the master node with hadoop user:

Does anyone know what the * following a filename in macOS terminal means?

All the file names in the folder:

/System/Library/Frameworks/JavaVM.framework/Versions/Current/Commands

have a * trailing them and I'm not sure why.

For example:

The GC time is too long in my spark streaming programme. In the GC log, I found that Someone called System.gc() in the programme. I do not call System.gc() in my code. So the caller should be the api I used.

I add -XX:-DisableExplicitGC to JVM and fix this problem. However, I want to know who call the System.gc().

I tried some methods.

1. Use jstack. But the GC is not so frequent, it is difficult to dump the thread that call the method.
2. I add trigger that add thread dump when invoke method java.lang.System.gc() in JProfiler. But it doesn't seem to work.

How can I know who call System.gc() in spark streaming program?