Collecting thread dumps

 
Solved Solved
0 2 1,260
1 ACCEPTED SOLUTION

2 REPLIES 2

1. While API traffic is stopped, login to your Message Processors (MP).

2. get the message processor pid using:

ps -ef | grep Micro

3. collect the jstack against the pid you got from step 2 (note your

location for java_home may be different, replace with your location)

sudo -u apigee /usr/java/jdk1.6.0_45/bin/jstack -l <pid> > jstack1.txt

4. repeat step 3 again 2 more times (for a total of 3 jstack output

collected against each MP) on the same MP renaming the output file

jstack2.txt, jstack3.txt, jstack4.txt, etc.

5. repeat steps 1-4 on any second, third, fourth MP servers in the

cluster. Name the files something different to distinguish these are

from other MPs.

6. restart your MPs.

7. collect the /opdk-home/var/log/apigee/logs/system.log files from each MP.

8. send the following data to Apigee support:

- collection of jstack data from each MP

- collect system.log files from each MP

- note the date/timestamp of the problem occurrence

We have a tool built into the OPDK to collect thread dumps and more, see:

https://community.apigee.com/articles/1907/collecting-diagnostic-data-from-apigee-edge-with-t.html