DO180 Comprehensive Review - Container exiting immediately

[scott4]

I've tried for about 3 days to get past this. I tried my Dockerfile and the solution's Dockerfile. The container builds, I ran the script to get the nexus files, and when I run the run-persistent.sh the container exist immediately.

Running podman logs only shows this:

nexus-start.sh: line 8: cd: /opt/nexus/nexus2: No such file or directory
Error: Could not find or load main class org.sonatype.nexus.bootstrap.Launcher

Any ideas? I reviewed the persistence section and everything seems right.

[scott4]

I had to change the curl in get-nexus-bundle to follow redirects. It was downloading the tar as an html file, and that broke the next steps. Maybe I screwed something up, not sure.

[Balakondareddy]

Hi
I am facing the same issue and tried with solutions as well, no luck.
Stuck at the same step.
Any help would be appreciated.
Thanks

[jordisola]

Hi everyone

I just tried to replicate the issue but it worked fine on my side.

What versions of the course are you using? Can you provide the output of the `curl` command?

KR

Jordi Sola

[scott4]

Like I described, the video instructs to get that nexus tarball, but if I recall, the url was correct, but there was a redirect on the server side to a CDN or something so the curl failed. I didn't throw an error until later in the process. 

OP: Make sure that you have the right nexus files and that they extract and install, as expected. As soon as I could get a tarball from curl, I think that fixed it. 

TL;DR: I think I just had to add -L to my curl.

[Balakondareddy]

Hi
I can't run the curl as the pod is getting exited after running ./run-persistent.sh.

I tried running ./run-persistent.sh after re-building the nexus images. Will that be a problem ?

Snippet of the steps performed.

1. [student@workstation image]$ ./get-nexus-bundle.sh
2. ######################################################################## 100.0%
3. Nexus bundle download successful
5. [student@workstation image]$ sudo podman build -t nexus .
6. STEP 1: FROM ubi7/ubi:7.7
7. --> Using cache b3110dfff0ed735f2b6dd2fff6815a3a7c040cd8a7db4b800c94499abb393174
8. STEP 23: COMMIT nexus
9. [student@workstation image]$
10. /home/student/DO180/labs/comprehensive-review/image
11. [student@workstation image]$ cd ../deploy/local/
12. [student@workstation local]$ ./run-persistent.sh
13. 37bab4a1053d8f091335daca184fe6f0894109dd57656182733726226b4b3f21
15. [student@workstation local]$ sudo podman ps -a
16. CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES IS INFRA
17. 37bab4a1053d localhost/nexus:latest sh nexus-start.sh 11 seconds ago Exited (1) 10 seconds ago vibrant_hermann false
18. [student@workstation local]$
19. [student@workstation local]$ sudo podman logs vibrant_hermann
20. nexus-start.sh: line 8: cd: /opt/nexus/nexus2: No such file or directory
21. Error: Could not find or load main class org.sonatype.nexus.bootstrap.Launcher
    ##
    
    

[scott4]

I think the curl might be in that ./get-nexus-bundle.sh maybe?

[scott4]

Oh, and I think you may want to build the container from scratch with --no-cache, maybe?

[jordisola · Accepted solution]

Hi

I should have been more explicit. You are right, I meant the `curl` command inside the `get_nexus_bundle.sh` script. I should be something like: 

> curl -L --progress-bar -O https://download.sonatype.com/nexus/oss/nexus-2.14.3-02-bundle.tar.gz

Note the "-L" parameter, indicating the curl command to follow redirects.

It really sounds like the script is silently failing, and the tarball is not being downloaded (IIRC, a HTML page is downloaded instead). This does not block creating the container image, but causes the later run issue.

P.S. Building the image using the --no-cache parameter is not necessary. If local files are changed, the layer including those files will also change, hence the cache will not hit.

[KevinRoche]

My container kept failing to stay up and running as well, but if I ran it as interactive and ran the nexus-start.sh script from there, it worked.

I finally figured out there was an error in the CMD line in the Dockerfile.
Changing it from

CMD ["sh", "start-nexus.sh"]

to 
CMD ["/bin/sh","nexus-start.sh"]

solved the problem.