Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: ts_main_telescope
-
Labels:
-
Story Points:1
-
Epic Link:
-
Sprint:TSSW Sprint - Oct 12 - Oct 26
-
Team:Telescope and Site
-
Urgent?:No
Description
Deploy the new LabVIEW docker image to the CTIO Nexus. This will install the xdotool to the Dockerfile. The .Xauthority file will be tested with different context to see I could use the Dockerfile to make the Xvnc plugin to work or not. If NI can provide a good way to avoid the prompt window of user login, we will not use the xdotool. The backup (worst solution) is to create the file and install the Caraya unit test manually on the local host. After this, upload the images to the CTIO Nexus to let Diego and Omar to work.
Based on the LabVIEW and Linux OS Compatibility, only the following three combinations have the good compatibility:
(1) CentOS 7 + LabVIEW 2018 SP1,
(2) CentOS 7 + LabVIEW 2020, and
(3) CentOS 8 + LabVIEW 2020.
PS. The NI ticket is here:
https://www.ni.com/my-support/s/case/5003q00001Lfj4bAAB/remove-start-up-dialog
Attachments
Issue Links
Activity
Based on the source code of Xvnc plugin, the .Xauthority will be created by itself :
if (useXauthority) { |
xauthority = createXauthorityFile(workspace, logger);
|
String xauthorityPath = xauthority.getRemote();
|
xauthorityEnv.put(XAUTHORITY_ENV, xauthorityPath);
|
if (context.getEnv().containsKey(XAUTHORITY_ENV)) { |
//We are probably doing a retry and context will complain if we try to set it again |
context.getEnv().remove(XAUTHORITY_ENV);
|
}
|
context.env(XAUTHORITY_ENV, xauthorityPath);
|
} else { |
xauthority = null; |
// Need something to identify it by for Launcher.kill in DisposerImpl. |
xauthorityEnv.put("XVNC_COOKIE", UUID.randomUUID().toString()); |
}
|
Therefore, I do not understand why the user need to create the .Xauthority in the first time. And the details behind is not clear to me.
Prepared three docker images by the Dockerfile already. The next step is the manual part. Created the following images:
cagvm3.ctio.noao.edu:8082/centos7/labview2020_pro:caraya_vnc |
cagvm3.ctio.noao.edu:8082/centos7/labview2018sp1_pro:caraya_vnc |
cagvm3.ctio.noao.edu:8082/centos8/labview2020_pro:caraya_vnc |
I got the error when uploading the third (final) image (cagvm3.ctio.noao.edu:8082/centos7/labview2020_pro:caraya_vnc). Need to check with CTIO in a latter time.
Tested the new docker image and there is no problem:
https://github.com/lsst-ts/ts_SALLabVIEW_test/blob/master/Jenkinsfile.labview
The test result is here:
http://cagvm3.ctio.noao.edu/jenkins/job/ts_SALLabVIEW_test/job/master/
Please help to review the PR:
https://github.com/lsst-ts/ts_Dockerfiles/pull/106
Two of the docker images had been uploaded to the CTIO Nexus. The third (final) one is uploading now. It failed two times of upload in the yesterday's night. If it is still the same, I will feedback to CTIO.
I updated the Jenkinsfile and you can see the test result:
http://cagvm3.ctio.noao.edu/jenkins/blue/organizations/jenkins/ts_SALLabVIEW/detail/master/6/
Thanks!
The images are uploaded to the CTIO Nexus already:
Approved the PR on https://github.com/lsst-ts/ts_Dockerfiles/pull/106#pullrequestreview-515788081 and reviewed the jenkins test results. Great work!
Installed the xdotool in the Dockerfile. Updated the details.
Reformatted the README.md.
I tested to use the environment variable of XAUTHORITY and it did not work.