New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
netdata/packaging/installer: Finish up netdata installer refactoring #5705
Comments
Last time the previous PR was tested, the following problem was identified when run stable and improved versions of netdata in brand new containers: Raw diff of logs between the two versions
From the above, the noticeable are the following (they are easier to understand with vimdiff
|
The chmod removal on the branched change was introduced intentionally by cde4264, depre According to #1292 bug, this addition was attending a problem back then.
|
The other difference that i am investigating next is the post-install start up of the service.
|
I 've spotted a couple more nits from the log file that i have committed already. Moving forward with my checks, i executed the following commands, so that i can then validate the generated folder structures from the installer
|
executing the tree command on both install folders and then vimdiffing them, i confirm that they generate an identical tree structure
|
I also confirmed the attribute details on each file by doing the following
The only differences are in some file sizes because of the different parent dir names in the paths primarily and also some timestamps because of the different execution time |
Now, in order to make sure we correctly evaluate the user/group ownership details we should run the content listing from within the container.
|
Refactor code | Stable
So even though captured contents of files are fine, with the previous commands we also captured the output of the installation. It seems that we have a difference on these commands. I will have to check whether this is again one of the intentional changes |
We finally nailed the root cause for this difference also! we deep dived and we discovered that when we switched to To mitigate this, we basically renamed BASH function pidof to safe_pidof which is actually more accurate name too and thus we prevented the loop. PR is already updated.
|
Merged all fixes, closing |
Feature idea summary
This issue is a wrap-up task of PR #5505
We will use it to track any further problems identified during testing and also make it the link between the old PR and the new that i will be producing with any further fixes we may come up with.
CC: @cakrit
Expected behavior
netdata-installer.sh runs as before, with the fixes included.
The text was updated successfully, but these errors were encountered: