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
execute-as-user code change #517
Conversation
2. modified ProcessJob to call execute-as-user 3. modified AzkabanProcess so that log consumption starts earlier (and don't get gobbled up during execution failures) 4. Made type=command read sysProps as well
… default. 2. Changing the C code to print less verbose log outputs
Update Gradle wrapper version to 2.7.
* Add gradle-errorprone-plugin to build.gradle and enable for all Java builds. * Fix compile-time errors raised by error-prone. * Add @ignore to PythonJobTests.testPythonJob because it appeared to hang. Fixes azkaban#501
Use error-prone for Java build.
Conflicts: azkaban-common/src/main/java/azkaban/jobExecutor/ProcessJob.java
@@ -101,7 +101,7 @@ | |||
private ExecutingManagerUpdaterThread executingManager; | |||
// 12 weeks | |||
private static final long DEFAULT_EXECUTION_LOGS_RETENTION_MS = 3 * 4 * 7 | |||
* 24 * 60 * 60 * 1000l; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It seems that your diff base is old since this patch contains a number of changes from #508. Can you please rebase?
Can you rebase against master instead? That way, you do not have all the old commits polluting your diff. Ideally, I think it would be better to squash all of the commits into one instead of keeping all the intermediate patches:
|
Hey David, Do you mean to merge the code from master branch? Or do you mean for me to squash my commits so the log looks more clean? Thanks! |
Hi John, I took a look at your fork and the revision history for this PR. It looks like your master branch diverged from the azkaban/azkaban master a while ago and the commits for this PR are interleaved with more recent commits from upstream. I actually tried cloning your fork of Azkaban, adding azkaban/azkaban as upstream, and doing Here is how you can do this:
Once you have done this:
Until this PR is merged, if there are new commits that go into master before your commit, I would recommend rebasing so that your patch will come after those new commits:
After your patch is merged, sync your master branch with upstream using:
After your patch is merged, I would recommend doing development in feature branches in your fork and keeping your master branch identical to upstream/master at all times, and when new changes land in master, rebase your feature branch on top of master. To illustrate:
This way, your master branch will always be identical to upstream, and your commits are always on top of master and not interleaved with older commits. It is good to rebase often, whenever new commits land in master. This will greatly reduce the number of painful merge conflicts and will allow you to work on more than one change by using different feature branches. The only caveat is that if you want to push your feature branch to your remote branch (such as if you want to continue working on a different machine), you would need to do a force push after you do a rebase:
I think force-pushes for feature branches in this case are fine, but force pushes for master branches should be avoided whenever possible. :) Anyway, sorry for the hassle. I think we should try to rebase and squash commits before merging pull requests. This way, we would reduce the likelihood of particularly nasty merge conflicts, keep the revision history cleaner, and make it easier to track down why a certain line of code is changed. This is also closer to what Apache projects do. Feel free to email me if you run into any issues. I am usually online in the evenings as well. |
Hey David, Thanks for your such detailed instructions! I've copied them down somewhere, as it will come in handy when merging into master :) I'll do the squash, and that should reduce a lot of cluttering on the logs. We plan to merge multiexecutor-canary into master after my execute-as-user feature request. That will be another whole big operation, and we can watch out to use rebase at that time. Thanks! |
Thanks for the clarification. I didn't notice that your diff-base was Sounds good. Feel free to let me know if you need any help with the rebase. |
No description provided.