User data cleaner - Githubissues

catalyst / moodle-local_datacleaner

Reduce, filter, and anonymize moodle data for non-prod environments

https://moodle.org/plugins/local_datacleaner

19 stars 17 forks source link

User data cleaner #12

Closed Nicosoft closed 7 years ago

Nicosoft commented 8 years ago

I cannot run the user cleaner. It keeps running in 'dryrun' mode ... or am I doing something wrong? I am using v2.3.5

Both say: 'Would ..."

Please see the screenshot. -- screenshot 1: DryRun Mode dryrun_mode

-- screenshot 2: Run Mode run_mode

brendanheywood commented 8 years ago

Hmm, In theory it should tell you why it failed if it does a dry run. Can you double check all these conditions are true and let me know which condition isn't being correctly reported back:

It can only be run from the CLI. There is no GUI.
We store the hostname in the cleaning configuration data. If the hostname matches production, DataCleaner will not run. If this data is missing then it will not run.
Typically a refreshed database will be from a nightly snapshot and so the data should be slightly stale. If a non admin user has logged in recently, that's a sign this Moodle is being used, and the DataCleaner will not run.
If cron has run recently, DataCleaner will not run. This should only be run on a data washing instance, cron should not be needed here.
It can only be run if and only if a 'local_datacleaner_allowexecution = true;' has been added to config.php

Nicosoft commented 8 years ago

I double checked the conditions and I believe all the conditions are respected.

I an using Xdebug to follow the execution. In the screenshot 1, you can see the values used in function safety_checks() in --run mode. -- screenshot 1: 1_run_values

When executed in --dryrun or --run the function safety_checks() returns FALSE. So I believe this is correct.

I have more concern when the object cleaner_users/clean is constructed. When executed in --run mode, cleaner_users\clean->dryrun = TRUE; I believe the correct value should FALSE not TRUE. (Please see the screenshot 2) -- screenshot 2: 2_run_construct

So obviously when the method execute() is called, because of the value of $dryrun, method stops before any update. -- screenshot 3: 3_run_execute

What do you think?

roperto commented 7 years ago

Hi Nico,

The data scrambler was refactored and a few issues with dryrun fixed, as a result I cannot reproduce that issue anymore.

Please, if you find out that there is a similar issue with the new version feel free to reopen it or create a new one, I will gladly look into it.

Cheers,

Daniel