[wp-trac] [WordPress Trac] #62176: No stable error detection after Automatic Updates with more than one update

[WordPress Trac]

#62176: No stable error detection after Automatic Updates with more than one update
-------------------------------------------------+-------------------------
 Reporter:  georgwordpress                       |       Owner:  (none)
     Type:  defect (bug)                         |      Status:  new
 Priority:  normal                               |   Milestone:  Awaiting
                                                 |  Review
Component:  Upgrade/Install                      |     Version:  6.6
 Severity:  normal                               |  Resolution:
 Keywords:  2nd-opinion has-patch reporter-      |     Focuses:
  feedback                                       |
-------------------------------------------------+-------------------------

Comment (by georgwordpress):

 Hi @costdev
 Thanks for the quick reply!

 My journey started a while ago.
 I'm a big fan of the automatic updates because of the ever faster
 exploitation of vulnerabilities.
 If the risk of activating faulty updates can be minimized by automatic
 rollback - all the better.

 Yes - I fell into all sorts of traps at the beginning that make it
 difficult to reproduce the problem:
 Special features of an existing installation, manual triggering of the
 cron job etc.

 Why the latter can make a difference is partly due to the cookies sent by
 the loopback.
 Depending on how I start the cron job, my admin cookies may be sent along
 with the loopback.
 Then the wp load runs a little differently and the homescreen also looks a
 little different than with a loopback request as an anonymous user.
 I have therefore also ensured in the PR that the loopback request is now
 always carried out as an anonymous user.

 But that's just a side note.

 And to avoid any misunderstandings:
 after my research in the old tickets, I was able to understand how much
 effort was put into testing by everyone involved! Because I've done a lot
 of testing myself in the meantime, I can understand this very well.

 @costdev
 And I especially have a lot of respect for Your efforts!

 There is no question that everything was tested very carefully before
 commissioning!

 But it is an edge case that rarely occurs, especially because the current
 sleep(2) has really reduced the probability.

 I will report later on exactly how I am currently testing.

 But before anyone gets nervous, you should sit back and realize what this
 is all about.

 The normal operation of a WP installation is not at risk!
 Even with automatic updates of core and plugins, I have had very good
 experiences in recent years.
 With careful selection of quality-assured plugins, the release of a plugin
 with fatal errors is very, very unlikely.

 Part 3 of the project is only about further minimizing the operational
 risk by automatically detecting a buggy plugin update and preventing the
 update.

 The current detection cannot detect all errors anyway.
 Plugin errors that only occur when logging into the admin backend cannot
 be detected by the loopback request either - are therefore outside the
 scope of part 3.

 But I hope you can see from my efforts that I am a big fan of the
 automatic rollback project!

-- 
Ticket URL: <https://core.trac.wordpress.org/ticket/62176#comment:5>
WordPress Trac <https://core.trac.wordpress.org/>
WordPress publishing platform