Please fix how you upgrade the database
-
Currently whenever there is a database upgrade, Broken Link Checker will try to upgrade the database in the init hook. This is a major problem because that means if you have a high trafficked multisite instance, thousands of requests are triggering that database hook script for multiple sites at the same time.
We’re a high traffic university system with over 800+ sites in our multisite install, so for us this always leads to a db lockout error.
Every time there is a BLC update, I have to comment out the upgrade script, push it up to our production environment, then write a command line script that manually runs that upgrade code in a foreach loop over each site. It’s not very ideal.
Can your dev team please create a upgrade method that’s safe for large multisite installs? Maybe there can be some network level queue table? On init, if that blog id is not top of the queue, add it to the queue. If it is on top of the queue, run the upgrade script. That’s a half baked idea, but I think anything that makes those updates sequential across a multisite install would be a huge improvement.
- The topic ‘Please fix how you upgrade the database’ is closed to new replies.