Site migration is not working good

[yaazkal]

Hi, using BOA HEAD, boa info here

Site is unreachable (404 error) after migrating to a new platform because nginx vhost file is not updated and also there is somewhere a bad platform related to the site after migration.

Sample scenario:

• A platform called "oldplatform" (Drupal 8.7)
• Site called example.com
• A platform called "newplatform" (Drupal 8.8).

After migrating example.com from "oldplatform" to "newplatform" site is unreachable with a 404 error. re verifying the platform and the site doesn't seems to solve the issue, so you have to manually update the corresponding nginx vhost file at /data/disk/o1/config/server_master/nginx/vhost.d/example.com with the new database, username and password as it is not effectively updated (can see that information at the settings.php file). Side note: I guess that information was meant to not be there, just in the vhost file.

In Aegir dashboard all seems ok (site shows that the plaftorm is newplatform) but if you try to flush caches of the site, you will end up getting something like this task log when provision tries to load the oldplatform, so the task is failed.

The Aegir database looks ok, I mean, the registry points to the correct new platform.

If you run drush @example.com status in shell, all seems fine too. The alias drush file seems also correct (site and platform aliases) and running tasks like flushing caches works fine in shell.

I tried restarting redis, flushing aegir cache and I still get task failure in aegir dashboard.

I still can't find where in the world is still pointing to the old platform.

[yaazkal]

Maybe https://github.com/omega8cc/boa/issues/1454 is related

[omega8cc]

Drupal 8.8 is not yet fully tested and supported in Aegir. Thanks for the bug report!

[yaazkal]

@omega8cc thanks for the answer. It happened to me before migrating from older versions of 8.x too, I can't remember wich one as I don't always create a new platform for upgrades even if it sounds no good in the aegir workflow ;)

[omega8cc]

This could be still a matter of Drush 8 issues with latest Drupal 8 versions. It worked for us before, so perhaps Drush still has something to fix to stay compatible, not sure, we will try to reproduce the problem.

[yaazkal]

I've access the shell today and now it seems it can't find the @example.com alias :) Seems the daily tasks deleted the alias file, oh wait, also the nginx vhost file.

The site is still present in aegir, verifying the site solves the problem. I hope the site still exist tomorrow.

[yaazkal]

I had to report that the 404 error also happens if also upgrade a D7 site, like to a the new platform with the recent 7.69 release.

Nginx and drush alias files are correct but a manual nginx restart and site verification is needed.

[zanami]

Same here

Migrated D8.6 to D8.8 — everything is nice and green drush @site status doesn't show database credentials site doesn't work (redirects to install.php and returns 404)

service nginx restart and site verify helps

[serrato-dan]

I've had this issue a few times as well and it just happened this morning. I was running a migration of site from platform A (8.9.13) to platform B (8.9.16) and everything is green on the dashboard for a successful migration. Site give 404 error and url pushes to domain.com/core/install.php. Ran verify on the site and it fixed it quickly. Not ideal, but verify has been the solution for the handful of times this has occurred.

[tourtools]

Same thing happened to me yesterday both for D6 and D7 migrations: root pointing to install.php, after manual trigger the website verification all went to its place.