Hi Ignacio,
Any time you encounter a 500 error like this, I suggest taking a look in the webserver error logs so we can learn more about the nature of the error. See:
Please share any related error message you find - hopefully it will give us more insight into what's happening, so we can provide further suggestions.
In the meantime, if you haven't already, then you may want to try the second solution mentioned in the original
solution thread - editing the Nginx
fastcgi_read_timeout value to increase the setting allowing the loaddata call to complete. If you do try this, remember two things:
- This needs to be done before the data is loaded - so you may need to ensure you have a backup of your sql data, and then drop/recreate the database, run the purge task to clear any residual data, and then implement the recommended changes before reloading your data, running the upgrade task, etc.
- After making any changes to the Ngnix configuration file, you'll want to reload Nginx.
Implementing these reminders will look something like this:
First, make sure you have a copy of your sqldump outside of AtoM - we are going to delete ALL data currently held there, and try reloading it. If this is a brand new installation and you're not loading any legacy data from an earlier version, then don't worry, there's nothing to lose.
We'll need to access the MySQL command prompt, which means you'll need to know the database name and user that were established during the installation process. Instructions on how to access the MySQL command prompt, and how to find this information if you don't remember it, can be found here:
- mysql -u username -p -e "DROP DATABASE IF EXISTS atom;"
- mysql -u username -p -e "CREATE DATABASE atom CHARACTER SET utf8mb4 COLLATE utf8mb4_0900_ai_ci;"
Replace username with the database user name used during installation. You'll be prompted for the password used during installation as well.
Additionally, after dropping and recreating the database, but BEFORE loading your backup, you can also try to run the tools:purge command-line task. This will flush any residual data from the db, which can sometimes be useful after multiple installation attempts. Remember, be sure you have a copy of your sqldump outside of AtoM before running this - everything will be deleted! Run the following from the root AtoM installation directory - typically this is /usr/share/nginx/atom if you have followed our recommended installation instructions:
When you run the task, AtoM will purge everything and then prompt you for a new site name and description and admin user credentials. If you do have a backup you intend to load, then it doesn't really matter what you enter at this stage, because they will be overwritten once you load your SQL dump again.
Now we can make the changes to the Nginx configuration file recommended in option B of the
solution thread I linked to in my last post.
We'll want to reload Nginx for the changes to take effect after:
- sudo systemctl reload nginx
Now you can follow the rest of the Upgrading instructions to load your data (e.g. load the sqldump, run the upgrade task, restart services, reindex, etc). If you have no data to load and all other normal installation steps have been completed, then you can try re-launching the web installer by returning to the loadData page.
Note that some users report that adjusting the PHP execution times can help as well. See:
You may also find that running a resource monitor will help to know when it's safe to advance to the next page if you continue to encounter timeout issues. See for example this part of the thread:
We have some recommendations on how to run such tools here:
I'm sorry that this has been a frustrating process! We are hoping to remove the web installer completely for the 2.7 release to avoid these kinds of issues - since 90% of the installation process is done in the command line, there's no reason such a resource intensive process at the end should need to be executed via a web browser. Up until now we've relied on the web installer provided by the Symfony framework, but given these continued installation issues, our hope is that removing it and moving to a command-line-only installation process will avoid these issues in the future.
Let us know what you find in the error logs, and if you try this second option, let us know how it goes.