Help needed — SuiteCRM 8 install on Oracle Cloud (Ubuntu) — login/500 errors

📦 Install & Upgrade Help
1

Sorry in advance for a detailed post. It took me 5 days to actually install SuiteCRM to the point that I could actually reach the installation screen. Now I’ve been trying to log in for 2 days. I believe this is a common issue from what I have seen here.

I am not technical, so unfortunately have been relying on AI assistants to provide CLI commands based on publicly available SuiteCRM advice, forums, and guides. I am at a loss and am turning this community for help. A clean install is definitely an option, although considering how painful it was to get the system working in the first place, I might just have to give up on SuiteCRM.

Summary of problems encountered:

  • Fresh install shows “Login credentials incorrect” immediately; suitecrm.log contains “[FATAL] Temp Useris not retrieved … boolean given”.

  • Common causes found and attempted fixes:

    • File/dir permissions wrong (many files owned by root). Fixed with chown -R www-data:www-dataand chmod 775 on cache/var/etc.

    • Admin password resets via CLI (suitecrm:admin:reset-password) failed because bin/consolemissing on this package; fallback manual SQL hash update used.

    • Missing/incorrect DB manager: logs showed “unable to load DB manager for: mysqli”. config.php’sdb_manager needed forcing to MysqliManager and include/db/MysqliManager.php sometimesmissing — restored from upstream when so.

    • Missing DB columns: users table lacked totp_secret/totp_enabled/totp_secret_created → causeduser load to return boolean. Added columns and set ROW_FORMAT=DYNAMIC to avoid row-sizeerrors.

    • PHP compatibility/extensions: system was on PHP 8.1; SuiteCRM 8 expects PHP 8.2 andextensions (intl, mbstring, xml, curl, zip, bcmath, gd). Installed php8.2 and extensions, switchedApache to PHP 8.2.

    • Symfony/bootstrap issues: public/index.php showed raw PHP or Symfony errors when .envmissing or vendor missing. Fixed by creating .env with APP_ENV/APP_SECRET/DATABASE_URLand ensuring vendor/ exists (composer install if missing).

    • Apache config / DocumentRoot wrong: Apache pointed at repo root instead of /public (resulted inraw code or wrong path /public/public). Fixed VirtualHost to DocumentRoot/var/www/html/suitecrm/public and Directory accordingly.

    • Missing var/cache (and var/log) directories: Symfony crashed silently when var/cache missing —created directories and set ownership/permissions.

    • Internal Server Error (500) persisted after above; likely causes remaining: missing vendordependencies (composer install required), missing PHP extensions, bad .htaccess / mod_rewriteor redirect loop, or silent fatal before logging (needed to enable display_errors temporarily tocapture trace).

What I’ve already tried (commands/actions)

  • chown -R www-data:www-data /var/www/html/suitecrm

  • chmod -R 775 on cache/var/etc; created /var/www/html/suitecrm/var/{cache,log}

  • Checked/edited config.php db_manager → MysqliManager

  • Restored include/db/MysqliManager.php from SuiteCRM 8.3.0 when missing

  • Added totp_* columns to users table and ROW_FORMAT=DYNAMIC

  • Reset admin password via suitecrm CLI (when available) or manual bcrypt SQL update

  • Installed PHP 8.2 + extensions; disabled php8.1, enabled php8.2 in Apache

  • Created/verified .env (APP_ENV, APP_SECRET, DATABASE_URL)

  • Ensured DocumentRoot → /var/www/html/suitecrm/public and updated Apache site config

  • Created vendor directory / ran composer install when possible

  • Enabled PHP info test.php to verify PHP processing

  • Temporarily enabled display_errors in public/index.php to capture fatal trace (recommended)

Key logs / error lines seen

  • [FATAL] User update error: Temp User is not retrieved at ID 1, boolean given

  • [FATAL] unable to load DB manager for: mysqli

  • Apache: AH00112/ AH00558 warnings when DocumentRoot incorrect

  • Browser: raw PHP source displayed (when Apache served wrong directory) or generic 500 InternalServer Error (when PHP crashed before logging)

Immediate help I need from the community

  1. If you’ve successfully installed SuiteCRM 8 on Oracle Cloud (Ubuntu 22.04) — any known nuances withthe OCI image that break vendor/composer installs or file ownership?

  2. Advice or checklist for verifying vendor/composer dependencies in a constrained/free-tier environment(composer memory/timeouts, proxy, or network limits).

  3. Tips to capture fatal PHP/Symfony errors when Apache logs remain empty (best way to get a stack tracesafely on a public VM).

  4. Any SuiteCRM-specific patches/hotfixes (8.x) known for Oracle Cloud or for issues with PHP 8.2 +Ubuntu 22.04 that I should apply (links or patch names appreciated).

  5. If you’ve seen the totp_* schema issue before — was it caused by an installer bug or an upgrade path?Any safer schema migration SQL recommended?

Environment details (if helpful)

  • Oracle Cloud Free Tier VM (Ubuntu 22.04 LTS)

  • Apache 2.4.52

  • PHP originally 8.1 → switched to 8.2

  • MySQL/MariaDB local (named schema: suitecrm)

  • SuiteCRM 8.x (installation variant uses Symfony frontend + legacy backend)

  • Tried both CLI and manual fixes; saw missing bin/console in this package

What I can provide on request

  • Exact outputs / tail of /var/log/apache2/error.log,/var/www/html/suitecrm/public/legacy/suitecrm.log

  • Output of php -m, ls -la of vendor/ and include/db/, current config.php db_manager line, and schemadescribe users;

  • The composer install output (if requested)

Thanks — any focused pointers or a minimal set of commands to verify vendor integrity + capture the fatal trace would be hugely appreciated.

2

Your installation sounds like a nightmare and I don’t think you really have those 10 problems, you probably just one big underlying problem, that for some reason is being harder to find and fix than usual.

Some of the things like missing modules and missing file permissions, can allow for a partial, broken installation. And then you’re left picking up the pieces of that process. The real fix is to start with everything in place, and aim to achieve a full, error-free execution of the entire installation process.

This isn’t that hard to do; SuiteCRM is installed every month by hundreds of people. It’s true some of them have problems, but not all of them; and not almost all of the possible problems, like it seems to be your case.

If it is Ubuntu and you follow the official instructions (not others) you should be able to get a clean install.

1 Like
3

Thank you, I’ll try again. This is what happens when a complete novice uses LLMs which often get things wrong. I’ll appreciate everything that the SuiteCRM team are doing as well. Thanks!

4

@rcword LLM’s are particularily bad at SuiteCRM advice in my experience. They get confused between SugarCRM, SugarCRM CE and SuiteCRM and because there are so many releases of SuiteCRM the LLM’s hallucinate things that don’t exist (but think it should) and often gives outdated info or info thats applicable to SugarCRM not SuiteCRM.

1 Like
5

Hello,

a SuiteCRM installation usually takes a few minutes and works extremely smooth, if the server is correctly setup.
I prefer a vanilla LAMP system and had most issues with web admin interfaces, shared hosting and “special / secure servers”.

SuiteCRM is self hosted and requires technical knowledge and that should be solid (you’re working on the brain / core data of your company and its revenue generating processes).
There are plenty of SaaS offerings that don’t require any technical knowledge - that is one option.
You can always purchase the hosted SuiteCRM from the website itself and/or approach the service providers around SuiteCRM to help with setup / hosting / implementation.

1 Like
6

Thanks for the responses.

I have started a fresh install and am focused on getting the foundations right, following the official instructions.

I’m up to step 4 (update the vhost configuration). I’m stuck for now, so will come back to it when I have a moment.

Bastian, thanks for the video link, I’ll look at that too. I’m on a zero budget for now and am in no rush, so I’ll plod along for the time being. I just want to set up a beta environment to begin with and have people who will be involved in that lined up already.

Thanks again.

7

Hi @rcword ,
Fresh installation is a good move.

From what you described earlier, it looks less like several separate SuiteCRM issues and more like the first installation got into a partially broken state because some foundations were not fully correct at the start. Once files, vendor dependencies, permissions, PHP version, vhost path, and database schema start getting manually adjusted one by one, it becomes very difficult to know what the real cause is and what is only a side effect.

For a clean SuiteCRM 8 install, I would suggest keeping it very simple:

1. Start with a clean Ubuntu/LAMP setup.
2. Follow only the official SuiteCRM installation guide for your exact version.
3. Make sure Apache points to the correct public directory.
4. Confirm the required PHP version and extensions before running the installer.
5. Avoid manually adding missing database columns or copying individual files from another package unless you are 100% sure why they are missing.

The missing `totp_*` columns and missing files suggest the install may not have completed cleanly, or the package/dependencies were not in a consistent state. In that situation, I would avoid trying to repair the existing installation and would focus on getting a clean install completed without manual patches.

LLMs can be useful for explaining concepts, but for SuiteCRM installation commands, they can easily mix up SuiteCRM versions, SugarCRM, old SuiteCRM 7 behaviour, and SuiteCRM 8/Symfony behaviour. For installation, the official docs and community guidance are much safer. This is something I have observed personally, and as a domain expert, we ensure that we get the correct solution with proper syntax, etc.

Since you are at the vhost step now, I would recommend sharing your Apache vhost config, SuiteCRM folder path, and Apache error log if you get stuck again. That will make it much easier for the community/me to point out the exact issue.

8

Thank you for everyone’s support. I’m afraid my Oracle instance is now just 1 OCPU with 1Gb memory now, so I have to find an alternative before taking this further. I’ll revisit this then.

Thanks again and all the best.

9
10

Thanks. I did have 6Gb on Oracle, ytried to change it and ended up with a smaller offering… DOH! For now I’ll continue using SuiteCRM 7 in MAMP. If my pilot is successful, I’ll get a paid VPS and will install SuiteCRM 8 on there.