Could not find driver error - Docker version

Self-Hosted | v5
1

Suddenly getting an error trying to start the container. This is the error I get each time I try to start.

2023-06-03T13:57:07Z [INFO] [Entrypoint]: Initialising Invoice Ninja...

   INFO  Configuration cached successfully.  


   INFO  Caching the framework bootstrap files.  

  config ........................................................... 41ms DONE
  routes .......................................................... 124ms DONE


In Connection.php line 760:
                                                                               
  could not find driver (SQL: select * from information_schema.tables where t  
  able_schema = ninjav5 and table_name = accounts and table_type = 'BASE TABL  
  E')                                                                          
                                                                               

In Connector.php line 70:
                         
  could not find driver

The container exits immediately so I canā€™t really try any other troubleshooting. Iā€™m using a MariaDB back end.

2

Hi,

Maybe this will help:

3

Wouldnā€™t a change like that need to be done every time the container is updated? Itā€™s not my container itā€™s the one from the unRAID apps.

4

@david do you have any thoughts?

5

Looks like the system is missing the mysql-client library to run the migrations.

6

Is that a one time thing thatā€™s needed?

7

@nironwolf

It is used to perform database migrations, so typically each release may contain migrations that are needed for database updates, so it should always be present.

8

Makes sense. The unRAID app template is pointing to this version maihai/invoiceninja_v5 which seems to be updated nightly. Is this provided by InvoiceNinja or a 3rd party?

9

Iā€™m not familiar with that link, the official Dockerfile is here:

https://hub.docker.com/r/invoiceninja/invoiceninja

1 Like
10

So I setup everything with the official docker container and I get the same error message.

I had been running along updating MariaDB and I think maybe and update from the 10.6 to 10.11 back end could be causing this. Should I roll back my updates for MariaDB? I had been faithfully hitting update out of security concerns, but itā€™s not exposed to the outside world so I may be ok with reverting. However none of the other software that connects to it seems to be having issues, so Iā€™m a bit confused.

11

Running php -m in the container I get this output:

[PHP Modules]
Core
ctype
curl
date
dom
fileinfo
filter
ftp
hash
iconv
json
libxml
mbstring
mysqlnd
openssl
pcre
PDO
pdo_sqlite
Phar
posix
readline
Reflection
session
SimpleXML
sodium
SPL
sqlite3
standard
tokenizer
xml
xmlreader
xmlwriter
zlib

[Zend Modules]

I see that mysqli specifically isnā€™t on the list. But I donā€™t know enough about php modules to know if mysqlnd is an equivalent. I also donā€™t see ā€˜pdo_mysqlā€™ which is mentioned in the stack exchange link you sent.

So I also tried setting up a new MariaDB user and a fresh, empty database. It gives the same error.

12

After setting up new containers, restoring databases, downgrading database, playing with .env files for the APP_KEY to fix MAC errors, and moving the old documents folder around and chown-ing it to the right user:groupā€¦

Itā€™s working again. But only if I use the docker container for 5.5.123. It runs fine with the original MariaDB container running 10.11. But if I update it to 5.5.124 the connection error returns.

tl;dr Had to downgrade to 5.5.123 (Which I couldnā€™t do with the container in the unRAID App Store because he doesnā€™t make anything but a ā€œlatestā€.)