SeAT - forum: Migrations fail when installing mattfalahe/manager-core - Page 1

Hey folks, New to SeAT (previously used AA but thought I'd try this). I did an install with Docker this weekend and it went well. I then added mattfalahe/structure-manager and all was well. I then added mattfalahe/manager-core and it failed with a migration / db error:

front-1      |   2026_01_01_000005_create_manager_core_appraisal_items_table ....... 4ms DONE
front-1      |   2026_01_01_000006_create_manager_core_plugin_registry_table ...... 27ms DONE
front-1      |   2026_01_02_000001_create_manager_core_market_prices_table ......... 2ms FAIL
front-1      |
front-1      | In Connection.php line 829:
front-1      |
front-1      |   SQLSTATE[42S01]: Base table or view already exists: 1050 Table 'manager_cor
front-1      |   e_market_prices' already exists (Connection: mysql, SQL: create table `mana
front-1      |   ger_core_market_prices` (`id` bigint unsigned not null auto_increment prima
front-1      |   ry key, `type_id` int not null, `market` varchar(50) not null, `price_type`
front-1      |    varchar(10) not null, `price_min` decimal(20, 2) not null default '0', `pr
front-1      |   ice_max` decimal(20, 2) not null default '0', `price_avg` decimal(20, 2) no
front-1      |   t null default '0', `price_median` decimal(20, 2) not null default '0', `pr
front-1      |   ice_percentile` decimal(20, 2) not null default '0', `price_stddev` decimal
front-1      |   (20, 2) not null default '0', `volume` bigint not null default '0', `order_
front-1      |   count` int not null default '0', `strategy` varchar(50) not null default 'o
front-1      |   rders', `updated_at` timestamp not null) default character set utf8mb4 coll
front-1      |   ate 'utf8mb4_unicode_ci')
front-1      |
front-1      |
front-1      | In Connection.php line 587:
front-1      |
front-1      |   SQLSTATE[42S01]: Base table or view already exists: 1050 Table 'manager_cor
front-1      |   e_market_prices' already exists

The table is seemingly trying to get created twice...? I thought this might be a DB consistency issue, so I nuked the volumes and recreated it from fresh but with both structure-manager and manager-core listed as plug-ins from the get-go. Same result. Am I caught in an inconsistency due to these modules having new releases dropping right now? Or is this something else I borked? I changed nothing from the default install instructions, and I'm running on Rocky Linux 10. Any suggestions? Thanks!

manager core is not released yet.

just remove from .env and u should be fine

ah ha, okay, thanks for the quick reply! Yeah it worked without it. But the UI itself prompted me to install it

yes because I should have released it yesterday but I didnt had enough time

Alright! No worries then, and I guess I just chose the day to try it when you were half way through releasing

Appreciate the work and the clarification

it look like u did created some tables though

what did u put in your env? just mattfalahe/manager-core ? or did u add :dev-dev to it?

SEAT_PLUGINS=mattfalahe/structure-manager,mattfalahe/manager-core

I also tried it with the reverse order, in case that mattered

order doesnt matter

just give me a sec I will check main branch and will help u delete whatever u have in your tables created

It is a fresh install for testing, so I am happy nuking it

so when you will want to install the release it will go ok

I will stick with the regular structure-manager for now and wait for the release announcement

SM on its own works perfectly fine. MC is just an addon that will speed up your notifications + it will also share additional data with my other plugins

Cool

I'm new to the SeAT ecosystem but thought this looked cool

We use AA mainly but are thinking about switching

at the moment I am just testing if my promise of 2 minutes holds in edge cases. This is an example so all notifications are between 20s up to 2 minutes

that is with MC installed

Do you have to actually go toast sovhubs to test this? Or can you simulate it?

no we just have some events like those so its just easier to setup and check time between eve - seat and eve - MC in parallel

SM has diagnostic page that can test this but its aways better to test on real events.

diagnostic page is accessible only with admin permission to the plugin at structure-manager/diagnostic

okay

Cleanup after faild MC run docker compose exec mariadb mariadb --user=seat -p --database=seat Password to your DB is in .env file.

-- =============================================================
-- Wipe Manager Core (main) install — exact-names version.
-- =============================================================

-- ---- 1. Drop the 6 tables main's migrations create ----
-- FK_CHECKS off in case appraisal_items has an FK to appraisals;
-- IF EXISTS so partial states don't error.
SET FOREIGN_KEY_CHECKS = 0;
DROP TABLE IF EXISTS manager_core_appraisal_items;
DROP TABLE IF EXISTS manager_core_appraisals;
DROP TABLE IF EXISTS manager_core_market_prices;
DROP TABLE IF EXISTS manager_core_plugin_registry;
DROP TABLE IF EXISTS manager_core_price_history;
DROP TABLE IF EXISTS manager_core_type_subscriptions;
SET FOREIGN_KEY_CHECKS = 1;

-- ---- 2. Remove the 12 migration-history rows main wrote ----
DELETE FROM migrations WHERE migration IN (
    '2026_01_01_000001_create_manager_core_market_prices_table',
    '2026_01_01_000002_create_manager_core_price_history_table',
    '2026_01_01_000003_create_manager_core_type_subscriptions_table',
    '2026_01_01_000004_create_manager_core_appraisals_table',
    '2026_01_01_000005_create_manager_core_appraisal_items_table',
    '2026_01_01_000006_create_manager_core_plugin_registry_table',
    '2026_01_02_000001_create_manager_core_market_prices_table',
    '2026_01_02_000002_create_manager_core_price_history_table',
    '2026_01_02_000003_create_manager_core_type_subscriptions_table',
    '2026_01_02_000004_create_manager_core_appraisals_table',
    '2026_01_02_000005_create_manager_core_appraisal_items_table',
    '2026_01_02_000006_create_manager_core_plugin_registry_table'
);

-- ---- 3. Remove the 2 scheduled commands main's ScheduleSeeder wrote ----
DELETE FROM schedules WHERE command IN (
    'manager-core:update-prices',
    'manager-core:cleanup'
);

log into db

paste the code above

and you will be clean before MC install when its ready

Verification

-- Should return 0 rows each:
SHOW TABLES LIKE 'manager_core_%';
SELECT * FROM migrations WHERE migration LIKE '%manager_core%';
SELECT * FROM schedules WHERE command LIKE 'manager-core:%';

Nice

Actually I'll test that now

let me know if that worked

I am no expert at this

It seems to but then it seems to recreate the tables

did u remove manager core from .env?

then restarted cleanly?

ah

after that run that db ?

ok let me check

u need to stop your stack first because it got stuck in the loop for now

so u delete and composer tries again to create

Yep, confirming it worked

Once I removed the manager-core from .env

MariaDB [seat]> SHOW TABLES LIKE 'manager_core_%';
Empty set (0.001 sec)

MariaDB [seat]>
MariaDB [seat]> SELECT * FROM migrations WHERE migration LIKE '%manager_core%';
Empty set (0.002 sec)

MariaDB [seat]>
MariaDB [seat]> SELECT * FROM schedules WHERE command LIKE 'manager-core:%';
Empty set (0.002 sec)

This is after it was stuck in the loop

Thanks so much for this. And sorry for taking you away from actually releasing it 😂

no worries

always happy to help