[Bug]: Cannot open account after upgrading to v24.12.0 from v24.8.0

[DracoC77]

Verified issue does not already exist?

• I have searched and found no existing issue

What happened?

I attempted to import a QFX file from Chase credit cards into the wrong account, realized I was loading the QFX file into the wrong account and clicked cancel. Now anytime I try to navigate to my chase account (not the one I had open when I mistakenly tried to import the QFX file), it sends me to a fatal error screen with the following errors:

Note that I CAN get to the other accounts just fine, just when I click on that specific account it errors out:

"Fatal Error
There was an unrecoverable error in the UI. Sorry!"

TypeError: Cannot read properties of null (reading 'num1') at IC (https://--redacted--/static/js/index.JHeog5K2.js:98:43758) at https://--redacted--/static/js/index.JHeog5K2.js:98:123249 at Array.map () at https://--redacted--/static/js/index.JHeog5K2.js:98:123180 at Object.cM [as useMemo] (https://--redacted--/static/js/index.JHeog5K2.js:57:23922) at Ye.useMemo (https://--redacted--/static/js/index.JHeog5K2.js:10:6208) at yRe (https://--redacted--/static/js/index.JHeog5K2.js:98:123141) at v (https://--redacted--/static/js/useAccountPreviewTransactions.EORsFDeG.chunk.js:1:117) at Jo (https://--redacted--/static/js/wide.k1vKSsoJ.chunk.js:1:89660) at kk (https://--redacted--/static/js/index.JHeog5K2.js:57:19539)

How can we reproduce the issue?

Unclear root cause, UI fatal error each time I go to account

Where are you hosting Actual?

Docker

What browsers are you seeing the problem on?

Chrome

Operating System

Windows 10

[youngcw]

Looks like the db has bad values. Do you have a backup? That would be the easiest fix

[DracoC77]

I just tried an older image snapshot of my whole server and took a backup file from that (previously on v24.8.0), imported into v24.12.0 and looks like the problem happened when I went from v24.8.0 to v24.12.0...

In the dashboard after the update I can see a different summary balance for the troubled account, but when I go into the account it gives me the same error messages... so there's something interacting between my database for that account and v24.12.0 specifically.

[FIXED] Also trying to go back docker version to v24.8.0 gives me "Rejection: Error: Missing migration file: 1718889148000-openid.js" error when I try to start the docker container... even after I deleted all directories AND docker images (weird). [FIXED... I missed a .migrate file in the actual-data folder, doh!]

[DracoC77]

Ok was able to downgrade my actual-server to v24.8.0, my backup file from the same time-frame (v24.8.0) works just fine on the troubled account when running v24.8.0 actual-server.

When I try to load the latest backup file (recently upgraded/migrated from v24.8.0 to v24.12.0 in server upgrade) into v24.8.0 actual-server, the troubled account still gives me the same fatal error message.

Is it possible/worth trying to inspect the sqlite database files directly to look for the error? Any guidance on what to look for?

[matt-fidd]

Can you reproduce on https://edge.actualbudget.org? The trace looks awfully familiar from the schedules bug

[DracoC77]

Just spun up a new local instance of edge docker image and the trouble account seems to be working now. Will re-confirm on my "prod" setup before closing the bug

[DracoC77]

Confirmed that latest edge build seems to fix this problem with the "corrupted" database. Trouble account now loads normally without a fatal error. Based on the error codes likely a dupliciate of 4062