In some cases, a database error may occur. In the application, such an error usually manifests as an error message related to SQL, indexes, a database procedure, or similar issues.
Troubleshooting procedure
An important step before making any changes to the database is to back up your data. We recommend backing up regularly (own/local server; for cloud, we handle backups on our side), and ideally keeping older backups available as well.
We recommend creating a backup copy of the entire PostgreSQL data directory.
Perform a disk integrity check (check disk, fsck, ...)
The next step is to repair PostgreSQL — reindex, remove duplicate records, etc. (In isolated cases, this step can be performed on our end as a service intervention. However, the resolution should be carried out by your database specialist.)
The final step is to verify functionality by creating a current backup of the billed company in Flexi and performing a test restore from the backup.
Possible causes
In isolated cases, a PostgreSQL service outage has occurred at the client's site
Windows Server 2022 Standard
The log displays a general message: "WinStrom-PostgreSQL was unexpectedly terminated."
A standard service restart (server and PostgreSQL) did not resolve the issue.
Ultimately, the issue was resolved by starting PostgreSQL manually using pg_ctl start and waiting approximately 10–15 minutes for it to start. The default timeout of 300s was not sufficient. The service was subsequently able to start successfully.
ERROR: could not create unique index
The issue is caused by a change in local storage following an OS update — PostgreSQL currently has no way to detect an incompatible glibc update (source). As a resolution, it is usually sufficient to clear the wcolumn table.
Error example:
ERROR: could not create unique index "ddoklint_kod_key" Detail: Key (kod, modul)=(PE+0001/2021, POK) is duplicated.: updateDatabaseInternal