Duplication after multiple imports - isn't this a bug?
As an organization, we have multiple people maintaining/adding/managing what is currently about 50 vaults - one per client - and each with many entries to manage that client's credentials for all their vendors. The vaults have between 20 and 100 entries each and each entry may correspond to many accounts, sometimes hundreds of accounts. As we set up accounts, we also have automation that accesses the credentials for each account through a 1P Connect Server. To do this, each account has the UUID for it's corresponding credential entry configured in our automation jobs.
On the business side, they have mistakenly submitted duplicate data to be added to 1Password. This duplicate information was loaded into the 1Password vault through the "Import Data" function. That import completes without error - even when the imported entries duplicate existing entries. This is not the behavior that I would expect. For each entry in the import, 1Password creates a new entry with an identical Title but with a new UUID.
In the 1P user interface, the system will not display these multiple entries. The UI displays only one entry with a specific Title, so the duplication is hidden. It appears there is no way to access the original entry through the UI. Worse, there is no indication that there is any duplication in the system. The UI just displays the most recent entry for each Title without any notice of duplication.
I've verified the existence of the duplicates by exporting the vault data through using the CLI.
This is a serious problem for us because it breaks our Integrations. The carefully constructed automations fail when there is a credential update, because those updates are now stored in the most recent copy of the entry but the automation is still referencing the UUID of the original entry. This is only evident if we extract the UUIDs from our automation configuration and compare them to the UUID of the current entry being displayed in the UserInterface.
I don't understand this architectural design. If the UI will only display one entry with a specific title, then from my users' perspective, that title is effective acting as a primary key - even if the system is using the UUID as the actual primary key. When we access an entry by that title, we're expecting to see the change history for that item from the creation of the original entry. Why allow an import to override that without providing a warning? Or better yet, an option that would error out duplicate titles. I can imagine no scenario where we would want to allow ghost versions of earlier entries to persist in the database but never be accessible in the UI.
I know that we can access these older versions of an entry via the CLI, but that is extremely cumbersome vehicle for cleaning up old data. It requires I commit a developer to the job. We should've been warned before this mess was created.
Do you have any tools that can help with this clean up? Or suggestions about how to approach this?
And I request that the import tool be modified to allow a user to block the creation of entries with duplicate Titles. I suggest there be a switch to ask if an importer wants to allow the creation of duplicate entries and I suggest that it should default to 'No'.
Thanks for any insight you can provide.
1Password Version: 8.10.3012
Extension Version: 2.8.1
OS Version: Windows 10, v21H2
Browser:_ Chrome
Comments
-
Hi @mhiggins . I'm sorry to hear you are going through such a frustrating experience. I can't seem to reproduce the issue where a duplicated item, only shows one of the items with the same title. To clarify by UI, are you referring to the 1Password desktop app or the 1password.com web app. Additionally, when duplicating these items were they placed into a new vault or the same vault? Providing this additional info will hopefully help us debug your issue. In the meantime, unfortunately the best solution to removing these duplicates would be use the CLI tool.
0
