The symptoms that you are describing sound like the users do not have a valid mapping to the PI Data Archive. In order for a user to be able to search for a tag, that user must be mapped to through a mapping to an identity with read permissions to the PI Tag.
If I were to guess, it sounds like users are seeing data through a trust instead of a mapping. You can use Network Management Statistics under SMT>Operation>Network Management Statistics to see if the user is connecting as trust or mapping when they are getting data from PI Vision (listed under the w3wp.exe application).
To create a mapping for new users, you can use SMT>Security>Mappings and Trusts.
Thanks for the answer.
Actually the users are mapped to piadmin, so they should have enough permissions. Network management Statistics shows that they are logging on with piadmin.
We updated PIVision and also changed the useraccounts running pi web api, pi web api crawler services and the application pools.
It seems to me like this change somehow werent succesfull, and this causes the crawler not to crawl for some reason. (The users have no problem accessing data from datalink or an existing display)
Could you please let us know if you are running PI Web API with a service account?. Would suggest you to run similar account to PI vision and check if you are able to access PI and AF in PI vision. If you are using service account then it should have access to PI Archive database(Point and Data Security)
Also request you to login to PI Vision Admin website and check Data Server Allowed section in Overview and PI User mapped.
I will refer you to this KB article for you to get your crawler working.
Do let us know the cause if you are able to resolve it.
Have you set the Kerberos delegation set up for PI Vision server? In order to properly impersonate users account priviledge by PI Vision's service account, the following portion of configuration needs to be done. Creating PI Mapping is only the first step. Also, Users need to authenticate to PI/AF using WIS before PI Trusts. PI Web API Crawler requires users to be mapped to PI Mappings and the auth. order may be important.