New PAT token format not working on first generation

[Maxime Lardier]

Following the recent update to PAT tokens I came across an issue where the first token generated after the update is not working.

I couldn't use the token to communicate with the API, I had an error :

"Basic authentication with passwords is deprecated. For more information, see: https://developer.atlassian.com/cloud/confluence/deprecation-notice-basic-auth/".

After looking here on the forum I saw a user that went through the same issue, it was solved by deleting and recreating the token. By generating new tokens I managed to make it work.

In the end any token I generate now (past 4 tokens generated) work as expected, but the first one generated still doesn't work.

Is this issue widespread ? Is there any way to track when this issue will be solved ?

[David Bakkers]

Hello @Maxime Lardier 

Firstly, you are asking a question, not starting a discussion, so this really should have been raised as question.

In response to the question, there is nothing to be 'solved' as the solution is... generate new tokens and they will have the required length to work properly.

Don't over-complicate the problem. One token is, for whatever reason, broken and there is nothing you can do to fix it, all you can do is delete it and generate a new one.

[Maxime Lardier]

Hello David,

The first token was already created with the new length and still wasn't working.

It is not one token broken. It looks like a lot (maybe all) first tokens generated in the new format are bugged, we were able to reproduce this issue on 3 different accounts. So there is indeed a bug on the creation of new format tokens and something to be looked at.

This thread is to track the issue and let other people know in case they encounter the same problem.

[David Bakkers]

You said "In the end any token I generate now work as expected" then you said "but the first one generated still doesn't work". So, only one was broken and all others were OK

Now you are saying "a lot are broken" and "maybe all are broken". First only one was broken, now all 'might be' broken, now a lot are broken, now newly generated ones do not work as expected.

I can't replicate your particular issue. All my old 24 char tokens still work as expected. All the new, longer, random length tokens I generate work perfectly. All the companies I work with who use Jira Cloud and are all generating the new, longer, random length tokens are not having this problem.

If you are having token generation problems with specific Jira Cloud instances, I suggest you raise a support ticket with Atlassian for them to investigate for those instances.

Good luck.