Weird problem with permalinks

The CMS Questions
1

Hi,

I’m experiencing a strange issue with permalinks.

I have two environments, UAT and Production. Content is copied from UAT to Production via rsync. The content and files are identical; I have verified this.

In UAT, I have a page link in some text field /@/page/hjwkfhvkdvlya44x, and this link works. If I click it, I am redirected to that page.

In Production, I have the same page with the same link (because it’s rsynced), but there, the same link returns a /error.

Even stranger: Another permalink on that same page works, but many of them don’t. They appear the same, and the content is the same. I have no idea what is happening here :sweat_smile:

Has anyone experienced similar problems and could offer any tips? :sweat_smile:

Cheers!

2

Permalinks only resolve the UUID from the UUID cache. Please make sure that the UUID cache is populated on both ends, e.g. with syncing or Uuids::generate() | Kirby CMS

3

It’s synced, i validated all files on both ends, they are identical.
When I use permalinksToUrls it works (so I have a temporary solution)

4

Have you tried repopulating the UUID cache on the server? Cause it really sound alike that is the source of your issue.

(Cause permalinksToUrls doesn’t have that limitation, that will actually loop through your content files if it isn’t cached)

5

Random thought: are the two environments identical? Because, for example, one issue I sometimes stumble into is that my local dev environment treats upper/lower case differently than my actual server.

6

I’ll doublecheck later today when I can access the server again. Thx!

7

It’s a ploi with build in staging, which means it are basically 2 folders on the same server and rsync so I doubt that this can be the problem :sweat_smile: Thx for thinking along though!

8

doing kirby uuid:populate didn’t change anything. (also tried generate but didn’t change anything also, the pages already have a uuid)

Normally I’m good in fixing problems like this but can’t really wrap my head around this one. There’s no reason why it shouldn’t work :man_facepalming: It works locally, it works on uat, it just don’t works on prod (which is basically a site on same server as uat :sweat_smile:)

9

are you sure the cli command is flushing the cache before populating?