Michael Collins Posted June 11 Posted June 11 After upgrading the front end of my site seems to have lost all styling: https://youririshheritage.com/community/ I get this console error: Uncaught ReferenceError: ips is not defined at community/:133:4 Any ideas what went wrong and where to go from here?
Jim M Posted June 11 Posted June 11 You'll want to try and clear your cache by going to ACP -> Support -> Clear System Caches The ACP may look odd but it will function.
Michael Collins Posted June 11 Author Posted June 11 @Jim M Thanks but I've already tried that a few times now, and again just now, but it hasn't solved it. Also cleared caches on the server too.
Jim M Posted June 11 Posted June 11 1 minute ago, Michael Collins said: @Jim M Thanks but I've already tried that a few times now, and again just now, but it hasn't solved it. Also cleared caches on the server too. You will want to check your S3 bucket. All references are getting a 403 Access Denied. Which means either the files aren't getting created or you are caching the whole page and referencing old URLs.
Michael Collins Posted June 11 Author Posted June 11 @Jim M ok thanks, any guidance on what to check in particular?
Jim M Posted June 11 Posted June 11 1 minute ago, Michael Collins said: @Jim M ok thanks, any guidance on what to check in particular? This would be outside our scope of support as it is S3 related and hosting related. You would want to ensure that your S3 bucket is writable and readable by the software plus readable publicly, if you are not using a CDN on your bucket.
Michael Collins Posted June 11 Author Posted June 11 (edited) @Jim M Ok, but as far as I am aware we are just using AWS to store images/uploads, our developer set it up last month. Why would this be causing the issue we're having now after upgrading? Edited June 11 by Michael Collins
Jim M Posted June 11 Posted June 11 3 minutes ago, Michael Collins said: Why would this be causing the issue we're having now after upgrading? Likely not setup correctly and the first time which these resources have been rebuilt. Unfortunately, the permissions of S3 is beyond our scope of support.
Michael Collins Posted June 11 Author Posted June 11 What I meant was, is that definitely the cause of the issue as I don't understand why image storage would cause that. I see now though that the "Theme resources" are also there, I'm guessing thats why the front end looks so strange? Switched it back to the "file system" option and it looks back to normal, thanks for the help, will have to try sort out AWS.
Jim M Posted June 11 Posted June 11 2 minutes ago, Michael Collins said: I see now though that the "Theme resources" are also there, I'm guessing thats why the front end looks so strange? This would be correct. You have not stored just images there. Michael Collins 1
Michael Collins Posted June 24 Author Posted June 24 Hi again, Upon investigation here we noticed that the S3 folder was still connected but the files had been removed, I have performed a manual re-upload of the media and it is now working again (the connection was fine all along by the looks of things). We're trying to understand how this occurred to avoid it in the future as it still appears to have been something to do with the update itself. We updated Invision around 10am on 11/6/24 and switched the storage settings back over to local sometime after 5pm on that someday when we thought we couldn't re-establish a connection. The S3 bucket timeline shows us that the bucket was mostly deleted by 10.23am on 11/6 (after the update but before we had changed any settings) and upon review of the database I have found entries for "container_deletion" with a timestamp of 1711368417 (Tue Jun 11 2024 10:51:41 GMT+0000). I know there is a discrepancy here with regards to the timestamps but perhaps this is servers in different timezones, not sure. What is clear to us is that the bucket was mostly deleted ahead of us changing settings within the storage settings, seemingly triggered by the update itself. I'll attach a couple of screenshots for reference but I hope this is something you can take a look at on your end and advise if anyone else has reported anything similar. If the update has somehow triggered a delete call to the bucket that's obviously a big problem for us. Thanks
teraßyte Posted June 24 Posted June 24 1 hour ago, Michael Collins said: If the update has somehow triggered a delete call to the bucket that's obviously a big problem for us. The latest version fixes a bug I reported a few months ago: Files marked for removal from the bucket were never actually removed because the task wasn't activated properly. In your case clearing the system caches from ACP should have rebuilt the CSS files if they were deleted by error (it shouldn't have happened, though).
Recommended Posts