Hello everyone, if we try to move a post from one thread to another, we sometimes get a ‘500 internal server error’ Is there some tutorial how to troubleshoot issues like this? Where i can find the log files to get details about the issue? Thx für every reply :slight_smile: Greetings, Ronny

[image] Moving posts returns 502 bad gateway Self-hosting Okay, so the error 500 should be fixed on latest and I just commited another fix so that the actual error appears in logs in case there’s another bug in there. If you get an error 502 then the moving takes too long and times out. I made lots of changes over the last couple of months to move more metadata when moving posts. I tried to optimized it, but it highly depends on the amount of data and might still be too resource intensive for some server to do all the work in under 30 seconds. The so…

We are also seeing 500 errors when moving posts. I managed to trigger it by moving a post to another topic, then attempting to move it back to the original topic. Here are the details: SyntaxError: Unexpected token ) at t (_application-1aa5dba826bc3ed623e07f37fc542db9746ecf4a5fcf6b519e5921036bface8d.js:12468) at _application-1aa5dba826bc3ed623e07f37fc542db9746ecf4a5fcf6b519e5921036bface8d.js:34707 at f (_ember_jquery-d430881a3fb1403871256e5a02423c4b20a78793685e92088613ca9a701baf88.js:70136) at T (_ember_jquery-d430881a3fb1403871256e5a02423c4b20a78793685e92088613ca9a701baf88.js:70309) at E (_ember_jquery-d430881a3fb1403871256e5a02423c4b20a78793685e92088613ca9a701baf88.js:70295) at y (_ember_jquery-d430881a3fb1403871256e5a02423c4b20a78793685e92088613ca9a701baf88.js:70231) at t.invoke (_ember_jquery-d430881a3fb1403871256e5a02423c4b20a78793685e92088613ca9a701baf88.js:64637) at e.t.flush (_ember_jquery-d430881a3fb1403871256e5a02423c4b20a78793685e92088613ca9a701baf88.js:64527) at e.t.flush (_ember_jquery-d430881a3fb1403871256e5a02423c4b20a78793685e92088613ca9a701baf88.js:64731)

Thank you for your reply! In the meantime i found the causing problem in the Table post_timings in the database. In this table is stored, to which post the users already read in the threads. Under (for me) unknown circumstances it can happen, that the server tries to write entries in the table, that already existing. I’ve searched in the logs and found a hint to this database table. Unfortunally i don’t know anymore, which logs i searched… :confused: The solution for me, was simply to delete the content of the table. All users lost their progress of reading threads, and some people had to write a new entry in existing threads just to overwrite browser cached stuff.I have a very little forum with just approx. 20 users, i don’t think, this is applicable for a forum with more users :wink: Greetings Ronny

Thanks Ronny, and yikes! I don’t think my 10k users will be very happy about losing their read history. I’ve also received reports from many users that their posts are not being marked as read properly in the past weeks. This explains that too, I guess. Is there any way to clean this up without deleting the table, @sam or others? :slight_smile:

[image] DiscourseMetrics: posts are not being marked as read This resolved bug probably explains the unread states: [image] Topics incorrectly stuck in unread state on Android in Chrome 77 Bug Per: Unread tab does not update after reading There were quite a few people mentioning issues with topics not being marked read on Android. We have done some root cause analysis and found that the page visibility api changed the way it operates in Android Chrome 77. It was discussed in brief here: https://github.com/pgilad/react-page-visibility/issues/9 There is a fancy graph that explains the new lifecycle here: The symptoms of the issue is. You use your android You flip to do som…

I just tried to reproduce this error in my local environment. I was able to move a post to a new topic and then move it back. [image] rldml: In the meantime i found the causing problem in the Table post_timings in the database. In this table is stored, to which post the users already read in the threads . Under (for me) unknown circumstances it can happen, that the server tries to write entries in the table, that already existing. I’ve searched in the logs and found a hint to this database table. Unfortunally i don’t know anymore, which logs i searched… I can’t tell if the commit being mentioned here is causing the error. It doesn’t alter the post_timings table and it’s part of an experimental feature we’ve been working on (disabled by default). Just to be sure, I tried moving posts when the feature is enabled but it also worked correctly. If someone could share some logs, that could be really helpful.

I take a look, if i have my old logs about the datatable error. But the deletion of the table content fixed the problem for sure (with the side effects i mentioned)

Sorry, i’ve searched for it, but cannot find it anymore… :confused:

For what it’s worth, I’m also getting this 500 error when trying to move a particular post from one thread to another. I haven’t tried anything fancy, so if someone from Discourse wants logs, let me know what to look for and I’ll see if I can do that.

Thanks, I’ll try it when the next beta ships.

Moderator move post, get 500 internal error, how to troubleshoot

Product Support

Remah (Just another happy Discourse user) October 2, 2019, 12:25pm 5

This resolved bug probably explains the unread states:

Topic		Replies	Views
Error 500 when moving posts Bug	9	527	June 22, 2026
Can not move any reply into an existing topic Bug	4	1062	October 9, 2019
Moving posts returns 502 bad gateway Self-hosting performance , server-resources	71	6438	January 12, 2025
500 Error: There was an error moving posts Bug	3	463	October 25, 2023
"There was an error moving posts" Bug	2	191	December 12, 2024

Moderator move post, get 500 internal error, how to troubleshoot

Related topics