[bug] Timeline Internal server error with no logs #508
Labels
No labels
approved, awaiting change
bug
configuration
documentation
duplicate
enhancement
extremely low priority
feature request
Fix it yourself
help wanted
invalid
mastodon_api
needs docs
needs tests
not a bug
planned
pleroma_api
privacy
question
static_fe
triage
wontfix
No milestone
No project
No assignees
5 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: AkkomaGang/akkoma#508
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Your setup
OTP
Extra details
Ubuntu 22.04
Version
2023.02 (3.6.0-0-g71d0899)
PostgreSQL version
14.6 (Ubuntu 14.6-0ubuntu0.22.04.1)
What were you trying to do?
Load the home timeline or that of a specific user that seems to cause it. It again seems to be a user from bird.makeup and unfollowing that user makes my home timeline work again.
What did you expect to happen?
The timeline to load.
What actually happened?
UI shows "Error fetching timeline: Internal Server Error" and /api/v1/accounts/{id}/statuses and /api/v1/timelines/home return 500:
Logs
There is nothing in the server logs for this.
Severity
I cannot use the software
Have you searched for this issue?
without more information about the user in question it will be impossible to act on this
we'd need to know the data formats that were in use
i also occasionally have this problem. I'll see if i can grab some logs the next time it happens. the issue will usually clear itself up after more posts are federated, so it must be something with a post corrupting a specifc piece of db data
Sorry but it's quite difficult to get more information because there is nothing in the server logs and any attempt to retrieve or even determine the ID of the post causing the issue is also broken.
I can provide which user I'm seeing it on: https://headpat.cafe/users/ATLSmTvkPCuwrIqwFM (warning NSFW content)
it's retweets
lookie lookie
if we don't exclude retweets we get the error:
https://headpat.cafe/api/v1/accounts/ATLSmTvkPCuwrIqwFM/statuses?with_muted=true&exclude_replies=1&limit=20&exclude_reblogs=false
but if we do, we don't get it:
https://headpat.cafe/api/v1/accounts/ATLSmTvkPCuwrIqwFM/statuses?with_muted=true&exclude_replies=1&limit=20&exclude_reblogs=true
so something about how this bridge handles converting retweets is messing up the rendering - it's probably giving us a null user or something
i'll have to figure out how to make it show me the AP for that
can you run the following SQL and send the output across?
can you give me guidance on how to do that on an OTP install?
psql akkoma
(or psql akkoma_prod, or whatever you called your database)then just paste it in
Here is the output:
(16 rows)
wonderful, thanks a lot!
that should assist, i'll try loading it into my debug instance and trying to reproduce
i am sadly not able to replicate this in any capacity
all posts load and render just fine
you may want to turn up your logging to debug, that may help
Right. I actually have the same now. I went and had dinner, server has been up and running flawlessly for 4 days straight, when I came back to computer from dinner it can't load home timeline. Notifications is loading. Right out of the blue this happened.
I tried to re-apply the latest stable update (according to official guide, which I was already on, but hey, had to try), including doing the front-end updates and have restarted the server twice. Same thing.
No idea what to do next.
again with no data to go on there's no way anyone can help
turn logs up to maximum, see if the API route returns anything with reblogs turned off etc etc
As I was looking in the log and noticed a few strange things I got annoyed by the fact it said "pleroma" (yeah, I know, a little thing) so I figured I should change that in admin-fe too, and reload the logs. I changed it to "akkoma" and then just went to verify it didn't work still and...well...it did. The home feed that is. It was back to working. No idea what happened. Sorry for being quick on the trigger here. I've had wine, is that a decent excuse...?
Trying that same endpoint again and now I'm getting an error in the logs. It seems to be the same error as in #482 so maybe that was it the whole time.
Also ran into a some errors today. Here's what popped in my logs at the time:
@floogulinc Based on this thread and problem description, I could say this could be caused by the same underlying reason as this one:
#487
My timeline is once again unusable with the same errors in the log as in #482.
@floatingghost would you be able to make a new OTP release that includes the fix for this issue from
0c77be9308
since it wasn't included in 2023.03 and this basically breaks the main functionality of the software?it's on the develop branch, you're free to switch to that if you think it'll help