Not sure if this is a bug or not.
Since we updated to v1.8.0.beta5 +68 we’ve been seeing some occasional failures to load new posts.
So far all I can say is:
- It happens with new posts only.
- It seems to happen when you click on a new post immediately after the notification appears.
- It is the same for public as well as private posts.
- The post is stored, it just doesn’t get rendered during the page display, so you see everything normal… just without the post or topic that you want to see.
- Unrelated to how many users are logged on.
- Refresh makes no difference.
- Happens on Edge and on Chrome.
- Waiting 10-30 sec after the notification appears, or if a page comes up empty then just going to to any other topic, waiting 10-30 seconds and trying again will load the page.
Is anyone else seeing the same thing?
Have you noticed this on meta? What plugins do you have installed?
This is on our site forums.jag-lovers.com (v1.8.0.beta5 +68)
Discourse Whos Online
(new) indicates the plugins we added when we did the update.¨
/logs shows a couple of these:
TypeError: e.attrs.userCustomFields is null
Window Location: https://forums.jag-lovers.com/t/xk120-fhc-dhc-floor-rear-corner-clips/351224
But nowhere near as many of them as I have had it happen in the last 24 h or so.
If you run in Safe Mode, does the issue go away?
Just trying that now. Of course, the meatbags are all on break so it may take a moment for some new posts to arrive.
Grumble grumble grumble one meatbag insinuates that the system swallowed his post and they all go silent…
Still in safe-mode, but they’ve all gone silent.
This is almost certainly the third party plugins you are using. I don’t recommend using third party plugins.
So, the empty posts thing seems to have gone away on it’s own. It was intermittent, but fairly regular, but for the last 16h I haven’t seen it, and only a couple of other users saw it last night - one of whom was running in safe-mode, which would rule out plugins, no?
(insert Bob Hoskins voice: “They don’t fix themselves Sir, no, they don’t fix themselves.”)
Some kind of timing issue from sidekiq job load after we updated from the earlier release?