This is being caused by a conflict in addons. @Nifty-Chicken will need to report this to the addon authors:
Error: Call to undefined method SV\BookmarkEssentials\XF\Entity\User::canCreateImsSeries()
When? Now? Earlier today? Yesterday?
Everything looks fine for the last 6 hours in terms of what we can see.
I'm in the UK, and everything is loading instantly.
Looking at my external monitoring, something has been happening with the network being slow, but then as stated, it's been resolved around 6 hours now. It's more obvious from certain monitoring locations (San Francisco was really obviously impacted, but Amsterdam wasn't)
Nov 21 12:15:20 roost rg-listener: [e273e7091b] could not read request: WAN timeout
Nov 21 12:15:21 roost rg-listener: [45ef825df1] could not read request: WAN timeout
Nov 21 12:15:24 roost rg-listener: [307cdb89bb] could not read request: WAN timeout
Nov 21 12:15:35 roost rg-listener...
Also, as a precaution, I've just restored the database backup I took prior to starting any work this morning. There are only 5 posts in that thread:
MariaDB [DB1]> SELECT COUNT(*) as POSTS FROM `xf_post` WHERE `thread_id` = "1496543" AND `message_state` = "visible";
+-------+
| POSTS |...
No posts should have been deleted. The work done to fix the emojis was to replace all the old posts again prior to the import with their corrected versions. It shouldn't have deleted any posts, as it was a replace function with the unique post ID which is stored in the database.
That thread was...
The rebuild does appear to be fixing the issue:
https://www.backyardchickens.com/threads/the-nfc-b-day-chat-thread.1050266/page-10315
Posts in that thread are now in the correct order.
It's the same for everyone. Until I confirm the rebuild has been run and finished, it will still be happening. There are 1.4 million threads that need to be rebuilt.
I know why it's happening now. A thread knows it's first and last post ID, so when we restored certain tables information from yesterday, it also reset these values, so when people started posting again today, the posts are being marked with the post positions following on from the previously...
It will be threads which had posts from yesterday. As stated, I'm investigating, and I can see what the issue is, a thread and post cache rebuild should fix it, as each post in a thread is given a specific position ID.
OK, so I can see why it's doing that:
The post numbers are the same #103,132
But the post ID's are unique. I think we need to rebuild the thread and post information. There is a cache rebuild process for this available, which I'm testing on my own server now.