Is it intentional that submitted edits that show up in the edit history do not display on the website for minutes? This stale data behavior started within the past week, and makes editing almost impossible.
It’s gotten so bad now that the edit system is throwing phantom failed dependencies:
edit #153293950
edit #153295724
And MagicISRC sometimes claims “The ISRCs have been successfully submitted.” but they did NOT make it into the edit history…
I noticed this as well, I’d never seen edits “lag” behind the live site before.
Earlier today some of my edits actually failed to go through at all, in the same edit as this one (which took some refreshes to see applied in the overview page) I added two other url relationships to that artist, but neither of those went through.
My best guess is that this is probably related to the ongoing ai scraping / server overload issues, but this is the first time I see something like this happen.
I’m also having significant issues with this, primarily seems to hang on barcodes, but maybe that’s just because that’s what I am focusing on at the minute.
It has nothing to do with barcodes, it’s the entire database:
Yes, server must be overloading or something. Yesterday was particularly slow for me as well but all my edits seem to have gone through.
The only thing that was really giving me problems was uploading cover art.
I’ve never seen this type of server lag before either. If you enter an auto edit and then hit F5 a bunch of times it appears to be completely random whether the change will be displayed or not.
I guess there are two MB servers in load balancing, and only one of them is late.
There appears to be some caching involved. ISRCs are typically not available in Picard immediately but have some latency, like up to a minute. That’s how it works for me for as long as I use them (around a year).
Regarding load balancing, if there is one it must be async for this to manifest shortly.
The same goes for barcodes. When I used MET, if I wanted it to recognize a release, I had to add it on MB side, but then It is invisible to MET up to several minutes.
So, I think caching, as a timespan is too long to be load balancing.
There was a duplicate cache container running on another server for the past 48 hours which was not intended. I removed it and flushed the cache on the other, so things should be back to normal. I’m sorry for any failed edits/frustration this caused!