Thanks for the additional example! Jason is abroad for a conference right now, but I’ll meet with him when he returns so we can continue tacking this together.
Dang it! I take it back- not a zombie encounter after all. So the real zombies rae just those 4 posted above.
1 Like
No worries! Thanks for the update.
Hi Anastasia, hope you’ve had a good break. I finally got tired of all the zombie encounters and deleted them, thinking we haven’t seen one for ages, maybe its resolved. and then it’s like I triggered that with thinking it! here is the latest one SeadragonSearch.org
Based on the submitter and media asset names, it looks like this was the original encounter it came from: SeadragonSearch.org
When I searched by submitter email address, I noticed that the unassigned encounter with broken images has different dates for the created and submitted columns.
I’ll bring these examples to Jason when I meet with him next week to see if we can dig into what’s causing this.
Hi @Anastasia, yes I agree, this looks like the OG. Actually that difference between created (=observed- I HATE this misleading field name!) and submitted is happening heaps across the board. and not just when these are actually genuinely different. I often get one day out, which I chalked up to the international date line haha. its a bit annoying though. The search brings up the incorrect date usually and the encounter has the correct date. I am very leery of the search functions, I’ll allow! I dont trust it.
I also found the search table confusing, so I’ve made a ticket to propose renaming “Created Date” to “Encounter Date”.
Hey @NeridaWilson Did you delete this most recent encounter with the broken images? We can’t troubleshoot what went wrong if we don’t have the examples for the devs to work from.
Also, I have a suspicion that this might also be a zombie, albeit with normal images. this was submitted this morning…and again now. Not sure
Perfect; thank you! I accidentally treated the sighting ID as the encounter ID and mistakenly thought it got cleaned up. 
I’m sorry for the confusion!
I’m working with Jon to figure this out. We may do some testing that triggers test submissions when we do. I’ll be sure to label them as tests in the comments and get them deleted as soon as we have what we need. I’ll give you a heads up before this happens so you’re not caught by surprise.
I’ll let you know if I find out whether this is a related issue or just a user who accidentally submitted twice.
The pair of encounters for the deceased seadragon were just the user submitting it a second time. You can see it’s not an exact match because they forgot to select a location ID in the first submission.
1 - SeadragonSearch.org
2 - SeadragonSearch.org
Jon said he’d need to see more examples of zombie pairs to move forward. Looking at the logs of the latest pair (1 | 2) they both came via the report an encounter page and one gave an internal error for having corrupt images.
ok thanks @Anastasia. Do we need to leave that zombie sitting in there for now? Or has Jon learned all he can? Happy to leave, just need to know, thank you!
Please leave it for us! It helps the devs dig into the backend stuff. The more examples we can work from, the better.
Thank you! I’ve shared it with the devs and will report back.
Just for my own reference, this is the original encounter without broken images: SeadragonSearch.org
What’s curious about this one is that the original submission and the “zombie” were submitted a week apart.
For the two most recent encounter pairs we looked at, both appeared to come in as form submissions, meaning that the request to create that second encounter came from the user’s browser. We don’t know yet if this is what’s historically caused the issue, but keep sharing examples as you find them so we can continue researching.
hi @Anastasia, I’m not 100% clear on the curious-ness of the points you made. In the first point you mentioned the time period between the original and the zombie. They have always happened after a certain period of time, which is not consistent.
I’m not sure how this timeframe relates to your second point. Which is more about where the zombie is triggered from? I did mention back in October "Each one was triggered by the same submitter making a new encounter, whic[h] has triggered this zombie of their last encounter. "
In the examples I remember, they appeared a day apart. I was only confirming that the data shows that these duplicates are coming in as unique submissions and not arbitrarily generated.
Ah, no. Maybe some of the examples were but usually they are quite some time period apart (because most people don’t submit encounters every couple of days). But thanks, now I better understand the point you were making!
1 Like