#monove-server-driven-terminal

1 messages ยท Page 1 of 1 (latest)

finite grove
#

Hello ๐Ÿ‘‹
so you didn't receive any webhook events for success or failure?

stray shuttle
#

that's problem #2. Problem #1 is that the API call received a failure as a response but then proceeded to collect payment

finite grove
#

Okay, interesting.. looking!

#

are you able to reproduce this behavior again? or was it just one time?

stray shuttle
#

just once. but it is significant because we are considering switching our 180+ terminals for new ones and using the server-driven flow so every bug is significant. We are the host platform for multiple Connected stores

#

We need to know whether this bug can be reproduced/identified by your team

#

we are switching terminals precisely due to connectivity issues and can't make that switch without the confidence that the new flow is more stable

finite grove
#

I'm not aware of any built in retries which could explain why the reader prompted for the amount
Taking a deeper look

stray shuttle
#

the double beep was unusual as it usually only beeps once

#

my hunch/concern is that our API's single request triggered 2 requests simultaneously. the first failed which would explain the API call failure and the second was successful.

#

our biggest concern is that our API had no idea at all that there had been a successful payment capture. this is worse than any client-side connectivity issues we're currently experiencing.

finite grove
#

Can you share your terminal process payment code snippet?
I see that the process_payment_intent call was made around 6:37:16 pm and an attempt to confirm the payment intent was made around 6:37:23 pm
It would be unusual for the confirmation request to trigger payment method collection

stray shuttle
#

Our API only made the failed API call: req_KAi7ciX55XFveK
I am almost 100% positive that the reader is what made the successful req_6B52JfxJugN2ep call

finite grove
#

yes, I'm almost certain as well but just want to rule out a few things.

stray shuttle
#

look at the request body, our API doesnt have that information

#

we make a single call to process_payment_intent and do nothing until we hear back from a webhook

#
  1. we post to payment_intents requesting an intent
  2. we then post the payment intent to process_payment_intent
stray shuttle
#

hi, just checking we're still connected

finite grove
#

yup still here
having to dig into reader logs ๐Ÿ™‚

stray shuttle
#

when we're done. i have another bug. should i report it in the main channel so you can open a new thread

#

perhaps you can look at the reader now as it is stuck on processing

finite grove
#

You can post it here

stray shuttle
#

still stuck like this

#

was testing all the error codes. this: req_3ozMssgNxTOQKb

#

online_or_offline_pin_required is actually a complicated error from our experience. what the client-side library does is to re-request the customer put their card in and does a retry of some sorts.

#

i'm wondering if this was coded properly for the server-driven integration.
i say this because it is interesting that every server-driven request to the reader results in a single success or failure with no retries.

finite grove
#

Yeah I think this could definitely be a bug. I came across a similar case earlier today with .02 and .03.

stray shuttle
#

would it be helpful for you if i leave the terminal in this state while you debug it?

#

we're very eager to help Stripe debug terminal issues

finite grove
#

I appreciate it ๐Ÿ™‚ this may need a deeper investigation though I believe we should be able to reproduce this

working on a colleague about the reader showing the payment method collection prompt first as that shouldn't be happening

#

and we can't easily reproduce that either

stray shuttle
#

i think the key was that double beep

#

which sounded like a double request from your servers

#

the second may have cancelled the first request making your server think the reader was offline

finite grove
#

Yeah so far It feels like we sent a message to the reader, couldn't get a reply so we errored but the reader got our message. Could definitely be an edge case scenario.

Quick question, where is your terminal registered exactly? What country?

stray shuttle
#

634 Northumberland Road
Teaneck, NJ 07666

#

usa

finite grove
#

okay, thanks! Okay seems like we'll need to get some more eyes on the first issue as it seems like a bug (a bad one for sure). We'll need to move this to emails.

Can you write in to our support team via support@stripe.com and provide the same details as you did earlier? Make sure to mention that you spoke with Hanzo on discord, I'll grab the ticket and move it to our queue.

stray shuttle
#

ok thanks

finite grove
#

Let me know once you write in so that I can grab the ticket

stray shuttle
#

i wont mention the endless Processing spinner. ok?

finite grove
#

You can/should.

#

I can use it to bump the issue to our team that handles reader softwares

stray shuttle
#

thank you!

finite grove
#

Its the same reader (WSC513142031615) with spinner issue right?

stray shuttle
#

yes

finite grove
#

gotcha. Can you share the PaymentIntent IDs for these test payments as well?
I can include it with my nudge

stray shuttle
#

ok

finite grove
#

Appreciate it ๐Ÿ™‚ lmk once you write in

stray shuttle
#

sent

finite grove
#

Thanks. looking

#

Got it.

#

I'll let you know once I have an update! Thanks again for your patience