Recheck for gateways if there isn't one saved

[benthecarman]

I am getting this currently: https://hermes-blinded-staging.fly.dev/lnurlp/test1234/callback?amount=10000

I assume because we don't durably save the fedimint db that we can lose the gateways we've been told about.

Finding a gateway is just lookups in the fedimint database so this shouldn't block requests.

[benthecarman]

Added a gateways route to try and make debugging this easier

[TonyGiorgio]

we don't durably save the fedimint db

We do currently, so that's not the issue.

[TonyGiorgio]

If this is running for awhile, what stops a current gateway from going down and this doesn't switch to another one?

[benthecarman]

If this is running for awhile, what stops a current gateway from going down and this doesn't switch to another one?

yeah this is why I was talking to justin about this fedimint/fedimint#4722

Maybe should just throw out the cache and lookup every time. Originally i thought the lookup was expensive but it's not

[TonyGiorgio]

If this is running for awhile, what stops a current gateway from going down and this doesn't switch to another one?

yeah this is why I was talking to justin about this fedimint/fedimint#4722

Maybe should just throw out the cache and lookup every time. Originally i thought the lookup was expensive but it's not

If it's a request to the federation every time then no.

[benthecarman]

If it's a request to the federation every time then no.

it looks up in our local db

[TonyGiorgio]

Then that doesn't really solve new/old gateways filtering in or out.

[benthecarman]

Then that doesn't really solve new/old gateways filtering in or out.

my understanding is that these are broadcast to us and saved to our db, let me ask though

[benthecarman]

Okay redid this so we periodically update our gateway db every hour and we always pick from that