Hello everyone! I’m having some issues with the gi...
# announcementsa
Alberto Costa
02/08/2021, 3:39 PMHello everyone! I’m having some issues with the github connector, it seems to be looping and printing the same error strings over and over again… has this appened to anyone else before? Also, is there a nice way to kill a running synchronization from the interface (or from somewhere else)?
c
Chris (deprecated profile)
02/08/2021, 3:41 PM
Cancelling a sync is not implemented yet... So i usually go for
docker kill <source/destination container for sync>✅ 1
a
Alberto Costa
02/08/2021, 3:46 PMah sounds good! will this become part of the interface eventually?
Alberto Costa
02/08/2021, 4:12 PMcause currently you have to kill the container 5 times am i right? it keeps respawning like doom enemies 😂
c
Chris (deprecated profile)
02/08/2021, 4:14 PM
haha, yes it’s the retry of the scheduler that does this… it’s being currently reworked though so you can expect some changes on this soon i guess
s
s
02/08/2021, 4:20 PM
@Alberto Costa is there an issue with the connector/is data not populating? Fwiw the underlying singer tap this connector is based on is very chatty in the logs
a
Alberto Costa
02/08/2021, 4:22 PMSome data is actually being synced, but it looks like the tap fails at some point, which triggers the normalization process. Once normalization is complete, a new Attempt starts
Alberto Costa
02/08/2021, 4:22 PMThe thing is that the new attempt starts from the beginning, wasting API quota
Alberto Costa
02/08/2021, 4:23 PMthe official singer tap is not very nicely implemented, is airbyte using that one wrapped in a docker container?
s
s
02/08/2021, 4:26 PM
Yes - this was one of the earlier taps we implemented before deciding to stay away from singer taps unless they were super high quality. We have it on the roadmap to implement an Airbyte native one but haven’t gotten to it yet.
s
02/08/2021, 4:27 PM
Just so I understand correctly: the sync fails before normalization but normalization is still triggered?
a
Alberto Costa
02/08/2021, 4:27 PMyes
Alberto Costa
02/08/2021, 4:27 PMthis is true also when you kill the container
Alberto Costa
02/08/2021, 4:28 PMthere’s a log that says
[32mINFO[m i.a.i.b.FailureTrackingConsumer(close):64 - {} - hasFailed: falseAlberto Costa
02/08/2021, 4:28 PMand then normalization triggers
s
s
02/08/2021, 4:29 PM
Has failed is false though? It’s a hard way of saying it succeeded 🙂
s
02/08/2021, 4:29 PM
Could you share the logs from the job so we can verify?
a
Alberto Costa
02/08/2021, 4:30 PMthat log was printed when i killed the container
Alberto Costa
02/08/2021, 4:30 PMand also when failing to sync before
Alberto Costa
02/08/2021, 4:31 PMso i think it misunderstand the concept of failure 😂
Alberto Costa
02/08/2021, 4:32 PMi can’t really share logs unfortunately since they contain all sort of data from our github 😕
Alberto Costa
02/08/2021, 4:32 PMi will see if i can clean it
s
s
02/08/2021, 4:48 PM
At least the very last bit of the log should have a stack trace or failure reason