hey, getting 504s when trying to refresh any data ...
# gooddata-cloud
t
hey, getting 504s when trying to refresh any data source, any ideas on how to debug/progress?
j
t
done, waiting for the refresh spinner...................
spinner
nah still 504ing
j
Could you provide the domain you are working with? I will check from my end
t
as in
<https://learnamp.cloud.gooddata.com/>
?
or something else
j
that works... give me a moment
Can you check the datasource on your end. We are only seeing the following error: exc": "java.util.concurrent.CancellationException: Cancellation received from client\n\tat kotlinx.coroutines.ExceptionsKt.CancellationException
it doesn't look like an issue on our end, as we cannot reproduce it in any other org.
Ok, so it looks like when you change the available sources before the initial table list has loaded, you will get that 500 timeout error. You need to wait until the full tables have been loaded and cached and then it should load without issue. Nevertheless, we will look into the performance of this as it is not an ideal error to stumble upon.
t
what do you mean by change the available sources before the initial table list has loaded?
change ie in redshift?
or change within good data or?
j
I mean the tables listed here:
You need to wait until they load completely, before trying to change one.
Are you still seeing an error after you let them all load?
t
they never load bud
that's the error
j
I'm not able to reproduce it
t
in our org?
j
yes, in your org
t
k, I should enter
edit
mode and wait?
j
yes
t
how long was it taking for you when doing repro, because I've waited like 5 mins earlier with no success
j
At first, it took some time for the tables to load, but after it had been cached there were no issues
t
\
so, assuming the above does work, how do I get it to not 504 on the initial load (and subsequent attempts to refresh)?
m
Hi Tomas, sorry for the delay in getting back to you here - My colleague Joe has been away for a few days. So we wanted to followup and see if you’re still running into the same issue now after following my colleagues suggestions?