Invalidate queries not working with retry

[kangelSL]

Hi,

I have a setup using the queryClient's defaultOptions.

The queries are set to retry and the mutations are expected to retry on a specific error.

This works in most cases as expected. However, there are cases where a useMutation call is followed by using invalidateQueries on a useQuery get request in order to trigger a refetch. This happens in both cases of the useMutation succeeding or failing.

My issue is, the useQuery call is not retrying on failures in this case when it should be. See network log below:

Where 1 is the failing mutation call (doesn't retry as expected)
And 2 is the failing get call triggered by the invalidate queries (doesn't retry, but should be)

I'm struggling to see anything at all in the documentation about this. Is invalidateQueries not designed to fire retries if it fails? How do I get it to apply the retry it does elsewhere, when the useQuery call is fired normally?

[TkDodo]

yeah it should retry. I would need to see a minimal reproduction please

[kangelSL]

export const queryClient = new QueryClient({ defaultOptions: { queries: { staleTime: 5 * 1000, retry: 3, throwOnError: true, refetchOnMount: false, refetchOnWindowFocus: false, }, mutations: { retry: () => functionToCheckErrorCode, }, }, })

And the invalidate on error/success is as follows:

queryClient.invalidateQueries({ queryKey: ["getKey"] })

[kangelSL]

@TkDodo If there's anything in the docs you can point me to, that would be useful also! There didn't seem to be any mention if retries work any differently on the invalidate queries firing

[TkDodo]

from what I can see this just works 🤷 . You have to create a stackblitz or codesandbox to show that it really doesn’t...

[kangelSL]

@TkDodo After further investigation, I've found that the POST request that doesn't retry beforehand is calling an error boundary. If I don't call the error boundary for the POST request, the invalidate queries correctly retries for the following GET request.

Alternatively, if I use resetQueries instead of invalidateQueries the retry works correctly.

Am I right in thinking the error boundary could be causing some kind of unmounting issue?

[kangelSL]

For anyone who runs in to anything similar - I found a way around my issue by using resetQueries on the invalidated queries