quarkus-cache might block indefinitely when backed by Redis and lockTimeout=0

[lxngr]

Describe the bug

When using qarkus-cache backed by quarkus-redis-cache and lockTimeout=0 the cache might block indefinitely. This is the case when a cache miss uses a "heavy weight" operation that itself uses a cache and that operation misses too (two chained cache misses).

Might not a be the best idea but works well when using the default cache implementation (Caffeine).

Expected behavior

The second "heavy weight" operation should be executed anyway and the cache should not block (like when using the default cache implementation).

Actual behavior

Cache blocks indefinitely.

How to Reproduce?

I've build an reproducer based on the cache example (APPLICATION DATA CACHING) that uses two caches chained and both miss: https://github.com/lxngr/quarkus-cache.

Test with: "curl http://localhost:8080/weather\?city\=LA"

Output of uname -a or ver

Darwin Kernel Version 23.4.0 (macOS)

Output of java -version

openjdk version "17.0.10" 2024-01-16 LTS

Quarkus version or git rev

3.8.2

Build tool (ie. output of mvnw --version or gradlew --version)

Apache Maven 3.9.2

Additional information

No response

[quarkus-bot]

/cc @cescoffier (redis), @gsmet (redis), @gwenneg (cache), @machi1990 (redis)

[geoand]

Interesting scenario.

You mention lockTimeout=0 in the description, but I don't see that configured anywhere in the sample application you have attached. Can you clarify?

[lxngr]

I've left it out in my sample application as lockTimeout = 0 is the default.

[geoand]

Ah, you're right.

I am not sure what we can really do about this one as it's a weird case...