Jaromir Hamala

@mghildiy go ahead. what's your idea of a sensible behaviour?

@mghildiy it's complicated. this is about offheap/native memory exhausted. the problem is when memory is exhausted we cannot allocate anymore. but it appears that we require memory allocation to send...

hi @1-Arijit-choudhury, @adityagupta0251, it's best to join https://slack.questdb.com/ and ask question about the impl. or perhaps sync between each other - it's the best to join forced.

it turns out the `OperationExecutor` in master uses 10 recompilation attempts at most. so perhaps that's good enough for real-life and Azure CI runners, but not for Hetzner. Addressing this...

re-opening as that's happening repeatadly on Hetzner. even after increasing recompilation attempts. it suspends table so that's not great.

when this PR is merged then it should be also tested with jemalloc. see: https://github.com/questdb/questdb/pull/6477#discussion_r2577453557 this does not say we must support jemalloc+musl. this combination makes little sense to me...

a reproducer: https://github.com/jerrinot/jni-rs_repro The code is copied from a reproducer @RaphDal created for https://github.com/apache/opendal/issues/6869, but the OpenDAL library is removed.

hi @rib, many thanks for picking this up, very much appreciated! I also thought `FlsAttach()` would be the way to go, assuming you do not want to change programming model...

@coderabbitai dear :rabbit: , please review this little piggie.

@coderabbitai I, for one, welcome our AI overlords, as long as it means I am free from messing with build scripts. Review this change please, some context for you: https://maven.apache.org/surefire/maven-surefire-plugin/examples/shutdown.html