Kamil Braun

@aleksbykov @gleb-cloudius isn't this the issue that you discussed on #raft Slack channel? I think you resolved it and declared it's a non-issue?

> I can craete reproducer job with 2 nemesis and try to reproduce issue Yes please

@aleksbykov any luck with the above?

Similar failure: https://github.com/scylladb/scylladb/issues/18436 `abort_requested_exception` causing repair to fail, causing removenode to fail. Also with concurrent index drop! Perhaps the index drop triggers abort somewhere somehow?

Hypothesis: repair gets stuck on Raft read barrier (`start_operation`). The barrier times out and `abort_requested_exception` flies out (should be a different exception probably). Why the barrier times out? Perhaps the...

Actually in c01b19fcb3fa (used by this run) repair doesn't use Raft read barrier, I think. Instead it uses old schema pull code. This changed later in c6c8347493926b0d0dc1ae0d7a63be03296d2348

This may be a reincarnation of https://github.com/scylladb/scylladb/issues/15598 except we're getting abort_requested_exception instead of no_such_column_family for some reason

Possible source of the `abort_requested_exception` could be this `abort_on_expiry`: ```cpp future table_sync_and_check(replica::database& db, service::migration_manager& mm, const table_id& uuid) { if (mm.use_raft()) { abort_on_expiry aoe(lowres_clock::now() + std::chrono::seconds{10}); auto& as = aoe.abort_source();...

https://github.com/scylladb/scylladb/issues/15598#issuecomment-2113036024

Removed 6.0 milestone, it's not critical and most likely a reincarnation of existing issue which existed for more than a year already (https://github.com/scylladb/scylladb/issues/15598#issuecomment-2113036024) -- apparently the fix simply changed one...