| It turns out some jdbc operations do not respect thread interrupts (double-check that, austin.blatt and I have some context), and recent important stability/performance improvements (e.g. sync timeouts), were written under the assumption that they do (I'll plan to link some relevant tickets). Assuming interrupts won't always work, rewrite the code to accommodate that. Possibilities include (also) cancelling the statement (evaluate what pgjdbc supports there), setting a query timeout (investigate Statement.setQueryTimeout, or in-line per-transaction set commands to establish a timeout), etc. |