xtdb 2021-09-04 | Slack Archive

This is probably known, but starting a local crux node on java 16 ends the JVM with SIGABRT

Yep. There are some workarounds, but can't guarantee any will resolve for your situation: https://github.com/juxt/crux/issues/894

emccue14:09:34

Skimming that thread - seems like it bottoms out at JNI wierdness

emccue14:09:05

which is unfortunate

tatut04:09:56

fwiw, these env variables worked for me

;; CRUX_DISABLE_LIBCRYPTO=true
 ;; CRUX_ENABLE_BYTEUTILS_SHA1=true
 ;; to prevent crash "java is loading libcrypto in an unsafe way"

emccue22:09:22

this is with and without the add-opens for the concurrent map sins

emccue22:09:37

(c/start-node {})
2021-09-04 18:39:25,633 [nRepl-session-a0bb1109-6c2a-44f1-8841-1078801856ae] WARN  crux.cache - Could not open ConcurrentHashMap.table field - falling back to LRU caching. Use `--add-opens java.base/java.util.concurrent=ALL-UNNAMED` to use the second-chance cache. 

Process finished with exit code 134 (interrupted by signal 6: SIGABRT)

👍 2

2021-09-04

Channels