Alexandr Yankov
11/21/2025, 8:31 PMweekly chart, sharp drops are a restartKevin Cai
11/22/2025, 10:38 AMKevin Cai
11/23/2025, 8:52 AMstarrocks_be_jemalloc_allocated_bytes and starrocks_be_jemalloc_resident_bytesAlexandr Yankov
11/23/2025, 8:55 AMSo these metrics are not variables for configuration?Kevin Cai
11/23/2025, 9:00 AMhttp://<CN>:8040/metricsAlexandr Yankov
11/23/2025, 9:03 AMstarrocks_be_jemalloc_allocated_bytes 1541884120
starrocks_be_jemalloc_resident_bytes 1647411200Alexandr Yankov
11/23/2025, 9:04 AMI scaled the FE down to 3 pods and noticed that this only happens on the leader.Alexandr Yankov
11/23/2025, 9:06 AMWhy CN? If the memory is leaking on FE.Kevin Cai
11/23/2025, 9:09 AMKevin Cai
11/23/2025, 9:10 AMAlexandr Yankov
11/23/2025, 9:11 AMAlexandr Yankov
11/23/2025, 9:11 AMAlexandr Yankov
11/23/2025, 9:12 AMAlexandr Yankov
11/23/2025, 9:12 AMKevin Cai
11/23/2025, 9:13 AMKevin Cai
11/23/2025, 9:13 AMAlexandr Yankov
11/23/2025, 9:16 AMOkay, then I experimented, scaled FE to 3 pods to get a quorum and reduced the memory to 4 GB per pod. I actually changed JAVA_OPT to "-Dlog4j2.formatMsgNoLookups=true -Xmx3G -Xms1G -XX:+UseShenandoahGC -Xlog:gc*:${LOG_DIR}/fe.gc.log.$DATE:time" The result is the same.Alexandr Yankov
11/23/2025, 9:16 AMThere is even a file with profiling and a graph)Alexandr Yankov
11/23/2025, 1:25 PMt was solved in the following way: 2 Fepods and 4 CN. After that, the memory leak stopped. We need to somehow better understand how to build a quorum in Starrocks.Alexandr Yankov
11/23/2025, 1:25 PM