be异常掉线

【详述】be异常掉线,
【是否存算分离】否
【StarRocks版本】例如:2.5.19
【集群规模】例如:3fe+3be(fe与be混部)
【联系方式】1186127380@qq.com
【附件】
beINFO:
W0325 01:44:22.605867 30408 storage_engine.cpp:1052] traverse_rowset_meta and remove 312/17952 invalid rowset metas, path:/opt/StarRocks/be/storage duration:977ms
W0325 01:47:42.485762 30408 mem_hook.cpp:257] large memory alloc, query_id:00000000-0000-0000-0000-000000000000 instance: 00000000-0000-0000-0000-000000000000 acquire:1473738555 bytes, stack:
@ 0x4af97cf malloc
@ 0x80ba1b5 operator new()
@ 0x813270a std::__cxx11::basic_string<>::_M_mutate()
@ 0x590a2f5 rocksdb::WriteBatchInternal::DeleteRange()
@ 0x590a482 ZN7rocksdb10WriteBatch11DeleteRangeEPNS_18ColumnFamilyHandleERKNS_5SliceES5.localalias
@ 0x42e138c starrocks::TabletMetaManager::save()
@ 0x42d1b5a starrocks::TabletMeta::_save_meta()
@ 0x42d2106 starrocks::TabletMeta::save_meta()
@ 0x42a41eb starrocks::Tablet::save_meta()
@ 0x430c26a starrocks::TabletUpdates::remove_expired_versions()
@ 0x42c4f50 starrocks::TabletManager::start_trash_sweep()
@ 0x427abf4 starrocks::StorageEngine::_start_trash_sweep()
@ 0x44f4e70 starrocks::StorageEngine::_garbage_sweeper_thread_callback()
@ 0x8134fa0 execute_native_thread_routine
@ 0x7f664b821ea5 start_thread
@ 0x7f664ae3cb0d __clone
@ (nil) (unknown)
W0325 01:47:46.171340 30408 storage_engine.cpp:1032] tablet not found, remove rowset meta, rowset_id=02000000006b84f6284427cef55a2687f7a9b296a978ccb8 tablet: 301406470

be.out

dmesg -T 看下,是不是OOM了

有oom

OOM的信息发下

cat /proc/sys/vm/overcommit_memory 看下这个是多少

oom0325.log (292.7 KB)

[root@bogon log]# cat /proc/sys/vm/overcommit_memory
0

这个改成1, 文档里有明确说明

感谢,我再对比改下 检查环境配置 | StarRocks

系统改完生效之后,需要重启be吗?

不需要