Hi, I have a 3 DC multisite setup. The replication is directional like HKG->SGP->US so the bucket is replicated from HKG to SGP and the same bucket is replicated further from SGP to US. The HKG > SGP connection is pretty fast 12.5millions objects (600GB) transferred in 6.5 hours. Once the OSD crashed in SGP, it stopped the complete chain replication and made PG_DAMAGED cluster error. The pg can be repaired but the sync never started back only bucket sync disable/enable helped. I got OSD crash also in HKG BUT in ASH. ASH no any error, the replication speed is 2 millions objects in 6.5 hours which is like 90GB of data. This is the crash of the osd: { "backtrace": [ "(()+0x12b20) [0x7f597d3fbb20]", "(gsignal()+0x10f) [0x7f597c0667ff]", "(abort()+0x127) [0x7f597c050c35]", "(()+0x9009b) [0x7f597ca1c09b]", "(()+0x9653c) [0x7f597ca2253c]", "(()+0x96597) [0x7f597ca22597]", "(()+0x967f8) [0x7f597ca227f8]", "(()+0x19d24) [0x7f597e168d24]", "(tcmalloc::allocate_full_cpp_throw_oom(unsigned long)+0x146) [0x7f597e18b0d6]", "(rocksdb::Arena::AllocateNewBlock(unsigned long)+0x43) [0x5632f08ccb93]", "(rocksdb::Arena::AllocateFallback(unsigned long, bool)+0x4b) [0x5632f08ccc3b]", "(rocksdb::ConcurrentArena::AllocateAligned(unsigned long, unsigned long, rocksdb::Logger*)+0xb4) [0x5632f07fae94]", "(()+0x1103e7e) [0x5632f085ae7e]", "(rocksdb::MemTable::Add(unsigned long, rocksdb::ValueType, rocksdb::Slice const&, rocksdb::Slice const&, bool, rocksdb::MemTablePostProcessInfo*)+0xcf) [0x5632f07f6f8f]", "(rocksdb::MemTableInserter::PutCFImpl(unsigned int, rocksdb::Slice const&, rocksdb::Slice const&, rocksdb::ValueType)+0x452) [0x5632f08520e2]", "(rocksdb::MemTableInserter::PutCF(unsigned int, rocksdb::Slice const&, rocksdb::Slice const&)+0x17) [0x5632f0852e97]", "(rocksdb::WriteBatch::Iterate(rocksdb::WriteBatch::Handler*) const+0x480) [0x5632f084ac20]", "(rocksdb::WriteBatchInternal::InsertInto(rocksdb::WriteThread::WriteGroup&, unsigned long, rocksdb::ColumnFamilyMemTables*, rocksdb::FlushScheduler*, bool, unsigned long, rocksdb::DB*, bool, bool, bool)+0x149) [0x5632f084ebe9]", "(rocksdb::DBImpl::WriteImpl(rocksdb::WriteOptions const&, rocksdb::WriteBatch*, rocksdb::WriteCallback*, unsigned long*, unsigned long, bool, unsigned long*, unsigned long, rocksdb::PreReleaseCallback*)+0x1acd) [0x5632f078a03d]", "(rocksdb::DBImpl::Write(rocksdb::WriteOptions const&, rocksdb::WriteBatch*)+0x21) [0x5632f078ac11]", "(RocksDBStore::submit_common(rocksdb::WriteOptions&, std::shared_ptr<KeyValueDB::TransactionImpl>)+0x8c) [0x5632f074180c]", "(RocksDBStore::submit_transaction(std::shared_ptr<KeyValueDB::TransactionImpl>)+0x87) [0x5632f0742027]", "(BlueStore::_txc_apply_kv(BlueStore::TransContext*, bool)+0x426) [0x5632f0226376]", "(BlueStore::_kv_sync_thread()+0x176f) [0x5632f024bc1f]", "(BlueStore::KVSyncThread::entry()+0x11) [0x5632f0273791]", "(()+0x814a) [0x7f597d3f114a]", "(clone()+0x43) [0x7f597c12bf23]" ], "ceph_version": "15.2.9", "crash_id": "2021-03-04T14:55:45.094048Z_3d481fd3-7573-4cb7-9b22-20784b418e64", "entity_name": "osd.5", "os_id": "centos", "os_name": "CentOS Linux", "os_version": "8", "os_version_id": "8", "process_name": "ceph-osd", "stack_sig": "9643c370a20c0d34f5e8965ae4461e2a7cf709ab4183929239bc263d0e1eef94", "timestamp": "2021-03-04T14:55:45.094048Z", "utsname_hostname": "hostname", "utsname_machine": "x86_64", "utsname_release": "4.18.0-240.10.1.el8_3.x86_64", "utsname_sysname": "Linux", "utsname_version": "#1 SMP Mon Jan 18 17:05:51 UTC 2021" } Any idea what I should tune? Thank you. ________________________________ This message is confidential and is for the sole use of the intended recipient(s). It may also be privileged or otherwise protected by copyright or other legal rules. If you have received it by mistake please let us know by reply email and delete it from your system. It is prohibited to copy this message or disclose its content to anyone. Any confidentiality or privilege is not waived or lost by any mistaken delivery or unauthorized disclosure of the message. All messages sent to and from Agoda may be monitored to ensure compliance with company policies, to protect the company's interests and to remove potential malware. Electronic messages may be intercepted, amended, lost or deleted, or contain viruses.