...
After successfully copied a segment file to remote storage, RLM will append a set of index entries to 3 local index files: remotelogindex, remoteoffsetindex, remotetimeindex. These index files are rotated by RLM at a configurable time interval (or a configurable size).
(active segment) {log.dirs}/{topic-partition}/0000002400013.index {log.dirs}/{topic-partition}/0000002400013.timeindex {log.dirs}/{topic-partition}/0000002400013.log (inactive segments) {log.dirs}/{topic-partition}/0000002000238.index {log.dirs}/{topic-partition}/0000002000238.timeindex {log.dirs}/{topic-partition}/0000002000238.log {log.dirs}/{topic-partition}/0000001600100.index {log.dirs}/{topic-partition}/0000001600100.timeindex {log.dirs}/{topic-partition}/0000001600100.log (active remote segment) {log.dirs}/{topic-partition}/0000001000121.remoteoffsetindex {log.dirs}/{topic-partition}/0000001000121.remotetimeindex {log.dirs}/{topic-partition}/0000001000121.remotelogindex (inactive remote segments) {log.dirs}/{topic-partition}/0000000512002.remoteoffsetindex {log.dirs}/{topic-partition}/0000000512002.remotetimeindex {log.dirs}/{topic-partition}/0000000512002.remotelogindex |
Each index entry of the remotelogindex file contains the information of a sequence of records in the remote log segment file. The format of a remotelogindex entry:
magic: int16 (current magic value is 0) length: int16 (length of this entry) crc: int32 (checksum of this entry) startOffset: int64 (the Kafka offset of the 1st record) lastOffset: int64 (the Kafka offset of the last record) firstTimestamp: int64 maxTimestamp: int64 dataLength: int32 (length of the data) rdiLength: int16 rdi: byte[] (Remote data identifier) |
RDI (Remote data identifier) is the "pointer" or "URI" of the remote data. The format of RDI depends on the implementation. For example, RDI can be HDFS file path and offset, or S3 key and offset. When reading the remote records, RLM will use RDI to retrieve the remote data.
...