Alex Behm has posted comments on this change. Change subject: IMPALA-5500: Reduce catalog update topic size ......................................................................
Patch Set 2: (7 comments) http://gerrit.cloudera.org:8080/#/c/7268/2/be/src/catalog/catalog-server.cc File be/src/catalog/catalog-server.cc: Line 330: if (!status.ok()) { take out of if/else since it's common to both cases Line 367: uint8_t* compressed_buffer_ptr = &compressed_buffer[0]; comptessed_buffer.data() Line 370: // Base64 encode the compressed catalog object and store it in the topic item. Is base64-encoding really necessary? Why don't the raw bytes work? http://gerrit.cloudera.org:8080/#/c/7268/2/be/src/catalog/catalog-server.h File be/src/catalog/catalog-server.h: Line 200: /// Serializes 'catalog_object' and compresses it using Snappy compression. remove second "compression" http://gerrit.cloudera.org:8080/#/c/7268/2/be/src/service/impala-server.cc File be/src/service/impala-server.cc: Line 1417: Status ImpalaServer::DecompressAndDeserializeCatalogObject(const TTopicItem& item, Why not separate decompression from deserialization? Seem like separate concerns. Then the code would be something like: buffer = raw buffer if (compaction is on) buffer = decompress(); DeserializeThriftMsg() ... http://gerrit.cloudera.org:8080/#/c/7268/2/be/src/util/compress.h File be/src/util/compress.h: Line 103: SnappyCompressor(MemPool* mem_pool = NULL, bool reuse_buffer = false); Why these changes? Shouldn't these always be created using CreateCompressor()? http://gerrit.cloudera.org:8080/#/c/7268/2/tests/custom_cluster/test_compact_catalog_updates.py File tests/custom_cluster/test_compact_catalog_updates.py: Line 45: "select count(*) from functional.alltypes") Check result? Maybe also use another query on a parquet table. -- To view, visit http://gerrit.cloudera.org:8080/7268 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I2f725cd8596205e6101d5b56abf08125faa30b0a Gerrit-PatchSet: 2 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Dimitris Tsirogiannis <[email protected]> Gerrit-Reviewer: Alex Behm <[email protected]> Gerrit-HasComments: Yes
