This CL mainly changes:
1. Reducing the rpc timeout problem caused by rpc waiting for the worker thread of brpc.
1. Merge multiple fragment instances on the same BE to send requests to reduce the number of send fragment rpcs
2. If fragments size >= 3, use 2 phase RPC: one is to send all fragments, two is to start these fragments. So that there
will be at most 2 RPC for each query on one BE.
3. Set the timeout of send fragment rpc to the query timeout to ensure the consistency of users' expectation of query timeout period.
4. Do not close the connection anymore when rpc timeout occurs.
5. Change some log level from info to debug to simplify the fe.log content.
NOTICE:
1. Change the definition of execPlanFragment rpc, must first upgrade BE.
3. Remove FE config `remote_fragment_exec_timeout_ms`
As in issue, the combination and schema change at the same time may lead to version intersection.
Describe the overview of changes.
1. Do not do compaction before schema change is actually executed.
2. Set tablet as bad when it has version intersection.
3. Do not do schema change when it can not find appropriate versions to delete in new tablet.
4. Do not change rowsets after compaction if the rowsets of the tablet has changed.
* handle ColumnDictory in evaluate_or
We need to handle ComlumnDictory in evaluate_or, otherwise delete handler
would triger a core dump.
* handle ComlumnDictionary in evaluate_and
Because there is only one difference between evaluate_and and
evaluate_or, that is or and delete, I merge two macros into one.
Delete handlers also trigger evaluate_and, i am not sure if column
dictionary would be used in evaluate_and.
* clang format
* fix short circut for evaluate_and and evaluate_or
* clang format
ZSTD compression is fast with high compression ratio. It can be used to archive higher compression ratio
than default Lz4f codec for storing cost sensitive data such as logs.
Compared to Lz4f codec, we see zstd codec get 35% compressed size off, 30% faster at first time read without OS page
cache, 40% slower at second time read with OS page cache in the following comparison test.
test data: 25GB text log, 110 million rows
test table: test_table(ts varchar(30), log string)
test SQL: set enable_vectorized_engine=1; select sum(length(log)) from test_table
be.conf: disable_storage_page_cache = true
set this config to disable doris page cache to avoid all data cached in memory for test real decompression speed.
test result
master branch with lz4f codec result:
- compressed size 4.3G
- SQL first exec time(read data from disk + decompress + little computation) : 18.3s
- SQL second exec time(read data from OS pagecache + decompress + little computation) : 2.4s
this branch with zstd codec (hardcode enable it) result:
- compressed size: 2.8G
- SQL first exec time: 12.8s
- SQL second exec time: 3.4s
1. Fix Lru Cache MemTracker consumption value is negative.
2. Fix compaction Cache MemTracker has no track.
3. Add USE_MEM_TRACKER compile option.
4. Make sure the malloc/free hook is not stopped at any time.
* [Enhancement][Vectorized]build hash table with new thread, as non-vectorized past do
edit after comments
* format code with clang format
Co-authored-by: lidongyang <dongyang.li@rateup.com.cn>
Co-authored-by: stephen <hello-stephen@qq.com>
