Support persistence of configuration items modified at runtime via HTTP API.
```
FE:
GET /api/_set_config?key=value&persist=true
BE
POST /api/update_config?key=value&persist=true
```
The modified config will be saved in `fe_custom.conf` or `be_custom.conf`.
And when process starts, it will load `fe.conf/be.conf` first, then `fe_custom.conf/be_custom.conf`.
Thread [Connect-Scheduler-Check-Timer] fail to iterate the connectionMap , because other threads will modify the connectionMap at the same time.
Fe's connection which is timed out can't be released because of thread [Connect-Scheduler-Check-Timer] not working.
When selecting candidate rowsets to do the cumulative compaction,
some rowsets may not be selected because the protection time has not expired.
Therefore, we need to find the current longest continuous version path in the candidate rowsets.
We create some objects which are only used for unit tests, it's not necessary,
and it may cause create duplicate instances for some classes.
This patch remove unnecessary instance of class BlockManager and StoragePageCache.
If the materialized slot is empty in inline view, the scanner will throw exception 'no materialized slot!'.
When the outer query does not need any real columns of the inner query at all, such as the outer query only contains constant calculations,
all slots in the inner query will not be materialized.
For example:
```
select c1 from (select 'xx' c1, k1 from test) t1;
```
But this does not mean that the inner query (select 'xx' c1, k1 from test) does not need to be executed.
Because the number of rows in the outer result needs to be determined by the number of rows in the inner query.
In this case, the inner query scan node needs to add a column to ensure the correct result.
At the same time it can avoid the scan node reporting errors.
At present, when some rpc errors occur, the client cannot obtain the error information well.
And this CL change the RPC error returned to client like this:
```
ERROR 1064 (HY000): errCode = 2, detailMessage = there is no scanNode Backend. [10002: in black list(A error
occurred: errorCode=2001 errorMessage:Channel inactive error!)]
ERROR 1064 (HY000): failed to send brpc batch, error=The server is overcrowded, error_text=[E1011]The server is
overcrowded @xx.xx.xx.xx:8060 [R1][E1011]The server is overcrowded @xx.xx.xx.xx:8060 [R2][E1011]The server is
overcrowded @xx.xx.xx.xx:8060 [R3][E1011]The server is overcrowded @xx.xx.xx.xx:8060, client: yy.yy.yy.yy
```
This is the last PR of proposal #4308
1. Add a new FE config `enable_http_server_v2` to enable new HTTP Server implementation. The default value is false.
2. Add a new FE config `http_api_extra_base_path` so that we can set base path for Frontend UI.
3. Refactor the new HTTP API response body. The return http status code is always 200, and using internal code in response body to indicate the certain error.
1. Support limit clause push down both odbc table and mysql table.
2. Code refactor of ODBC Scan Node, change `build_connect_string` and `query_string` from BE to FE to make it easily to modify
Currently, there are M threads to do base compaction and N threads to do cumulative compaction for each disk.
Too many compaction tasks may run out of memory, so the max concurrency of running compaction tasks
is limited by semaphore.
If the running threads cost too much memory, we can't defense it. In addition, reducing concurrency to avoid OOM
will lead to some compaction tasks can't be executed in time and we may encounter more heavy compaction.
Therefore, concurrency limitation is not enough.
The strategy proposed in #3624 may be effective to solve the OOM.
A CompactionPermitLimiter is used for compaction limitation, and use single-producer/multi-consumer model.
Producer will try to generate compaction tasks and acquire `permits` for each task.
The compaction task which can hold `permits` will be executed in thread pool and each finished task will
release its `permits`.
`permits` should be applied for before a compaction task can execute. When the sum of `permits`
held by executing compaction tasks reaches a threshold, subsequent compaction task will be no longer allowed,
until some `permits` are released. Tablet compaction score is used as `permits` of compaction task here.
To some extent, memory consumption can be limited by setting appropriate `permits` threshold.
Support new synatx CREATE TABLE [IF NOT EXISTS] [db_name].table_name AS [db_name2].table_name2;
to create a new table from existed table with same table schema.
ISSUE: #4355
