[Enhancement] Support use local file to accelerate the broker load

[mchades]

What type of PR is this：

• Enhancement

Which issues of this PR fixes ：

Fixes #

Problem Summary(Required) ：

When there is obvious network latency(e.g. ping command cost time greater than 40ms) between BE and broker file(such as in different idc), broker load ran into slow.

This PR set use_local_cache in broker load option to accelerate it.

LOAD LABEL test_db.label1
(
    xxx
)
WITH BROKER "mybroker"
(
    xxx
)
PROPERTIES
(
    "timeout" = "3600"
    "use_local_cache" = "true"
);

test result:

orc broker load（file count：101，total data size：13GB，table field count:25）

BE&broker file in different idc	use Optimize orc load random I/O	use local cache	elapsed time
NO	NO	NO	7min
NO	YES	NO	6.5min
NO	NO	YES	6.5min
NO	YES	YES	6.5min
YES	NO	NO	30min timeoute
YES	YES	NO	30min timeoute
YES	NO	YES	11min
YES	YES	YES	11min

Checklist:

• I have added test cases for my bug fix or my new feature
• I have added user document for my new feature or new function

[xiaoyong-z]

maybe set config::use_local_filecache_for_broker_random_access_file in broker load's PROPERTIES is a better choice? like the following:

LOAD LABEL test_db.label1
(
    xxx
)
WITH BROKER "mybroker"
(
    xxx
)
PROPERTIES
(
    "timeout" = "3600"
    "use_local_filecache_for_broker_random_access_file" = "true"
);

Then user don't need to restart BE if they want to change use_local_filecache_for_broker_random_access_file

[mchades]

maybe set config::use_local_filecache_for_broker_random_access_file in broker load's PROPERTIES is a better choice? like the following:

LOAD LABEL test_db.label1
(
    xxx
)
WITH BROKER "mybroker"
(
    xxx
)
PROPERTIES
(
    "timeout" = "3600"
    "use_local_filecache_for_broker_random_access_file" = "true"
);

Then user don't need to restart BE if they want to change use_local_filecache_for_broker_random_access_file

ok, but use_local_filecache_for_broker_random_access_file seems like too long to use, I take use_local_cache instead

[mchades]

run starrocks_fe_unittest

[sonarqubecloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
1 Code Smell

0.0% Coverage
0.0% Duplication

[wanpengfei-git]

[FE PR Coverage Check]

😞 fail : 14 / 19 (73.68%)

file detail

	path	covered_line	new_line	coverage	not_covered_line_detail
🔵	com/starrocks/load/loadv2/LoadJob.java	2	4	50.00%	[234, 329]
🔵	com/starrocks/analysis/LoadStmt.java	3	6	50.00%	[265, 266, 267]
🔵	com/starrocks/planner/FileScanNode.java	4	4	100.00%	[]
🔵	com/starrocks/load/loadv2/LoadingTaskPlanner.java	4	4	100.00%	[]
🔵	com/starrocks/load/loadv2/LoadLoadingTask.java	1	1	100.00%	[]

[chaoyli]

If an orc file is split into tasks and has some parallelism, will it download one file for many files?