-
Notifications
You must be signed in to change notification settings - Fork 3.5k
[opt](cloud) Support cached cloud partition version for high frequency query #50439
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[opt](cloud) Support cached cloud partition version for high frequency query #50439
Conversation
Thank you for your contribution to Apache Doris. Please clearly describe your PR:
|
run buildall |
run buildall |
afed61d
to
eda01ce
Compare
run buildall |
TPC-H: Total hot run time: 34544 ms
|
TPC-DS: Total hot run time: 193061 ms
|
ClickBench: Total hot run time: 29.86 s
|
run buildall |
ab071d3
to
d025c52
Compare
run buildall |
TPC-H: Total hot run time: 34065 ms
|
TPC-DS: Total hot run time: 192107 ms
|
ClickBench: Total hot run time: 29.2 s
|
PR approved by anyone and no changes requested. |
run buildall |
TPC-H: Total hot run time: 33852 ms
|
TPC-DS: Total hot run time: 190995 ms
|
ClickBench: Total hot run time: 29.5 s
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
PR approved by at least one committer and no changes requested. |
…y query (#50439) Support new session variable to get rid of frequent get_version request to meta-service. set `cloud_partition_version_cache_ttl_ms` with a suitable value as needed, it means a partition's version may be cached on FE at most X milliseconds. pro: reduce RPC/IO to meta-service and fdb, increase query throughput and reduce query latency (plan phase) con: visibility may be delayed when doing point query; data consistency (snapshot) may be broken when doing query wi involved multiple partitions and there is continuous data import;
…y query (apache#50439) Support new session variable to get rid of frequent get_version request to meta-service. set `cloud_partition_version_cache_ttl_ms` with a suitable value as needed, it means a partition's version may be cached on FE at most X milliseconds. pro: reduce RPC/IO to meta-service and fdb, increase query throughput and reduce query latency (plan phase) con: visibility may be delayed when doing point query; data consistency (snapshot) may be broken when doing query wi involved multiple partitions and there is continuous data import;
What problem does this PR solve?
Support new session variable to get rid of frequent get_version request to meta-service.
set
cloud_partition_version_cache_ttl_ms
with a suitable value as needed, it means a partition's version may be cached on FE at most X milliseconds.pro: reduce RPC/IO to meta-service and fdb, increase query throughput and reduce query latency (plan phase)
con: visibility may be delayed when doing point query; data consistency (snapshot) may be broken when doing query wi involved multiple partitions and there is continuous data import;
Check List (For Author)
Test
Behavior changed:
Does this need documentation?
Check List (For Reviewer who merge this PR)