Uh oh!
There was an error while loading. Please reload this page.
[fix](cooldown) Fix bug in follow cooldowned data - #32798
Conversation
doris-robot
commented
Mar 25, 2024
Thank you for your contribution to Apache Doris. Since 2024-03-18, the Document has been moved to doris-website. |
PR approved by anyone and no changes requested. |
clang-tidy review says "All clean, LGTM! 👍" |
clang-tidy review says "All clean, LGTM! 👍" |
dataroaring
commented
Mar 25, 2024
run buildall |
PR approved by at least one committer and no changes requested. |
doris-robot
commented
Mar 26, 2024
TeamCity be ut coverage result: |
platoneko
commented
Mar 26, 2024
run buildall |
clang-tidy review says "All clean, LGTM! 👍" |
doris-robot
commented
Mar 26, 2024
TeamCity be ut coverage result: |
doris-robot
commented
Mar 26, 2024
TPC-H: Total hot run time: 37748 ms |
doris-robot
commented
Mar 26, 2024
TPC-DS: Total hot run time: 181761 ms |
doris-robot
commented
Mar 26, 2024
Load test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G' |
Proposed changes
The single replica cooldown protocal in Doris must maintain the invariant that cooldowned rowsets of a replica exactly corresponds to the cooldown meta id, however skipping rowsets with the same version when following cooldowned data would break this invariant. Modifying the code to skip rowsets with the same rowset id can maintain the invariant.
This PR also fix another potential data loss problem.
Further comments
If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...