[client] Get tableInfo via the Admin API instead of via metadata updater#2016
Conversation
|
@swuferhong is this PR ready to review? It seems there are tests are failing. |
1764fac to
688de05
Compare
wuchong
left a comment
There was a problem hiding this comment.
We can put TableInfo into WriteRecord, and then we don't need TableInfoCache.
73a02b6 to
fa9445a
Compare
fa9445a to
f103569
Compare
wuchong
left a comment
There was a problem hiding this comment.
@swuferhong I made a little changes to the PR and addressed my comments. Please take a look.
|
|
||
| @Override | ||
| public int hashCode() { | ||
| return Objects.hash(tableId); |
There was a problem hiding this comment.
The TableKey is super hack, because it is used as a map key, but the TableInfo is not considered for comparison. If we want to carry the TableInfo in the Map, it should belong to the values. However, I think we don't need the TableInfo to check log/kv. I made a change for this.
| @AfterAll | ||
| protected static void afterAll() throws Exception { | ||
| conn.close(); | ||
| } |
There was a problem hiding this comment.
I think we can revert all the test changes.
| int bucketId, | ||
| PhysicalTablePath physicalTablePath, | ||
| int schemaId, | ||
| TableInfo tableInfo, |
There was a problem hiding this comment.
We should avoid passing large context objects; pass only the required data.
Purpose
Linked issue: close #2014
Get
TableInfovia the Admin API instead of via metadata updater. This PR does not aim to fully removeTableInfofrom themetadataUpdaterand theCluster—that work will be done in 483. This PR only changes the way table info is retrieved.Brief change log
Tests
API and Format
Documentation