[
https://issues.apache.org/jira/browse/HBASE-29734?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jianfei Cheng updated HBASE-29734:
----------------------------------
Description:
In MoveWithAck, [the maxWait timeout was calculated inside the retry
loop|https://github.com/apache/hbase/blob/master/hbase-server/src/main/java/org/apache/hadoop/hbase/util/MoveWithAck.java#L83-L84]
using a fixed startTime calculated at the out side of the retry loop, causing
it to remain constant across all retry attempts. This means that each retry
didn't really check isSameServer and marks all retries failed.
To fix it, calculate maxWait based on the current time instead of sartTime
was:
In MoveWithAck, [the maxWait timeout was calculated inside the retry
loop|https://github.com/apache/hbase/blob/master/hbase-server/src/main/java/org/apache/hadoop/hbase/util/MoveWithAck.java#L83-L84]
using a fixed startTime calculated at the out side of the retry loop, causing
it to remain constant across all retry attempts. This meant that each retry
didn't really check isSameServer and marks all retries failed.
To fix it, calculate maxWait based on the current time instead of sartTime
> MoveWithAck calculates the maxWait timeout wrong.
> -------------------------------------------------
>
> Key: HBASE-29734
> URL: https://issues.apache.org/jira/browse/HBASE-29734
> Project: HBase
> Issue Type: Bug
> Components: util
> Reporter: Jianfei Cheng
> Priority: Major
>
> In MoveWithAck, [the maxWait timeout was calculated inside the retry
> loop|https://github.com/apache/hbase/blob/master/hbase-server/src/main/java/org/apache/hadoop/hbase/util/MoveWithAck.java#L83-L84]
> using a fixed startTime calculated at the out side of the retry loop,
> causing it to remain constant across all retry attempts. This means that each
> retry didn't really check isSameServer and marks all retries failed.
> To fix it, calculate maxWait based on the current time instead of sartTime
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
