[
https://issues.apache.org/jira/browse/HADOOP-18905?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ConfX updated HADOOP-18905:
---------------------------
Description:
Graceful fence timeout of FailoverController in ZKFailovercontroller equals to
`ha.failover-controller.graceful-fence.rpc-timeout.ms` * 2. Since users are
unaware of this calculation, it thus has risks of overflowing to a negative
number if users set `ha.failover-controller.graceful-fence.rpc-timeout.ms` to a
large value.
To reproduce:
1. set `ha.failover-controller.graceful-fence.rpc-timeout.ms` to 1092752431
2. run `mvn surefire:test
-Dtest=org.apache.hadoop.ha.TestZKFailoverController#testGracefulFailoverFailBecomingStandby`
We create a PR that provides a fix by checking the timeout after multiplication
is at least 0.
was:
Graceful fence timeout of FailoverController in ZKFailovercontroller equals to
`ha.failover-controller.graceful-fence.rpc-timeout.ms` * 2. Since users are
unaware of this calculation, it thus has risks of overflowing to a negative
number if users set `ha.failover-controller.graceful-fence.rpc-timeout.ms` to a
large value.
To reproduce:
1. set `ha.failover-controller.graceful-fence.rpc-timeout.ms` to 1092752431
2. run `mvn surefire:test
-Dtest=org.apache.hadoop.ha.TestZKFailoverController#testGracefulFailoverFailBecomingStandby`
This PR provides a fix by checking the timeout after multiplication is at least
0.
> Negative timeout in ZKFailovercontroller due to overflow
> --------------------------------------------------------
>
> Key: HADOOP-18905
> URL: https://issues.apache.org/jira/browse/HADOOP-18905
> Project: Hadoop Common
> Issue Type: Bug
> Affects Versions: 3.3.6
> Reporter: ConfX
> Priority: Major
> Labels: pull-request-available
>
> Graceful fence timeout of FailoverController in ZKFailovercontroller equals
> to `ha.failover-controller.graceful-fence.rpc-timeout.ms` * 2. Since users
> are unaware of this calculation, it thus has risks of overflowing to a
> negative number if users set
> `ha.failover-controller.graceful-fence.rpc-timeout.ms` to a large value.
>
> To reproduce:
> 1. set `ha.failover-controller.graceful-fence.rpc-timeout.ms` to 1092752431
> 2. run `mvn surefire:test
> -Dtest=org.apache.hadoop.ha.TestZKFailoverController#testGracefulFailoverFailBecomingStandby`
>
> We create a PR that provides a fix by checking the timeout after
> multiplication is at least 0.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
