[
https://issues.apache.org/jira/browse/CASSGO-52?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
João Reis updated CASSGO-52:
----------------------------
Description:
If a node in `system.peers` is DOWN the `Conn.awaitSchemaAgreement` function
will complain about `gocql: cluster schema versions not consistent: [<uuid1>
<uuid2>]` instead of ignoring the node.
This can happen when a node is replaced but the `system.peers` still shows the
ghost node in there with a different schema version than the rest of the nodes.
Raised [here,|https://github.com/apache/cassandra-gocql-driver/issues/1736]
with a PR open
[here.|https://github.com/apache/cassandra-gocql-driver/pull/1738]
EDIT by [~joaoreis] : the driver should ignore DOWN nodes similar to what the
java driver does (check my comment below) and this will fix the issue with
ghost nodes even if we're not specifically trying to fix the ghost node issue
since the driver can't tell whether a node is a ghost node or not.
was:
If a node in `system.peers` is unreachable the `Conn.awaitSchemaAgreement`
function will complain about `gocql: cluster schema versions not consistent:
[<uuid1> <uuid2>]` instead of ignoring the unreachable node.
This can happen when a node is replaced but the `system.peers` still shows the
ghost node in there with a different schema version than the rest of the nodes.
Raised [here,|https://github.com/apache/cassandra-gocql-driver/issues/1736]
with a PR open
[here.|https://github.com/apache/cassandra-gocql-driver/pull/1738]
> awaitSchemaAgreement not ignoring DOWN node
> -------------------------------------------
>
> Key: CASSGO-52
> URL: https://issues.apache.org/jira/browse/CASSGO-52
> Project: Apache Cassandra Go driver
> Issue Type: Bug
> Components: Core
> Reporter: Felipe Pinheiro
> Priority: Normal
> Fix For: 2.x
>
>
> If a node in `system.peers` is DOWN the `Conn.awaitSchemaAgreement` function
> will complain about `gocql: cluster schema versions not consistent: [<uuid1>
> <uuid2>]` instead of ignoring the node.
> This can happen when a node is replaced but the `system.peers` still shows
> the ghost node in there with a different schema version than the rest of the
> nodes.
> Raised [here,|https://github.com/apache/cassandra-gocql-driver/issues/1736]
> with a PR open
> [here.|https://github.com/apache/cassandra-gocql-driver/pull/1738]
> EDIT by [~joaoreis] : the driver should ignore DOWN nodes similar to what the
> java driver does (check my comment below) and this will fix the issue with
> ghost nodes even if we're not specifically trying to fix the ghost node issue
> since the driver can't tell whether a node is a ghost node or not.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
