[
https://issues.apache.org/jira/browse/OAK-3727?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15097905#comment-15097905
]
Thomas Mueller commented on OAK-3727:
-------------------------------------
http://svn.apache.org/r1724565 (trunk)
The broadcasting cache (the TCP flavor) is enabled by default. To disable it,
use "broadcast=disabled", as in:
{noformat}
launchpad configuration (escaped):
persistentCache="crx-quickstart/repository/cache,size\=1024,binary\=0,broadcast\=disabled"
{noformat}
The configuration is stored in the "clusterNodes" collection. In case of
MongoDB, you may see:
{noformat}
> db.clusterNodes.find().pretty()
{
"_id" : "1",...,
"broadcastListener" : "10.132.4.224:9800",
"broadcastId" : "3dcc2c42-7aad-4473-95f1-35d832208d76"
}
{
"_id" : "2", ...,
"broadcastListener" : "10.132.4.224:9801",
"broadcastId" : "eab6e092-25de-4351-9aca-b047ac62040d"
}
{noformat}
This means the the listener is at IP address 10.132.4.224 and the given port. A
new, randomly generated broadcastId is used whenever a cluster node starts.
> Broadcasting cache: auto-configuration
> --------------------------------------
>
> Key: OAK-3727
> URL: https://issues.apache.org/jira/browse/OAK-3727
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: documentmk
> Reporter: Thomas Mueller
> Assignee: Thomas Mueller
> Fix For: 1.3.14
>
>
> The plan is, each cluster node writes its IP address and listening port to
> the clusterInfo collection, and (if really needed) a UUID. That way, it's
> possible to detect other cluster nodes and connect to them.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
