I'm taking Nagios Log Server 2024R2 for a spin and while most of it looks familiar I've hit a snag. On R1 configuring elasticsearch to show a green cluster status on a single instance setup was simply a matter of editing elasticsearch.yml and uncommenting the line that says index: number_of_replicas: 0.
When I try the same thing on opensearch the service wont start and I get an error in the log saying I have to configure it using the API instead of on a single node. I have no idea how to do that and I can't find any updated documentation for R2.
2024R2 Single Instance
-
jmichaelson
- Posts: 383
- Joined: Wed Aug 23, 2023 1:02 pm
Re: 2024R2 Single Instance
That was a deliberate design choice in 2024R2. A single instance cluster will always show a yellow status; see this KB article https://support.nagios.com/kb/article.php?id=469
We made that decision in part because of the underlying OpenSearch design decisions where in a single instance, the cluster health will always show yellow, and in part because it is intended to be a clustered application; out of the box a given license from Nagios allows at least four instances, allowing each customer the ability to cluster Nagios Log Server from the beginning.
We made that decision in part because of the underlying OpenSearch design decisions where in a single instance, the cluster health will always show yellow, and in part because it is intended to be a clustered application; out of the box a given license from Nagios allows at least four instances, allowing each customer the ability to cluster Nagios Log Server from the beginning.
Please let us know if you have any other questions or concerns.
-Jason
-Jason
Re: 2024R2 Single Instance
After a little bit or research I figured out how to use the API.
The following line results in an {"acknowledged":true}:
curl -XPUT --cacert /usr/local/nagioslogserver/opensearch/config/root-ca.pem -u nagioslogserver:xxxxxxxxxxxxxxxxxxxx -H 'Content-Type: application/json' 'https://localhost:9200/nagioslogserver/_settings' -d '{ "index.number_of_replicas" : "0" }'
The cluster health is still yellow however. Even after a reboot.
Is there really no way to fix this? It's really annoying and I only need a single instance.
The following line results in an {"acknowledged":true}:
curl -XPUT --cacert /usr/local/nagioslogserver/opensearch/config/root-ca.pem -u nagioslogserver:xxxxxxxxxxxxxxxxxxxx -H 'Content-Type: application/json' 'https://localhost:9200/nagioslogserver/_settings' -d '{ "index.number_of_replicas" : "0" }'
The cluster health is still yellow however. Even after a reboot.
Is there really no way to fix this? It's really annoying and I only need a single instance.
Re: 2024R2 Single Instance
Hi @scarygary,
The purpose of the cluster health icon showing yellow instead of green is to indicate the health of your environment. The suggestion that you have multiple nodes in your cluster is less about what you need for your purposes and more about securing your data. The yellow status indicates that your cluster does not have any redundancy, meaning that in the event of server issues, it is possible for you to lose data. The ideal way to fix this indicator would be to add a secondary node to your cluster, insuring redundancy.
The purpose of the cluster health icon showing yellow instead of green is to indicate the health of your environment. The suggestion that you have multiple nodes in your cluster is less about what you need for your purposes and more about securing your data. The yellow status indicates that your cluster does not have any redundancy, meaning that in the event of server issues, it is possible for you to lose data. The ideal way to fix this indicator would be to add a secondary node to your cluster, insuring redundancy.