perfSONAR User Q&A and Other Discussion

[Andrew Lake <>]

Hi,

By default perfSONAR installs a policy that keeps data for 6 months. If you don’t think that is working properly then you can run “psarchive ism_reset” to reapply this policy.

It may also be the case that its working and this is the amount of data being generated. If you want to more easily look at the Opensearch data and see what is taking up space I’d recommend doing the following:

1. Install the “perfsonar-dashboards” package on your archive host with either “dnf install perfsonar-dashboards” or “apt install perfsonar-dashboards”. This installs something called Opensearch Dashboards (different than Grafana) that is a UI for managing Opensearch. Basically their version of Kibana if you are familiar with Elasticsearch.

2. Grab the admin password from /etc/perfsonar/opensearch/opensearch_login on archive host

3. Go to https://<yourhost>/opensearchdash

4. Login with the admin username and password from step 2. 

5. Select the “Global Tenant” if it asks you about the tenant you should use.

6. Using the menu in the upper left, select Index Management.

7. If you click data streams it should tell you what indices are using the most space. Also, the “State Management Policies” you can edit to delete data sooner. You can also delete data directly through the UI. 

Hope that helps at least understand where the disk space is getting used the most. 

Thanks,

Andy

On October 17, 2024 at 3:16:19 PM, Matthan Mamaclay () wrote:

Hello,
Has anyone found a fix to this yet? We just had to rebuild our grafana
archiver after it was just rebuilt on 9/30/2024. We have 21 testpoints, VM
using 100G disk, on version 5.1.3.

70G /var
69G /var/lib/opensearch/nodes/0/indices
69G /var/lib/opensearch/nodes/0
69G /var/lib/opensearch/nodes
69G /var/lib/opensearch
69G /var/lib


Thanks!
--
To unsubscribe from this list: https://lists.internet2.edu/sympa/signoff/perfsonar-user