twin vapor
Hello, using nabox 3.2 and harvest 23.05.
From time to time, some storage system got disabled from monitoring, so need to go to the nabox -> systems -> and enable the on/off button for that system.
Is there a way to enable it by command line from nabox server?
On the old grafana there was a cli command that does this, but the new nabox has several docker containers running there and I cannot find how to do it.
Thanks
polar summit
hi @twin vapor if you ssh into nabox and run this it will restart a single poller. Replace $poller with the name of your poller. dc exec nabox-harvest2 /netapp-harvest/bin/harvest restart $poller
Also do you know why the system is disabled? That shouldn't happen. We can take a look at why that might be happening if you share you log files https://netapp.github.io/harvest/nightly/help/log-collection/#nabox
email them to ng-harvest-files@netapp.com
twin vapor
Thanks! Got into nabox-harvest2 and was trying to run: ./harvest status —config /conf/harvest.yml , but all clusters status seems as “not running” event they are all running from the gui view (nabox -> systems) . Trying to stop one of the clusters for test didn’t reflect a change on the gui. Any idea for this behavior?
polar summit
NAbox installs things a bit differently. Try this
docker exec -it nabox-harvest2 bash
to docker exec into the container that is running all Harvest pollers
then
cd /netapp-harvest
bin/harvest status
does that table look the way you expect? The pollers match what you expect and their statuses are correct?
As part of the container, NAbox is setting the environment variable HARVEST_CONF=/conf which causes Harvest to use the host mounted harvest.yml that NAbox manages. The harvest.yml file inside the container is /conf/harvest.yml which corresponds to the host file /opt/harvest2-conf/harvest.yml
twin vapor
I did this exactly - went to "docker exec -it nabox-harvest2 bash", then searched for all yml files and "/conf/harvest.yml" was the only yml file that contains that actual clusters configuration. But when I ran "/harvest status —config /conf/harvest.yml", all clusters appeared as "not running", which was not true - as the clusters were being monitored. I also tried to stop one of the clusters from monitoring using the "./harvest stop ..." but it doesn't seemed to be reflected on the gui view.
faint fern
@twin vapor Can you share you log files @ ng-harvest-files@netapp.com .
https://netapp.github.io/harvest/nightly/help/log-collection/#nabox
We have had similar problems in the past, where a process is being killed due to Out Of Memory (OOM) issues. How many clusters are you monitoring, and how much memory is allocated to NABox?