#QoS stats on Volumes
1 messages · Page 1 of 1 (latest)
QoS stats on Volumes
@viscid thorn Could you share below.
1: Share your custom.yaml
2: Harvest start up logs. You can share them to us via email to ng-harvest-files@netapp.com
Hi Rahul, I've sent the email. thanks!
@viscid thorn Thanks received the logs. Have you restarted harvest after making changes to custom.yaml?
I see that logs are missing Harvest start up logs.
nabox-harvest2 | 2023-02-27T21:09:45Z INF ./poller.go:202 > Init Poller= configPath=/conf/harvest.yml logLevel=info version="harvest version 22.11.1-1 (commit 689c1cfb) (build date 2022-12-08T09:28:01-0500) linux/amd64\n"
Could you share logs after restarting Harvest.
I did a "dc down" and then scraped the log with dc logs -f nabox-harvest2 (after reboot). I've sent another output. Let me know if there is a better way
that works ok - instead of following (-f) you could capture everything like this dc logs nabox-harvest2 > nabox-harvest2.log mentioned here https://nabox.org/documentation/troubleshooting/ compress as needed
Yes and need to wait few minutes between dc down and log collection so that we capture full start up logs.
Sounds good. That is what I sent most recently. Let me know if that's not sufficient.
That is missing information. Notice the logs from another nabox setup after dc down It shows the templates being loaded.
Logs shared from your side have just below lines
nabox-harvest2 | 2023-03-01T15:18:00Z INF ./poller.go:202 > Init Poller= configPath=/conf/harvest.yml logLevel=info version="harvest version 22.11.1-1 (commit 689c1cfb) (build date 2022-12-08T09:28:01-0500) linux/amd64\n"
nabox-harvest2 | 2023-03-01T15:18:00Z INF ./poller.go:239 > started in foreground Poller= pid=95
Got it! Sent full log file. thanks!
Okay I can see that workload related templates were loaded by harvest
Could you check if you see metric qos_read_data in prometheus?
I can see data is being collected as well
Thanks for that information. The counter is there in prometheus. Most are 0. When I search for specific volumes with non-zero counters I do see the latency graphs. The message about enabling QOS counters it stays on regardless .
@viscid thorn glad you're getting data now - yes this msg will stay since it's static - we haven't come up with a way to hide/show that depending on what is in your yaml config files
Thanks again. I was reviewing several SVMs and they were not showing data... because there wasn't any 😆 ! Why I didn't look at the cluster level because we have hundreds of volumes. Thanks for the education !