NAS Session Top Client

[devout veldt]

Hi All ,

I’m on NAbox 3.5.2 , to the client we need to represent which are the top clients accessing the session - if I translate into command then it’s statistics top client show - how can I enable it in Grafana

NAbox - 3.5.2

Netapp harvest - 23.11.0-1

Thank you

[stiff lintel]

Steps are documented for nabox4 https://github.com/NetApp/harvest/discussions/3130 Is it possible to upgrade to NAbox4?

[devout veldt]

Hi Chris thank you but my NABox is on version 3 as I have mentioned in subject , can you please give me link for version 3 thank you

@paper vortex I see you have written a lot of harvest doc and also on version 4 for the same , do you have for version 3 thank you

[paper vortex]

@devout veldt I'll add steps for NABox3 in the same discussion page.

[devout veldt]

Rahul thank you , just for info my harvest is of 23.11

[paper vortex]

Okay. You need to upgrade to the latest version of Harvest, as support for Top Clients was added in version 24.11. It works via the RestPerf collector. If you are using ZapiPerf collector then you will need to migrate to RestPerf as a prerequisite.

[devout veldt]

@paper vortex thank you - NAbox is OVA so I will install new Harvest it’ should be .rpm or .tar.gz thank you

[paper vortex]

tar.gz

[devout veldt]

Thank you

[devout veldt]

I believe I can simply upload the file on my NAbox maintenance windows and click install to get it done

I will wait for your response on the link update for top clients for NAbox 3

[paper vortex]

Yes.

You need to click on Reset Harvest dashboards as well in Maintenance window.

[devout veldt]

1. download the tar.gz
2. upload the file and click on install
   3 ) click on reset on reset harvest dashboard

[paper vortex]

Yes

Also make sure that your clusters are using RestPerf collector in harvest.yml for cluster above 9.12 version.

cat /opt/harvest2-conf/harvest.yml

It is easier to do in NABox4.

[devout veldt]

This is as like this

[paper vortex]

If ONTAP version is above 9.12.0
Then you can edit and set them as

collectors:
  - Rest
  - RestPerf
  - Zapi
  - ZapiPerf
  - Ems

[devout veldt]

Thank you so I will add RestPerf in the collectors

It’s a great help

[paper vortex]

np. Make sure RestPerf comes prior to ZapiPerf in order.

[devout veldt]

Thank a lot

[paper vortex]

I have added NABox3 steps
https://github.com/NetApp/harvest/discussions/3130

[devout veldt]

Thanks a lot Rahul 🙏

[devout veldt]

Hey Rahul I have migrated the data to NAbox4 as per the given document it’s says me to go to /etc/nabox/harvest/user/restperf but I do not find restperf directory so where should I do entries

[stiff lintel]

hi @devout veldt, try creating that directory if it does not exist

[devout veldt]

Thanks Chris for prompt response

Done restated the harvest

Created the restperf directory created file custome_volume.yml and directory 9.12.0 , under this subdirectory created file as custom_volume.yml

And entered all data as given in document

Restarted the harvest

And still I’m getting the screen as attached

[stiff lintel]

@devout veldt can you grab a support bundle and upload to https://upload.nabox.org/vygy-quvu-yayu

[devout veldt]

I can

[stiff lintel]

thanks

[devout veldt]

Done

Please have a review and let me know what can I change

[stiff lintel]

Did you capture the support bundle by clicking About in the left navbar and then the Download Support Bundle button?

The file you upload has a different shape than expected

But I can tell from the file you uploaded that you are running ZapiPerf instead of RestPerf. Check your /etc/nabox/harvest/harvest.yml file on nabox and move the RestPerf collector above the ZapiPerf one for your poller in question

Example

[devout veldt]

It’s giving me error

When I try to create support bundle using NABox

[stiff lintel]

what's the error?

@unique dawn NABox is failing to create a support bundle for Avi

[unique dawn]

What error ?

[devout veldt]

errors™: "TokenAuthMiddleware: no bearer token", "JWTAuthMiddleware: no Authorization header", "BasicAuthMiddleware: unable to read credentials"))

This as error

[devout veldt]

@unique dawn

[unique dawn]

That’s not necessary an error. Those messages can be normal. If you reload the page you get to the login page or the same page ?

[devout veldt]

Reload gives me same page

[unique dawn]

And no file is downloading ?

[devout veldt]

No

Aah now it’s working

Wow

I’m not sure why

It’s not giving me any error and support bundle started to download

[unique dawn]

Maybe it took a while to start downloading.

[devout veldt]

Now no error (as mentioned) on NAbox page it started to download

[devout veldt]

@unique dawn and @stiff lintel

Seems after migrating to NAbox 3 alert configured to send has stopped I have done in NAbox 3 entries in /usr/local/nabox/docker-compose but in NAbox 4 I could not find same directory , how can I configure in Grafana to send email alert via alert manager

And one more question how can I confirm that it’s is only having 90 days data

[stiff lintel]

that is documented https://nabox.org/faq/#change-default-retention

[devout veldt]

Thank you Chris , I have modified the file in .env.custom for retention

I have one question in NABOx3 Grafana alert manager can be configured to delay already delivered email alerts for some hours , i was going through this document but unable to find where I can do it in NAbox4 , as Grafana alert system is not sending email but the NAbox 4 itself is sending email ,

[stiff lintel]

@unique dawn should be able to help with that

[unique dawn]

That would require changing alertmanager which sits in read only partition within NAbox

So it’s not possible today except by using a docker compose override file and copying configurations to r/w area in NAbox. I should probably do a write up describing the process

[devout veldt]

Hey yan please- as this rules is firing and sending so many emails

[devout veldt]

One more thing - please also document how to tweak the alert body ; and group by alert name

[devout veldt]

In alertmamger.yml under /etc/nabox/alertmamager - I see under option route - group by cluster - can we change it to “ alertname “ to send notifications on being grouped together with alerts like vol full or power down or node fail

Also tweaking which alert to fire and which one not to fire

[unique dawn]

That’s the goal eventually but it requires a bit of work. I believe all current alert come from harvest definitions. I’ll have to check. Give me some time to document alerts cli customization but you should disable emails in the meantime if that’s becoming an issue.

[devout veldt]

Thanks Yann , I will wait for your update