Netdata package is now available in XCP-ng

Danp

FYI, after upgrading my XOA VM to the latest version, I was unable to turn on the new NetData plugin. Instead, I received the message --

Netdata is not available on this appliance, please deploy a new one

Subsequently had issues attempting to install a new XOA VM. Tried both the known methods --

https://xen-orchestra.com/#!/xoa
bash -c "$(curl -s http://xoa.io/deploy)"

The first just sat at a spinning graphic but never deployed. The second just stopped without any warning after the import attempt.

Further investigation shows that both methods were yielding the following error during VM.import --

The value given is invalid
field: bridge
value: xapi0

I ended up downloading the file from http://xoa.io:8888, renaming to an XVA file and then importing with my other XO VM.

Don't know if these issues will affect others, but hopefully these details will prevent someone else from tearing out their hair.

olivierlambert

what was the XCP-ng version where you tried to deploy it?
do you have a xapi0 on it?

I did tests on XS 6.5 and XCP-ng 8.0, and deploy worked on my side.

Danp

olivierlambert v8 fully patched. Here's the output from xe network-list --

uuid ( RO)                : a379a6e7-f099-f838-6514-486de4d56e80
          name-label ( RW): Pool-wide network associated with eth0
    name-description ( RW):
              bridge ( RO): xenbr0


uuid ( RO)                : 3d13d4c3-546f-3f51-19c9-792a2d899a94
          name-label ( RW): Bond 1+2+3
    name-description ( RW):
              bridge ( RO): xapi1


uuid ( RO)                : b14a0bdd-2f77-25f7-3a76-b830ff971624
          name-label ( RW): Pool-wide network associated with eth2
    name-description ( RW):
              bridge ( RO): xenbr2


uuid ( RO)                : 00b6604f-9098-023b-5880-0ef9f8e1b437
          name-label ( RW): Pool-wide network associated with eth1
    name-description ( RW):
              bridge ( RO): xenbr1


uuid ( RO)                : e3be0e1a-9872-4b90-c851-35eb30a86920
          name-label ( RW): Pool-wide network associated with eth3
    name-description ( RW):
              bridge ( RO): xenbr3


uuid ( RO)                : ea7dce53-81ff-3b7c-155f-966721119bac
          name-label ( RW): Private LAN
    name-description ( RW): Internal LAN network
              bridge ( RO): xapi0


uuid ( RO)                : c54ecffe-9467-5d5f-cb73-ec10dedef2b9
          name-label ( RW): Host internal management network
    name-description ( RW): Network on which guests will be assigned a private link-local IP address which can be used to talk XenAPI
              bridge ( RO): xenapi

olivierlambert

Okay I think I get it. The template with its VIF was build on a machine with xapi0 network. For some reason, it fails to import on a host without this network.

I'll try to modify the XML to use xenbr0 that should be for everyone. Why XAPI fails? Why importing "directly" works? IDK.

olivierlambert

Danp I regenerated XOA with xenbr0 instead of xapi0. Can you try to redeploy?

Danp

olivierlambert The "quick deploy" via web worked correctly this time

lawrencesystems

I am getting "netdata plugin is necessary" when I mouse over the "Enable Advances Live Telemetry" option. The host is a fully patched XCP-ng 8 and I have tested with both XOA Free and the the self compiled XOA both are running xo-server 5.53.0 & xo-web 5.53.3. I manually installed netdata-ui.x86_64 & netdata.x86_64 via command line and netdata is working fine and is accessible via http://host_address:19999

I did reboot the XCP-host and both the XOA and the problem persists, what should I check next?

Danp

@flipsidecreations There's a new Netdata plugin, which appears to only be available in XOA Premium edition. It doesn't appear that you will be able to enable the new Advanced Live Telemetry using the versions you are currently running.

lawrencesystems

Danp Thanks, that makes sense. Perhaps is should show "Premium Feature" instead of "netdata plugin is necessary" when using the free versions.

olivierlambert

@flipsidecreations said in Netdata package is now available in XCP-ng:

Danp Thanks, that makes sense. Perhaps is should show "Premium Feature" instead of "netdata plugin is necessary" when using the free versions.

This was explained in my 1st post in this very topic We just released in time, so we'll make it more clear for next release BTW, I think you have a XOA Premium available If it's not the case anymore, just let me know.

olivierlambert

Danp said in Netdata package is now available in XCP-ng:

olivierlambert The "quick deploy" via web worked correctly this time

Thanks for the feedback!

kevdog

Not specifically related -- however FreeNAS is doing away with Netdata I believe in the 11.3 release upcoming -- say there is big memory link when running. I don't have specifics but there development team was pretty adamant about the shortcomings of netdata -- so much so they are dropping it. I find it curious while they are dropping support, xcp-ng is adding support. Any testing for this condition?

akurzawa

This post is deleted!

olivierlambert

kevdog not aware of memory leaks with Netdata. I'm graphing on long term on my own hosts. In XCP-ng, we are using the latest version.

codedmind

Just some feedback, don't know if is normal.

Have a host, with a VM ON but nothing run, before netdata plugin 24% load (stacked values) after netdata install and reboot the host reports 29%, so netdata need 5% cpu ?

Thanks

stormi

You can check what amount of CPU usage netdata uses in htop... or in netdata.

Note that netdata runs with a very low priority so in theory the overhead should not slow your important tasks down.

lawrencesystems

kevdog The issue with Netdata that caused it to be removed from FreeNAS was specifically a memory leak in the FreeBSD version of Netdata.

https://github.com/netdata/netdata/issues/7356

ncmans created this issue in netdata/netdata

closed Re-introduce Netdata to FreeNAS #7356

kevdog

@flipsidecreations

Thanks for clarification on that issue -- Weird however -- I read through a bunch of those links referenced in your post, and the last post was 2017 which fully described the problem. One of the things in general I hate about github are many tickets are closed without any mention of how the problem was resolved -- or if it wasn't just due to lack of inactivity in the thread. Freenas now dropping package based on this thread information that is almost 3 years old.

Biggen

Rebooted my host today and the Netdata graphing is messed up in several sections. The individual VMs are showing broken graphs for every report. The graphs are just dots. No lines.

Also under the “Disks” section, the graphs are working but the numerical reporting is broken. It’s showing all zeros for Read, Write, and Utilization even though the graphs are showing activity.

I’m on iPad now looking at it so I can’t post screenshots. Later today I’ll post some captures so you can see what I’m talking about.

Any other reports of Netdata not working after reboot?

kevdog

Small nitpick -- you can tell me to go away. Anyway to install certs so I can view the data using https?