Pools don't appear under Settings > Servers on fresh build of XO

bern

Yesterday, I build a fresh installation of XO on a newly build VM. The build was successful without any error.

Under Settings > Servers I could add my two xcpng servers. But under "Pool" I don't see my pools.

MissingPools_Screenshot from 2023-04-17 11-26-54.png

If I disable and re-enable the servers, the pools appear for a few seconds and disappear. Also, the pools will appear by themselves but will stay only for a few seconds.

I have no problem with the XO Appliance.

Someone has an idea of my problem?

Thanks

olivierlambert

If it works with XOA, it's likely related to your XO from the sources setup then (you should test also on the "latest" channel)

Check your XO to be on HEAD for master branch
Rebuild
Retest

olivierlambert

Hi,

Are those hosts in the same pool or not?

I would check xo-server output but also your hosts (dmesg, free disk space on root etc.)

bern

In my setup, I have one host with one pool.

Host          Pool
xcpng01       xcpng01
xcpng02       xcpng02

dmesg shows no problem on both xcpng servers.

The space looks ok to me:
[13:43 xcpng01 ~]# df -h 
Filesystem                                                                                                 Size  Used Avail Use% Mounted on
devtmpfs                                                                                                   637M  4.0K  637M   1% /dev
tmpfs                                                                                                      649M  132K  649M   1% /dev/shm
tmpfs                                                                                                      649M  9.2M  640M   2% /run
tmpfs                                                                                                      649M     0  649M   0% /sys/fs/cgroup
/dev/sda1                                                                                                   18G   11G  6.7G  61% /
xenstore                                                                                                   649M     0  649M   0% /var/lib/xenstored
/dev/sda5                                                                                                  3.9G  231M  3.4G   7% /var/log
/dev/mapper/XSLocalEXT--8e7231ff--ff9c--09d4--e3c9--11aef0a9a01d-8e7231ff--ff9c--09d4--e3c9--11aef0a9a01d   70G   24G   43G  36% /run/sr-mount/8e7231ff-ff9c-09d4-e3c9-11aef0a9a01d
tmpfs                                                                                                      130M     0  130M   0% /run/user/0

[13:46 xcpng02 ~]# df -h 
Filesystem      Size  Used Avail Use% Mounted on
devtmpfs        2.0G   24K  2.0G   1% /dev
tmpfs           2.1G  400K  2.1G   1% /dev/shm
tmpfs           2.1G   11M  2.0G   1% /run
tmpfs           2.1G     0  2.1G   0% /sys/fs/cgroup
/dev/sda1        18G  7.3G  9.5G  44% /
xenstore        2.1G     0  2.1G   0% /var/lib/xenstored
/dev/sda5       3.9G  340M  3.3G  10% /var/log
tmpfs           411M     0  411M   0% /run/user/0

Does the following could point to something?

In /var/log/orchestra.log, the following messages are repeated continuously about every 2 minutes.

 _watchEvents Error: HTTP connection has timed out
    at ClientRequest.<anonymous> (/opt/xen-orchestra/node_modules/http-request-plus/index.js:49:23)
    at ClientRequest.emit (node:events:513:28)
    at ClientRequest.patchedEmit [as emit] (/opt/xen-orchestra/@xen-orchestra/log/configure.js:52:17)
    at TLSSocket.emitRequestTimeout (node:_http_client:840:9)
    at Object.onceWrapper (node:events:627:28)
    at TLSSocket.emit (node:events:525:35)
    at TLSSocket.patchedEmit [as emit] (/opt/xen-orchestra/@xen-orchestra/log/configure.js:52:17)
    at TLSSocket.Socket._onTimeout (node:net:580:8)
    at listOnTimeout (node:internal/timers:573:17)
    at processTimers (node:internal/timers:514:7) {
  originalUrl: 'https://192.168.9.90/jsonrpc',
  url: 'https://192.168.9.90/jsonrpc',
  call: {
    method: 'event.from',
    params: [ [Array], '00000000000014811213,00000000000014795804', 60.1 ]
  }
}
_watchEvents Error: HTTP connection has timed out
    at ClientRequest.<anonymous> (/opt/xen-orchestra/node_modules/http-request-plus/index.js:49:23)
    at ClientRequest.emit (node:events:513:28)
    at ClientRequest.patchedEmit [as emit] (/opt/xen-orchestra/@xen-orchestra/log/configure.js:52:17)
    at TLSSocket.emitRequestTimeout (node:_http_client:840:9)
    at Object.onceWrapper (node:events:627:28)
    at TLSSocket.emit (node:events:525:35)
    at TLSSocket.patchedEmit [as emit] (/opt/xen-orchestra/@xen-orchestra/log/configure.js:52:17)
    at TLSSocket.Socket._onTimeout (node:net:580:8)
    at listOnTimeout (node:internal/timers:573:17)
    at processTimers (node:internal/timers:514:7) {
  originalUrl: 'https://192.168.9.91/jsonrpc',
  url: 'https://192.168.9.91/jsonrpc',
  call: {
    method: 'event.from',
    params: [ [Array], '00000000000021389621,00000000000021275380', 60.1 ]
  }
}

192.168.9.90 -> xcpng02
192.168.9.91 -> xcpng01

olivierlambert

It seems you have connectivity issues between your XO and your hosts

bern

My two xcp-ng servers, my XOA VM and my XO VM are all in the same subnet.

olivierlambert

Can you try with a real updated XOA to compare and check if it's your XO install?

bern

My XOA is at the current version:

Current version: 5.80.2
node: 18.12.1
npm: 8.19.2
xen-orchestra-upload-ova: 0.1.4
xo-server: 5.110.1
xo-server-telemetry: 0.5.0
xo-server-xoa: 0.19.0
xo-web-free: 5.113.0
xoa-cli: 0.35.0
xoa-updater: 0.45.1

I have no problem with my XOA. The pool are available under Settings -> Servers.

Maybe I should consider erasing my XO VM and start over. Maybe a glitch occured somewhere

olivierlambert

If it works with XOA, it's likely related to your XO from the sources setup then (you should test also on the "latest" channel)

Check your XO to be on HEAD for master branch
Rebuild
Retest

bern

I did use the master by using:
git clone -b master https://github.com/vatesfr/xen-orchestra

So I deleted my vm and started over. Everything works fine now. Maybe I did something wrong or a glitch occurred without being noticed.

Thanks for our help!

al-indigo

I've stuck with the same problem and rechecked everything for couple of hours to make sure: I've made it just as said in docs for the "from the sources part".
The test build was bullseye Debian and I followed all the recommendations.

It seems that it's just an inaccuracy in documentation; it says exactly "Please always use latest Node LTS." but that's the problem itself: XO seems not to work with NodeJS 20.x (that is also LTS) with the exact same errors from this ticket.

So just keep in mind that "Please always use latest Node LTS." means really "use NodeJS 16.x or 18.x versions for now" or you will get the errors like these.

Danp

@al-indigo said in Pools don't appear under Settings > Servers on fresh build of XO:

NodeJS 20.x (that is also LTS)

Can you show where v20 is shown as LTS? AFAIK, only v18 is considered LTS at this time.

Andrew

@al-indigo I agree, Node 20.x does not work for XO at this time. Pools/Hosts appear and then disappear. Using Node 18.x (LTS, as required by XO), everything works great.

While 20.x is listed as Current, it is not the correct 18.x LTS version needed for XO.

olivierlambert

Indeed, only Node 18 is the current LTS, so our doc is correct

swaio

@olivierlambert in Debian 12 it is 20.9.0 ...

Output the latest LTS Node.js version available
> n --lts
20.9.0

cat /etc/debian_version
12.2

Danp

The docs were recently updated to show that Node v18 is required instead of LTS.

swaio

@Danp Thanks for the hint, of course I didn't check the documentation .
For updates I use a script "xo-update.sh" from https://github.com/Jarli01/xenorchestra_updater.git . There it only checks if "node -v" is equal to "node -lts", which does not work and recommends to run "sudo n lts".

Danp

@swaio Yes, I'm aware of the issue with that script ATM. I haven't decided on the best way to fix it. Perhaps display a warning instead of aborting the script.

Here are some steps that should allow you to successfully update using the existing script --

sudo n lts
curl https://raw.githubusercontent.com/Jarli01/xenorchestra_updater/master/xo-update.sh | sudo bash 
sudo n 18
sudo systemctl restart xo-server.service

HTH, Dan

itservices

@Danp
Hi everyone.

That "n 18" did it for me. My Xen Orchestra would not hold the connection, with the same "http connection time out".

Thanks a bunch

Regards
Marc