XOA can't start after installation and contradictory messages
While installing XOA the official way:
[host:~]# bash -c "$(curl -s http://xoa.io/deploy)" Welcome to the XOA auto-deploy script! Network settings: IP address? [dhcp] 10.1.0.4 Netmask? [255.255.255.0] Gateway? 10.1.0.254 dns? [22.214.171.124] NS.ip.address.here Importing XOA VMâŚ Booting XOA VMâŚ There are no suitable hosts to start this VM on. The following table provides per-host reasons for why the VM could not be started: host : Cannot start here [Host cannot attach to network: 77097537-8e7b-f375-abd6-d638ac42abcc (Pool-wide network associated with eth0)] There were no hosts available to complete the specified operation. Waiting for your XOA to be readyâŚ Your XOA is ready on https://10.1.0.4/ Default UI credentials: firstname.lastname@example.org/admin Default console credentials: xoa/xoa VM UUID: e58ce2ec-4fe4-5a9f-70b1-f6a5ca204946
is ready on https://10.1.0.4/?
No, it's not ready. It's shut and the Networking tab in XCP-ng-Center shows:
"Network 0 192.168.100.144".
Maybe the reason is that the address 10.1.0.4 which I'm trying to give to XOA is from one of the VLANs, but the installer tries to use a "Network 0" (which doesn't exist at all, as the NIC is Bond0+1). But why does it configure 192.168.100.144? It doesn't exist here. BTW, at a previous XOA installation the issue was the same, but the (bad) address was 192.168.100.190. Some tip here maybe.
Then I manually start the VM, login at the console and set the correct IP address, but for less experienced users these contradictory messages can be misleading.
- This is an edge case when you can't boot XOA because XAPI tries to find a match to "network 0" of the template (the xva file). Usually, 99% of users have a network 0. That's why you are the first to report this problem. In short, a non-handled error case.
- The "old" IP you can see is the one before changing, the one XOA had before being transformed into a template.
The thing is: this is how XVA files are working, and there's no pre-deploy option to select a network for a XVA.
@olivierlambert I see, thanks.
Anyway, maybe the installation script could check whether the XOA really started and not misleadingly inform
Your XOA is ready on https://10.1.0.4/?
As I said, it's an edge case we didn't think before. So, obviously, we'll improve this (but first we need to reproduce the issue). Can you print the output of
xe network-liston this host? I'm trying to understand why you don't have a network 0.
# xe network-list uuid ( RO) : 64aeff7a-42ed-e455-fcc0-1f6b259046cb name-label ( RW): VLAN4 name-description ( RW): VLAN4 bridge ( RO): xapi3 uuid ( RO) : 19d85500-c600-7476-13d4-6cb45b068186 name-label ( RW): Bond 0+1 name-description ( RW): bridge ( RO): xapi2 uuid ( RO) : f9718515-8820-96d3-0ab9-1195d1c27def name-label ( RW): VLAN21 name-description ( RW): VLAN21 bridge ( RO): xapi6 uuid ( RO) : adfcd92c-c127-8058-6f0e-fd0258383738 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 uuid ( RO) : 7c168d4b-6e5e-b2fa-249a-167792c9c641 name-label ( RW): VLAN7 name-description ( RW): VLAN7 bridge ( RO): xapi4 uuid ( RO) : c2fe838d-b71c-7520-8fc4-5998656c2a83 name-label ( RW): VLAN20 name-description ( RW): VLAN20 bridge ( RO): xapi1 uuid ( RO) : 77097537-8e7b-f375-abd6-d638ac42abcc name-label ( RW): Pool-wide network associated with eth0 name-description ( RW): bridge ( RO): xenbr0 uuid ( RO) : 7ea15b78-ff08-1285-f6d4-922fa35a22ca name-label ( RW): VLAN8 name-description ( RW): VLAN8 bridge ( RO): xapi5 uuid ( RO) : 6c8acc25-f1f8-8d8f-c5c3-ee4c519d9825 name-label ( RW): Pool-wide network associated with eth1 name-description ( RW): bridge ( RO): xenbr1 uuid ( RO) : 591ab537-8a76-1526-82ca-97908b86c444 name-label ( RW): VLAN3 name-description ( RW): VLAN3 bridge ( RO): xapi0
Thanks, we'll try to see if we could detect this case for an update on deploy script
MajorTom last edited by MajorTom
@olivierlambert You're welcome.
As a start, to prevent false positive message that the XOA VM is ready at some IP address, the script could check the power-state in the result of
# xe vm-list name-label=XOA uuid ( RO) : e58ce2ec-4fe4-5a9f-70b1-f6a5ca204946 name-label ( RW): XOA power-state ( RO): running
(whether running or not).
Low351 last edited by
I realize this is an old topic, but that make me realize it's unfortunate that this hasn't been fixed at this point.
I tried to use the command line to install XOA and it complains with the same error. In my case it's because I am using bonded interfaces just like the OP. Surprising that it doesn't come up more often but I guess if you set this up before you bond the interfaces it's not an issue.
@Low351 that's because the deploy script isn't the main/advised way to deploy XOA So it's lower in our backlog (vs the deploy UI from our website or even from within XCP-ng web page)
edit: I'll check with our teams if someone could add network selection quickly in the deploy script (which will hopefully solve the issue)
edit 2: investigating IF this will solve the issue first
low351 0 last edited by
@julien-f Yes, I realize that after. I finally just downloaded the ova and installed it that way, it's working fine. Thank you!
Low351 last edited by
@olivierlambert It would be nice to have but as you stated originally not a real priority. I installed using the .ova without issue. Thanks for responding.
FYI, we modified the deploy script to be able to automatically modify XOA's VIF to run on the management network.
So even in case of using only bonded networks like @Low351, this will now work.