topic Re: ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration in ExtremeCloud IQ- Site Engine Management Center

ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration

Mancon — Thu, 29 Aug 2024 09:26:40 GMT

Environment :
-ExtremeCloud IQ - Site Engine 23.4.12.3
-Server JRE Version: 1.8.0_372
-ZTP POC device : 5320-48P-8XE-FabricEngine (9.0.2.0)

I've been working on deploying a Zero Touch Provisioning solution and encountered some difficulties when adding a device to a site.

The first steps of the process goes well :
1. My device gets a DHCP address from the DHCP server
2. My device contacts XIQ-SE using DNS information the DHCP server provided
3. the device starts ZTP+ process and shows up in "Network > Discovered" menu on XIQ-SE

From now on, I don't know if I missing something or doing something wrong :
I select my device to configure it, I choose its site and click on "Sync from Site" button to get the good parameters, I then go to "ZTP+ Device Settings" to configure an IP previously set to N/A ("Use Discovered" parameter is set to "Disabled" since the device IP stays to N/A when choosing any other parameter, which isn't really clear why to me)

Then, I save configuration and manually reboot my Switch (I also tried to wait for it to reboot). ZTP+ status change from Pending to Staged, and it's at this state things starts to become problematic.

From the Operations menu, I can see that the configuration gets stucked at 80%. (As shown on image below)

When I go to ZTP logging (from Administration > Diagnostics > Server > Server Log)

I can see the following logging message (I will also join a screenshot from the log message to this post) :

2024-08-29 11:14:45,913 ERROR [java.lang.Class] ManagementService is invalid value for enumeration com.extreme.common.ezconfig.configblocks.VlanConfigRole. Supported values are : OutOfBandMgmt, VLAN
2024-08-29 11:14:50,533 ERROR [java.lang.Class] ManagementService is invalid value for enumeration com.extreme.common.ezconfig.configblocks.VlanConfigRole. Supported values are : OutOfBandMgmt, VLAN
2024-08-29 11:14:50,537 ERROR [com.enterasys.netsight.server.webapps.monitor.ezConfig.ZtpPlusDeviceDispatcher] ZtpPlusDeviceDispatcher.doConfigurationPut:: java.lang.NullPointerException

By the way, I tried to set the parameter "Management Interface" in "ZTP+ Device Settings" from my ZTP configuration template both to "Out-Of-Band" and "Default", which didn't solve the issue.

Any help would be appreciated

Re: ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration

Stefan_K_ — Thu, 29 Aug 2024 09:37:27 GMT

I suspect it's related to this: Solution: ZTP+ fails with VOSS 9.03+ firmware and XIQ-SE prior to 24.7.10 | Extreme Portal (site.com)

23.4 is fairly old, VOSS 9.0.2 is fairly new.

Re: ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration

Ludovico_Steven — Thu, 29 Aug 2024 12:29:55 GMT

> then go to "ZTP+ Device Settings" to configure an IP previously set to N/A ("Use Discovered" parameter is set to "Disabled" since the device IP stays to N/A when choosing any other parameter, which isn't really clear why to me)

Maybe you should give us a screen shot of what you are doing there...

The switch came in with a DHCP IP on some VLAN (probably onboarding VLAN 4048); that pull down determines how much of that discovered data (DHCP IP & VLAN) you want to use. If you set it to IP & Mgmt interface, it will keep both, the IP will be made a static IP.

If you are happy with the VLAN, it is usually better to select "Mgmt interface" and then provide an IP range on same IP subnet but outside the DHCP scope.

Or else you disable, as you did, now you have to provide the IP range + select the VLAN you want to use; on FabEng/VOSS the only option that will work here is if you set Mgmt interface pull-down to value "Management Service", but for this to work you need to have under the same Site services a service flagged as mgmt:

you need to set that up under Service Definitions, and then assign that definition to your site:

Maybe you did not do that, hence the error message...

Re: ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration

Robert_Haynes — Thu, 29 Aug 2024 12:59:43 GMT

Correction to my prior post.

For your specific issue it is addressed in 23.7.10 where ManagementService was added. In 24.7.10 a new service ManagementClip has been added. Essentially in both cases the switch is advertising a type that SE does not understand in the version of SE you are using.

Keeping up to date with SE as you're also upgrade / using more recent switch firmware is always advised.

Re: ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration

Mancon — Thu, 29 Aug 2024 14:02:11 GMT

For anyone having a similar issue

Using an older version of voss (8.8.2.0) helped me working with this issue. An update of XIQ-SE is mandatory to use most recent version of voss

Thanks everyone for your help

Re: ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration

Brent_Addis — Wed, 11 Sep 2024 09:07:50 GMT

Hi,

Are you able to provide more information here?

We are seeing the same issue onboarding 4220's. We are running a fresh install of ExtremeCloud IQ - Site Engine 24.7.10.52

Originally it was set to use the management and ip interface, however your information above has mentioned a new service "ManagementCLIP" which we have also tried with no difference.

Have tried:

The logs are very unhelpful:

Thankyou very much,

Re: ZTP+ Device configuration remains stucked at 80% when XIQ-SE sends staged configuration

Brent_Addis — Wed, 11 Sep 2024 09:36:21 GMT

Hi Ludovico,

We are having a same problem with a new out of box 4220 running switchengine. Is there any advice you can offer there? We have upgraded firmware to latest available (32.7.1.9) while SE is running Site Engine 24.7.10.52

Logs are not very helpful:

We have a client Demo we are trying to prepare for and this is killing our scheduling. Our Extreme Networks SE out of Australia has been poking at it however he has not seen this problem before.

Happy to provide any information you might need so we can work this out.