Update: I am also having the same problem with the new Nexus 9000v 7.0(3)I61 in VIRL. I can sometimes get the config extraction to work if i do one at a time with the new Nexus 9000v 7.0(3)I61 image (select the node, right click "extract configuration"), but even then it doesn't always work. "Extract configurations" for the CSR1000v always fails with the "Connection to node is unresponsive or broken".
I managed to extract the configurations from CSR1000V after I added the VIRL default credentials into the router config as if they woudl be built with AutoNetkit enabled.
username/password cisco/cisco and enable secret cisco
For Nexus 9000 it worked from the beginning. First time it starts, you have to define admin password. I ignored the complex password security and created username admin with password admin. I reckon extract config script relies on default credentials for Nexus. Try to put in admin/admin and test.
I have the same issue as other people in this post, mostly tested with IOSv.
The nodes in my simulations are unreachable at the start, i have to manually configure the management-interface(gi0/0) from CLI to make them reachable.
But everytime i try to extract configurations i get the following
(INFO) [Oct/05/2017 03:20:23] Attempt 1 out of 3 to download configuration on 3 nodes
(ERROR) [Oct/05/2017 03:20:23] Conection to node "iosv-1" is unresponsive or broken.
(ERROR) [Oct/05/2017 03:20:23] Conection to node "iosv-2" is unresponsive or broken.
(ERROR) [Oct/05/2017 03:20:23] Conection to node "iosv-3" is unresponsive or broken.
(INFO) [Oct/05/2017 03:20:23] Configuration extraction was skipped for all nodes.
(INFO) [Oct/05/2017 03:20:23] Done extracting configurations.
It actually ends with done extracting configurations but when i stop and start the simulation again they are back to default.
This issue started when i upgraded to Virl 1.3, does anyone know how to solve this or why it happens,
I check with IOSv and CSR1000V the Extract configuration worked as expected.
Procedure: New Lab> Build initial Configurations
INFO) [Oct/08/2017 20:28:11] Reachability status of node "iosv-1" changed from UNREACHABLE to REACHABLE
(INFO) [Oct/08/2017 20:28:20] Reachability status of node "iosv-2" changed from UNREACHABLE to REACHABLE
(INFO) [Oct/08/2017 20:29:50] Reachability status of node "csr1000v-1" changed from UNREACHABLE to REACHABLE
(INFO) [Oct/08/2017 20:30:14] Attempt 1 out of 3 to download configuration on 3 nodes
(INFO) [Oct/08/2017 20:30:18] Downloaded startup-updated config from node "iosv-2".
(INFO) [Oct/08/2017 20:30:18] Downloaded startup-updated config from node "iosv-1".
(INFO) [Oct/08/2017 20:30:19] Downloaded startup-updated config from node "csr1000v-1".
(INFO) [Oct/08/2017 20:30:19] Done extracting configurations.
Most reference platforms in my case are usually listed as unreachable. I do not use that automatic OOB management IP configuration feature, I use external access via FLAT network. I do not think VIRL extracts configuration via OOB management, it does it via the console. So for some reason it cannot get to the console for just that instance of yours.
As for not being able to use the extracted configs after a simulation restart, could you check that you do not have a tick in AutoNetKit for those instances? It might prevent it from using the extracted configurations.
It is a bit tricky with autoNetKit when it comes to small Linux VMs available to use in the simulations. You have to keep the autoNetKit tick on for them and actually build the autoConfig once, or else the default username and password are not provisioned.