Troubleshooting Cisco Intercloud Fabric Installation and Initial Setup

Issues with the Infrastructure Setup

This section includes symptoms, possible causes, and solutions for issues encountered while setting up the Intercloud Fabric infrastructure.

Symptom: After installing Cisco Intercloud Fabric, the GUI does not launch and is stuck on the loading page (with a picture of clouds).

Possible Cause: One of the services did not start correctly or took longer than expected to start.

Verification and Solution:

  1. Log in to Intercloud Fabric VM as shelladmin.

  2. Select option 3 to stop services.

  3. Select option 4 to start services and wait three to five minutes while services restart.

  4. Select option 2 to display the service status.

    After all services have restarted, the Intercloud Fabric launches.


    Note


    After restarting services in Step 3, you can repeatedly select option 2 to display the service status.


Symptom: There is an error when attempting to upload a Cisco Intercloud Fabric image tar file.

Possible Cause: This issue occurs if you use a browser other than Chrome. If you are using Chrome, it is possible that the image tar file did not completely upload.

Verification and Solution:

  1. If you are not using the Chrome browser, use Chrome and try again.

  2. If you encounter the error again, enable the root account by using the shelladmin credentials. The password for the shelladmin account is the same password that was entered when deploying the Intercloud Fabric appliance.

  3. Use the scp command to copy manually the tar file into the /opt/infra/uploads folder of Intercloud Fabric Director.

  4. On the Images page of the Infrastructure Setup dialog, click the Browse button and select the file from the local desktop.

  5. Click Next.

    Note


    Do not click the Upload button.


  6. Proceed with the remainder of the setup.

Symptom: The Intercloud image fails to successfully import into Cisco Prime Network Services Controller.
Error Message:
Handler failed with error - Image Upload Failed: End point timed out. 
  Check for IP, password, space, or access related issues.

Possible Cause: Cisco Prime Network Services Controller (Prime Network Services Controller) cannot connect to Cisco Intercloud Fabric.

Verification and Solution: If the import operation does not complete within a reasonable amount of time, check the Recent Jobs list in Prime Network Services Controller for the status and error message. The Task line item lists the last error encountered, if any, while importing and the number of retries attempted.

Symptom: After restarting Prime Network Services Controller services, the job for importing an Intercloud image shows as running, but the job does not progress.

Possible Cause: If Prime Network Services Controller services are restarted while the import operation is running, the task will not resume.

Verification and Solution: Cancel the job in Intercloud Fabric Director and import the image again.

Symptom: The Prime Network Services Controller server cannot resolve hostnames to IP addresses.
Error Message:
PNSC can't reach Public IP address

Possible Cause: The DNS server IP address is not entered correctly in the device profile. The DNS server resolves the hostname to an IP address. If ESX hosts were added to the VMware vCenter client using hostnames instead of IP addresses, those hostname-to-IP address mappings must exist in the DNS server.

Verification and Solution: Ensure that the DNS server IP address is entered correctly in the device profile, as follows:
  1. In the Policies tab, choose Intercloud Infrastructure Policies.

  2. If the DNS server IP address is incorrect, click Edit and enter the correct IP address.

For more information about configuring the DNS server IP address, see the Cisco Prime Network Services Controller User Guide.
Symptom: The Intercloud Fabric Director has an error regarding the shared secret.

Possible Cause: A shared secret that exceeds 11 characters was entered in the Intercloud Fabric Director.

Verification and Solution: Ensure that the shared secret is less than 12 characters.

Symptom: While deploying an Intercloud Fabric Director OVF template, VMware vSphere displays the error "The OVF package is invalid and cannot be deployed. The provided network mapping between OVF networks and the system network is not supported by any host."

Possible Cause: The VMware distributed virtual switch (DVS) is out of synchronization.

Verification and Solution: Ensure that the VMware DVS is properly synchronized with all hosts. For information about synchronizing the VMware DVS, see the VMware vSphere documentation.

Symptom: If the setting "Net.ReversePathFwdCheckPromisc=1" in the ESX host where the Intercloud Fabric Extender is hosted is not enabled, you might experience traffic loss or duplicate packages between enterprise and cloud VM traffic or Intercloud Fabric Switch module flap at the Intercloud Fabric VSM. This requirement applies only if you are using VMware virtual switch to host the Intercloud Fabric Extender; it does not apply if you are using a Cisco Nexus 1000V switch.

Possible Cause: If Intercloud Fabric Extender is hosted on a VMware vSwitch or distributed switch (VDS) and if the vSwitch or distributed switch is connected to multiple physical NICs, you must enable the setting Net.ReversePathFwdCheckPromisc=1 in the ESX host where the Intercloud Fabric Extender is hosted.

Verification and Solution: Enable the setting Net.ReversePathFwdCheckPromisc=1 in the ESX host where the Intercloud Fabric Extender is hosted. The setting is found under Host > Configuration > Advanced Settings > Net in the VMware vSphere UI.


Note


If the value of the Net.ReversePathFwdCheckPromisc configuration option is changed while the ESXi host is running, you must toggle (disable then re-enable) the Promiscuous Mode check box in the Intercloud Fabric Extender trunk port group security settings for the change to take effect.
  • For VMware virtual switch, you must set the trunk port group to allow All VLAN IDs in the VMware vSphere GUI.

  • For a security policy for the trunk port group on the VMware virtual switch, set the Promiscuous Mode, MAC Address Changes and Forged Transmits to Accept in the VMware vSphere GUI. This requirement applies only if you are using a VMware virtual switch. It does not apply if you are using a Cisco Nexus 1000v switch.