IsilonSD Edge: Troubleshooting

Note: This topic is part of the IsilonSD Edge - Isilon Info Hub.

 

This collection of common troubleshooting topics applies to IsilonSD Edge. For issues not covered here, post to the EMC Community Network (ECN) Isilon community with as much detail as possible about your specific problem.

 

Are 512e  and 4K native drives supported?

No, 512e  and 4K native drives are not supported in IsilonSD Edge. See the VMware Knowledge Base article Support statement for 512e and 4K Native drives for VMware vSphere and VSAN (2091600) for more information.

 

Are there any specific privileges required to register the management plug-in with VMware vCenter?

Full vCenter administration privileges are required to register the management plug-in with vCenter. Use root credentials to register the management plugin initially and optionally assign a sub-admin to deploy the clusters. If you are stuck in the registration process due to lack of privileges, try rebooting the management server instance using port 5480.

 

How do I address browser compatibility or browser refresh errors?

Sometimes browsers don’t get refreshed correctly or might have been interrupted during the IsilonSD cluster deployment process. In such cases, you might end up with browser-related errors. Try the following workarounds to fix these errors:

  • Use a different vCenter client session in another window.
  • Wait and refresh.

 

Why am I unable to see the management plug-in in the vCenter tab?

The cause for this issue is unknown and it does happen occasionally. Here are some tips to resolve the issue:

  • After registering the management plug-in, log out of vCenter and login again.
  • Stop and restart the vCenter Web Client service.
  • Manually add the DNS search domain to /etc/resolv.conf on the IsilonSD management console using an SSH client.

 

Why do I see hung instances of the management plug-in?

The management plug-in instances might hang under the following scenarios:

  • You have removed the management server without unregistering it.
  • You have more than one management server registered at a time.

Here are some tips to resolve this issue:

  • Unregister the unwanted management server instance before registering the new management server instance.
  • Unregister the management server extension.

 

How do I clean up the management server file extension?

  1. Go to https://<vcenter-ip>/mob.
  2. Click content > ExtensionManager. You should see "com.isilon.vonefs" as a management server extension.
  3. Click UnregisterExtension.
  4. Enter com.isilon.vonefs and then click Invoke Method to unregister the management server extension.

If the above steps do not clean up your previously installed IsilonSD Management plug-in files and you are unable to see the IsilonSD Management plug-in that you newly installed, perform the following additional steps:

  1. Unregister the vCenter server from the management server.
  2. Log out of the VMware vSphere Web Client.
  3. Clean up the IsilonSD Management plugin installation files by performing the following steps:
    1. log in to the VMware vCenter server and navigate to the location where the plugin installation files reside.
      On the Windows operating system, the plugin installation files reside at:
      %PROGRAMDATA%/VMware/vSphere%20Web%20Client/vc-packages/vsphere-client-serenity/
      On the Linux operating system, the plugin installation files reside at:
      /var/lib/vmware/vsphere-client/vc-packages/vsphere-client-serenity/
    2. Delete the files with the com.isilon.vonefs extensions to remove traces of the installed plugin bits.
  4. Restart the VMware vSphere Web Client.
  5. Register the vCenter server with the management server.
  6. Log in to the vCenter server.
  7. Click Administration > vCenter Server Extensions and check whether the IsilonSD Management plugin appears. If for some reason, the plugin is not seen, perform the following steps:
    1. Click Administration > Client plug-ins, and then click Check for new plug-ins. This might take some time. Wait until the process completes.
    2. Log out of the VMware vSphere Web Client and log back in.
    3. Navigate to the Manage tab to check whether you are able to see the IsilonSD Clusters tab.

As one last check, make sure that vCenter can access the management server using TCP port 9443.

 

There were alerts during the installation. Do I need to troubleshoot them?

  • External network alerts during installation are expected and can safely be ignored.
  • Because IsilonSD clusters require a higher level of protection than physical OneFS clusters, if you configure a new IsilonSD cluster or a new node pool in an IsilonSD cluster, the default suggested protection level for the newly created node pool is incorrect. You can ignore the suggested protection level.

If there are any other alerts, post to the EMC Community Network (ECN) Isilon community with as much detail as possible about the alerts.

 

Why do I see alerts about datastore usage in vCenter after deploying an IsilonSD cluster? Is it serious?

After deploying a cluster, you will see some alerts about the datastore usage in the Alarms pane in vCenter. This is because the cluster creation process formats and uses up all the space in each datastore allocated for a data disk. The space required does not increase subsequently. Therefore, you can ignore the alerts.

 

Why am I encountering SSL errors while deploying the management server?

SSL errors typically occur due to browser caching or due to DNS-related issues. There is another possible reason. If you have not specified the correct FQDN or IP address of the vCenter server in the System Name [FQDN or IP address] field in the Network Settings page on the VMware vCenter Server Appliance Deployment screen, you will not be able to proceed further even if you have accepted the SSL certificate. Reason being that vCenter encodes the system name (PNID) in the SSL certificate of the system so that the components can communicate with one another using this name. You cannot change the vCenter system name (PNID) after the vCenter instance is deployed. Refer to this article for specific details. To overcome this issue, re-deploy vCenter with the correct FQDN or IP address. Alternatively, perform these steps:

  1. Connect to the serial console of the management server through an SSH client.
  2. Run nslookup <vCenter_system_name (PNID)>
  3. If the nslookup command fails, modify the /etc/hosts file to add the entry that maps the vCenter system name(PNID) to the vCenter IP address.

 

What cleanup operations should I perform after an unsuccessful deployment of an IsilonSD cluster?

Unsuccessful deployments could result because of the following reasons:

  • Errors in your IsilonSD cluster configuration, virtual machine creation, or errors while using the datastore
  • Issues while joining the cluster or when setting up licenses
  • Possible timeouts

 

Perform the following cleanup operations:

  • Check the logs. Refer to IsilonSD Cluster Administration: Frequently Asked Questions for information regarding the logs.
  • If the cluster shows up within the plugin, delete the cluster, and put the storage back into the pool.
  • If the cluster does not show up within the plugin, delete the virtual machines from the selected ESXi hosts.
  • Remount the configuration.

 

How can I verify that the management server was successfully deployed?

Check your VMware vSphere inventory. If deployment succeeded, the management server virtual machine (VM) will be listed there. See the version-appropriate IsilonSD Edge With IsilonSD Management Server Installation and Administration Guide for details.

 

How do I address the problems that are encountered when creating an IsilonSD cluster?

Refer to the Isilon Troubleshooting Guide, Troubleshoot IsilonSD Edge Installation Problems, for guidance for troubleshooting cluster creation issues. If these guidelines don't address your issue, post to the ECN Isilon community with as much detail as possible about the problem.

 

How about setting protection levels for my clusters?

IsilonSD Edge requires a higher level of protection than OneFS that is running natively on Isilon hardware. When you configure a new IsilonSD Edge cluster or node pool, increase the protection level from the default [+2d:1n], or "tolerate failure of 2 drives OR 1 node",  to [+3d:1n1d], or "tolerate failure of 3 drives OR 1 node and 1 drive".

 

To increase the protection level from the OneFS web administration interface, choose:

[+3d:1n1d] Tolerate failure of 3 drives OR, 1 node AND 1 drive

from the Requested Protection menu in the Edit Node Pool Details dialogue for any configured node pools.

 

To increase the protection level from the command-line interface, run the following command where <node_pool_name> is the name of the node pool you want to update. This change should be made for all configured node pools:

isi storagepool nodepools modify <node_pool_name> --protection policy=+3d:1n1d

 

See the OneFS CLI Administration Guide for more information.

 

Why does the management server fail to power on?

Sometimes, the management server fails to power on with an alert similar to the following:


“Cannot initialize property ‘vami.netmask0.IsilonSD_Management_Server’. Network ‘port_group_name’ has no associated network protocol profile


In some network environments, you must configure a vCenter network protocol profile before you deploy IsilonSD Management Server on the ESXi host. For example, a network protocol profile might be required if your network relies on static IP addresses or a DHCP server. Refer to Create a network profile for steps to add a network protocol profile.