1. Using Oracle Cloud Infrastructure Storage Software Appliance
  2. Troubleshooting

11 Troubleshooting

This section provides solutions for problems you may encounter while using Oracle Cloud Infrastructure Storage Software Appliance.

Topics:

The provisioning tool keeps failing with an error that the Oracle Cloud Infrastructure Compute Classic endpoint is not valid. I'm sure that I've specified the correct endpoint in the configuration file.

Validating the Oracle Cloud Infrastructure Compute Classic endpoint is one of the first tasks that the provisioning tool does. If it fails at that point, and if you're sure that the endpoint that you've specified in the configuration file is correct, then the issue may be caused by formatting errors in the configuration file.

For example, if you edited the configuration file on a Windows computer and then did a binary transfer to the Oracle Linux host on which you're running the provisioning tool, then the file may contain extra line-ending characters.

To solve this problem, use the dos2unix file command to convert the file to the UNIX format. Alternatively, open the file on Linux using a text editor like vi, and run the set ff=unix command. Note that even if the line endings are fine in your configuration file, you can run the dos2unix file and set ff=unix commands safely.

If the invalid endpoint error continues, download the configuration file template afresh and edit it on your Oracle Linux host. If you prefer editing the configuration file in Windows, then when you transfer the file to your Linux host, do an ASCII (not binary) transfer. The steps to do an ASCII transfer depend on the file transfer client that you use. In FileZilla, for example, you should set Transfer type (in the Transfer menu) to ASCII.

I get a permission error like the following while creating the appliance.

{"message": "User /Compute-acme/jack.jones@example.com is not permitted to perform \"orchestration.add\" on orchestration:/Compute-acme/jack.jones@example.com/oscsa-boot"}

This error occurs if the Oracle Cloud Infrastructure Compute Classic user that you’ve specified in the configuration file doesn’t have the Compute_Operations role.

  • Ask your Oracle Cloud Infrastructure Compute Classic administrator to assign the Compute_Operations role to the user in Oracle Cloud My Services. See Managing User Roles in Managing and Monitoring Oracle Cloud.

  • Alternatively, update the configuration file to specify a user that has the Compute_Operations role. Then, try creating the appliance by using the updated configuration file.

I interrupted the provisioning tool before it finished running. What should I do now?

Run the tool again, with the same option that you specified earlier.

“Cannot satisfy both the placement and resource requirements” error during provisioning.

At times, you may get a long error message with the text Cannot satisfy both the placement and resource requirements, as highlighted in the following example, which has been truncated for readability: 

"errors":{"0":"re-launching instances: /Compute-acme/jack.smith@examp
le.com/vm"}},"obj_type":"launchplan","ha_policy":"active","label":...
..."state":"error","error_reason":"Cannot satisfy both the placement 
and resource requirements.",...,"info":{"errors":{"/Compute-acme/jack
.smith@example.com/gateway/vm":"error"}},"status_timestamp":"2016-01-
27T20:18:53Z","name":"/Compute-acme/jack.smith@example.com/gateway"}

This error occurs when the Oracle Cloud Infrastructure Compute Classic site doesn’t have sufficient CPU or memory resources to provision the appliance instance. Try specifying a smaller shape attribute in the configuration file. If the provisioning continues to fail, change the Oracle Cloud Infrastructure Compute Classic endpoint in the configuration file to a different site and try again. If the error persists, inform your Oracle Cloud Infrastructure Compute Classic administrator about the quota problem.

When I tried re-creating the appliance (by using the -r option), the provisioning tool exited with the following error.

{"message": "Orchestration \"/Compute-identity_domain/user/orchestration_name\" not found"}

This error occurs if you try re-creating the appliance after deleting it. It occurs because, when you delete the appliance (using the -d option), the orchestration that controls the instance is deleted from Oracle Cloud Infrastructure Compute Classic. If you want to create the appliance again after deleting it, then run provisioning tool with the -c option. See Creating the Appliance.

Error: Oracle Cloud Infrastructure Storage Software Appliance configuration file not found!

Make sure that the name and path of the configuration file that you specified when running the provisioning tool are correct, and run the tool again.

I forgot the appliance admin password. How do I reset or change it?

See Changing the Admin Password for the Appliance.

I’m unable to ssh to the Oracle Cloud Infrastructure Compute Classic instance that hosts my appliance.

See Can’t connect to an instance using SSH in Using Oracle Cloud Infrastructure Compute Classic.

The mount points on my client Oracle Cloud Infrastructure Compute Classic instances are unresponsive or unusable.

This problem occurs if the appliance instance is re-created. The appliance instance may have been re-created either manually (using the -r, -d, -u options of the provisioning script) or automatically by Oracle Cloud Infrastructure Compute Classic to recover from a crash of the instance or the physical node on which the appliance is provisioned.

  1. Verify that the appliance instance is running.

    Note that automatic re-creation of the appliance instance by Oracle Cloud Infrastructure Compute Classic may take a few minutes.

    If you can log in, using ssh, to the public IP address of the appliance instance as the opc user, or if you can access the login page of the management console (https://instance_public_ip_address), then the appliance instance is running.

  2. Find out the current private IP address of the appliance instance. See Finding Out the IP Addresses of the Appliance Instance.

  3. Do the following on each of the client Oracle Cloud Infrastructure Compute Classic instances on which you had previously mounted filesystems:

    1. Unmount each of the unresponsive mount points:

      sudo umount mount_point

    2. Mount the filesystem again.

      See Mounting Appliance FileSystems on Client Instances.

“Stale file handle” error on the NFS clients

If the Stale file handle error occurs and if you get the mount.nfs: Connection timed out error when you attempt to remount the appliance filesystem, then the appliance instance may have unexpectedly rebooted while under a heavy load.

To solve this problem:

  1. Stop the read and write operations of the filesystem. Disconnect the filesystem, and then connect it again. See Connecting a FileSystem.

  2. Stop the read and write operations of the filesystem. Unmount the filesystem, and then mount it again. See Mounting Appliance FileSystems on Client Instances.

If the problem persists, then run the following command to flush the entries from the kernel’s export table:

exportfs -f

Fresh entries will be added to the kernel’s export table, when you next mount the filesystems on the NFS client.

My application can’t write any more data to the appliance.

The filesystem cache may be full. Consider adding more data disks. See Adding Data Disks to the Appliance Instance.

The appliance instance isn’t responding and may have failed. How do I recover from this failure without losing data?

All the data that you’ve written to the appliance is safe. Data that’s been uploaded to your storage service instance is not affected by a failure of the appliance instance. Data that’s in the appliance cache is safe as well, because the block storage volumes that hold the cached data are intact.

Re-create the appliance using the same configuration file that you used to provision the appliance originally. See Re-creating the Appliance.

My appliance disks have crashed and are irrecoverable. What should I do?

  1. If you can log in to the management console, log in, and make a note of all the filesystem names, and their caching and encryption settings.

  2. Delete the appliance. See Deleting the Appliance.

  3. Provision the appliance again. See Creating the Appliance.

  4. Create the required filesystems, with the same names as in the old appliance. See Adding a FileSystem.

  5. Connect the filesystems to your storage service instance in the cloud. See Connecting a FileSystem.

    As long as the filesystem names are the same as they were in the old appliance, the filesystems will be connected to the correct containers that contain the data that was uploaded previously using the old appliance.

  6. Mount the filesystems on the client Oracle Cloud Infrastructure Compute Classic instances. See Mounting Appliance FileSystems on Client Instances.

Contacting Oracle for Support

  1. Go to https://support.oracle.com.

  2. In the Sign In pane, select Cloud Portal as the portal and click Sign In.

  3. On the Dashboard page, click Create Service Request.

  4. In the Create Service Request wizard, do the following:

    1. In the Service Type field, select your storage service instance in the cloud.

    2. In the Problem Type field, Select Issues accessing a Storage Software Appliance or Issues installing or upgrading a Storage Software Appliance, and select the appropriate problem subtype.

  5. Follow the prompts in the wizard to complete the service request.