User Interface Issues

This section describes known issues and workarounds related to the graphical user interface.

Moving Resources Between Compartments Is Not Supported in the Compute Web UI

The Compute Web UI does not provide any function to move a resource from one compartment to another. Operations to change the compartment where a cloud resource resides, can only be performed through the CLI. However, note that not all resource types support compartment changes. For example, none of the network resources can be moved.

Workaround: If you need to move a resource from its current compartment to another compartment, use the CLI. After successfully executing the CLI command, you can see the resulting changes in the Compute Web UI.

Bug: 33038606

Version: 3.0.1

Saving Resource Properties Without Modifications Briefly Changes Status to Provisioning

If you open the Edit dialog box in the Compute Web UI to modify the properties of a resource, and you click Save Changes without actually modifying any of the properties, the status of the resource does change to Provisioning for a few seconds. This is the normal response of the UI to a user clicking the Save button. It has no adverse effect.

Workaround: To prevent the resource status from changing to Provisioning if you have not made any changes to it, click the Cancel button in the dialog box instead.

Bug: 33445209

Version: 3.0.1

NFS Export Squash ID Not Displayed

In the Compute Web UI, the detail page of an NFS export does not display the squash ID in the NFS export options. The squash ID is required for anonymous access to the NFS export, but you can only retrieve it by editing the export options.

Workaround: To obtain an NFS export squash ID, go to the NFS Export detail page, scroll down to the NFS Export Options, and click Edit Options. Alternatively, look up the export options through the CLI.

Bug: 33480572

Version: 3.0.1

Scrollbars Not Visible in Browser

The browser-based interfaces of Private Cloud Appliance are built with Oracle JavaScript Extension Toolkit(JET) and follow Oracle's corporate design guidelines. Scrollbars are meant to remain hidden as long as you are not actively using the part of the screen where content does not fit within the space provided – for example: large tables, long drop-down lists, and so on. Not all browsers or browser versions display scrollbars in the intended way. For example, Google Chrome typically hides the scrollbars as intended while Mozilla Firefox does not hide them at all.

Workaround: The behavior of the scrollbars is by design. It applies to both the Compute Web UI and Service Web UI. In areas where content runs beyond the allocated screen area, scrollbars appear automatically where appropriate when the cursor is placed over the content in question.

Bug: 33489195

Version: 3.0.1

Authorization Failure When Retrieving Compartment Data

The Identity and Access Management service allows you to control users' access permissions to resources in a fine-grained way through policies. Those policies determine which operations a group of users is authorized to perform on resources of a particular type or residing in a particular compartment. In certain situations, the Compute Web UI is unable to hide all the resources that a user has no access to. Consequently, a user operation may result in a request for data the account is not authorized to access.

While using the Compute Web UI you may run into authorization failures in case your operation triggers an attempt to retrieve data that you have no permission for. In this situation an error appears in your browser, indicating that the application has stopped working due to account permissions. The compartment tree, in particular, is prone to this type of failure because it can display compartments that you are not allowed to access.

Workaround: When the error is caused by a compartment tree access issue, it is likely that the intended page is displayed when you click Try Again. Otherwise, contact your tenancy administrator to request additional permissions to access the required data.

Bug: 33497526, 33520207

Version: 3.0.1

Object List Is Not Updated Automatically

In the Compute Web UI you display the objects stored in a bucket by browsing through its directory structure. The list or table of objects is not automatically refreshed at regular intervals, so any object changes will only become visible when you refresh the page manually. There is no available function for the UI to poll the status of a bucket.

Workaround: To display the current list of objects in the Compute Web UI, refresh the page manually. This behavior is not specific to the object storage service; it may occur in other areas of the UI as well. If a resource list is not updated automatically at regular intervals, you should refresh it manually.

Bug: 33519215

Version: 3.0.1

Not All Resources Shown in Drop-Down List

When you need to use a drop-down list to select a resource, you may notice that not all items are shown if the list is very long. As you scroll through the list, more items are loaded, yet you may still be unable to find the item you are looking for. The reason for this behavior is that UI components are designed to respond quickly rather than slowing down the user due to long load times. You are encouraged to filter a long list by typing part of a resource name in the text field, instead of scrolling through a complete alphabetical list. This is characteristic of Oracle JavaScript Extension Toolkit, so it affects both the Compute Web UI and the Service Web UI

Workaround: Scrolling is not the preferred way to search for an item in a long drop-down list. Instead, start typing the name of the resource you are looking for, and the available list items will be reduced to those matching what you type.

Bug: 33583708

Version: 3.0.1

Volume Group Can Be Created Without Name

When you create a volume group in the Block Storage area of the Compute Web UI, you are not required to enter a name. If you leave the name field blank, the volume group appears in the list as Unnamed Item. However, if you do not provide a name when creating a volume group in the CLI, a name is automatically assigned based on the time of creation.

Workaround: This is not a code bug: the name is technically not a required parameter. To avoid having volume groups with meaningless names, make sure you provide an appropriate name at the time of creation, in the Compute Web UI as well as the CLI. If you accidentally created the volume group without specifying a name, you can edit the volume group afterwards and add the name of your choice.

Bug: 33608462

Version: 3.0.1

File Systems and Mount Targets Not Displayed

When users have access to the resources in a particular compartment, but have no permission to view the content of the root compartment of the tenancy, the Compute Web UI might not display the resources that a user is allowed to list. Instead, an authorization error is displayed. For the file system service specifically, file systems or mount targets in a particular compartment are not displayed, even if the user has full access to that compartment and the resources it contains.

This behavior is caused by the way the API request is made from the UI, using the OCID of the root compartment. In contrast, the CLI requires that you specify the OCID of the compartment that effectively contains the requested resources, so it is not affected by the same authorization issue as the UI.

Workaround: The tenancy administrator should ensure that users of the file system service have read access to the root compartment. Users who cannot list the file systems and mount targets they are authorized to use, should ask their tenancy administrator to verify their account permissions and make the necessary adjustments.

Bug: 33666365

Version: 3.0.1

Optional ICMP Security Rule Parameters Cannot Be Removed

When you add an ingress or egress security rule to the security list of a VCN, you can specifically select the ICMP protocol. The Compute Web UI indicates that selecting a Parameter Type and Parameter Code from the respective lists is optional. This is incorrect, because the Parameter Type is mandatory for ICMP rules.

If you specified both Type and Code in your ICMP rule, it is possible to remove the Parameter Code. Edit the security rule, place your cursor in the Code text field, and delete its content. This is how drop-down lists work in the UI; there is no "empty" option to select.

Workaround: When working with ICMP security rules, always specify the Parameter Type. To remove an optional parameter selected from a drop-down list, select and delete the content of the text field.

Bug: 33631794

Version: 3.0.1

Compartment Selector Not Available When Creating DHCP Options

When you create or modify DHCP options for a VCN through the Compute Web UI, there is no way to add the DHCP options to another compartment. Because the compartment selector is not available in the create and edit windows, the DHCP options are implicitly stored in the same compartment as the VCN itself. However, it is supported to store DHCP options in another compartment. If you wish to do so, please use the CLI.

Workaround: If you want DHCP options to be stored in a different compartment than the VCN they apply to, create the DHCP options through the CLI, or use the CLI to move them to the desired compartment.

Bug: 33722013

Version: 3.0.1

Fix available: Please apply the latest patches to your system.

Custom Search Domain Error Not Rolled Back When Operation Is Canceled

The Networking service allows you to control certain instance boot configuration parameters by setting DHCP options at the level of the VCN and subnet. One of the DHCP options you can control is the search domain, which is appended automatically to the instance FQDN in a DNS-enabled VCN and subnet.

The search domain must be specified in the format example.tld – where TLD stands for top-level domain. If you attempt to save an invalid search domain, two error messages are displayed: one appears immediately under the Search Domain field, and the other at the bottom of the DHCP Options window, indicating that "Some fields are incomplete or invalid". When you cancel the creation or modification of the DHCP options and subsequently reopen the DHCP Options window, the error messages and the incorrect value may still be displayed, even though the settings were not applied.

Workaround: If you close and reopen the DHCP Options window after a failed attempt to save new settings, and the same error messages and invalid value still appear, you may ignore the error. Refreshing the browser window may clear the error message. Enter a custom search domain in the required format: example.tld

Bug: 33734400

Version: 3.0.1

DHCP Options Error Message for Custom Search Domain Is Misleading

The search domain must be specified in the format example.tld – where TLD stands for top-level domain. However, the Compute Web UI does not validate this parameter; this is done by the Networking service when you save the DHCP options. The Compute Web UI checks that the value contains no spaces. If it does, an error message appears under the Search Domain field: "Must be in the format of example.tld". This is technically inaccurate as it merely indicates the value contains a space.

Workaround: Enter a custom search domain in the required format: example.tld. Spaces are not allowed in domain names. If the error message in question appears, correct the value you entered in the Search Domain field and try to save the DHCP options again.

Bug: 33753758

Version: 3.0.1

Unclear Error when Logging in to Service Web UI with Insufficient Privileges

In the Service Enclave, an administrator who is a member of the SuperAdmin authorization group can create other administrator accounts and determine to which authorization group those accounts belong. It is possible to set access restrictions that prevent certain administrators from logging in to the Service Web UI, even though they can still access the Service CLI.

If this situation occurs, the Service Web UI returns an error message that does not clearly describe the authorization problem – for example:

{"message":"AUTH_000008: An error occurred getting system config state.",
"errorCode":"AUTH_SERVICE_GET_CONFIG_STATE_ERROR",
"cause":["Caused by: com.oracle.pca.ui.common.server.exception.PcaConsoleException:
SERVICE_000002: Calling underlying service resulted in an error: 
Failed to get PcaSystem object from admin service [Bad Request]."],"csrfToken":null,"idps":null,"serviceError":null}

Workaround: If you receive an error similar to this example, and you should be able to access the Service Web UI, ask an appliance administrator with the appropriate permissions to correct the access restrictions for your administrator account.

Bug: 34522989

Version: 3.0.2

No Details Displayed for File System Cloned from Snapshot

When you create a file system, all relevant information about the file system is displayed in the Compute Web UI detail page. It contains practically the same information as the output from the OCI CLI command oci fs file-system get. However, when you clone a file system snapshot through the OCI CLI, to use as a new file system, the detail page of the clone file system does not provide the same data. Relevant missing fields include Source Snapshot, Parent File System, Clone Root, Hydration, etc.

Workaround: Use the OCI CLI instead if you need those details. The CLI displays all the data fields for a clone-based file system.

Bug: 34566735

Version: 3.0.2

Error Message when Exadata Network Is Deleted Successfully

When you delete an Exadata network using the Service Web UI, an error message might appear that says the network could not be deleted. Check the Exadata Network table for confirmation. The delete operation might succeed despite the error message.

Workaround: If the Exadata Networks table indicates that a network was deleted successfully, you can safely ignore the error message that indicates the opposite.

Bug: 35644436

Version: 3.0.2

Unable to Display Details of Instance Backup

When you create a backup of an instance, it is stored in an object storage bucket. When the backup operation has completed, and you try to view the details of the backup object, the system may return an HTTP 404 error.

Workaround: Manually reload the browser page. The backup details should be displayed. Additional error messages might appear in a pop-up but those can be ignored.

Bug: 34777856

Version: 3.0.2

IP Address List on VNIC Detail Page Not Updated

In the Compute Web UI you can manage the IP addresses assigned to a compute instance from the detail pages of the instance's attached VNICs. In the IP Addresses table in the Resources section of the VNIC detail page you can add and remove private and public IPs. However, the Reserve Public IP pop-up window is not always rendered correctly, and changes applied there are not displayed in the VNIC IP Address table.

Workaround: No workaround is available.

Bug: 34797160

Version: 3.0.2

No Error Displayed When Attempting to Reserve Public IP Address While All Public IPs In Use

In the Compute Web UI, when you try to reserve a public IP address when no more public IPs are available from the pool, no error message is displayed. The Reserve Public IP window hangs, but provides no feedback about the operation you are attempting. Since no IPs are available to reserve, the operation does not complete.

Workaround: Close the Reserve Public IP window or reload the browser interface page. An administrator needs to expand the public IP pool for the Private Cloud Appliance environment before you can reserve another public IP address.

Bug: 34832457

Version: 3.0.2

When Modifying Listener of Load Balancer, Existing Configuration Parameters Are Ignored

When you use the Compute Web UI to modify a listener configuration of a load balancer, several parameters are not preserved in the Edit Listener window. Although you are changing the configuration of an existing listener, certain parameters are replaced with default values, forcing you to enter and verify all the parameters again.

As an example, let's assume your load balancer has a listener configured for HTTPS at port 8443, with an SSL certificate and cipher suite set up. When you edit this listener through the UI, the protocol appears as HTTP and the SSL configuration is not shown. Next, when you select the already configured HTTPS protocol again, the custom port, TLS version and cipher suite are reverted to their default values instead of the currently configured parameters.

Parameters that are not preserved in the Edit Listener window: protocol other than HTTP, custom port number, SSL settings (TLS version and cipher suite). Other parameters are not reverted: backend set, host names, path route set, idle timeout.

Workaround: If you modify a listener configuration through the Compute Web UI, always ensure that you have selected the right protocol first. Then set all listener parameters as if you were creating it from scratch, without moving back and forth between the policy options in the UI. Alternatively, modify the listener configuration using the OCI CLI.

Bug: 36032894

Version: 3.0.2

Create Backend to Add Security List/Configure Automatically Using IP Address Does Not Display Egress/Ingress Lists

If you use the following method in the Compute Web UI to create backend servers to a backend set of a load balancer, the egress and ingress security lists/rules are not displayed:

Select IP Address under the Backend Servers section, enter the backend server IP address, and then select Configure Automatically under the Security Rules section.

Workaround: Create the backend servers using one of these other methods:

Select Computed Instances under the Backend Servers section when creating the backend, and select security rules Configure Automatically.
Select IP Address under the Backend Servers section when creating the backend, and select security rules Configure Manually. Then, manually add the ingress and egress security list rules to the security lists attached to the corresponding subnet.

Bug: 35570701

Version: 3.0.2

Tag Area Above Resource Tables Does Not Expand

Resource pages in the UI display a table listing resources of a particular type. Just above the table you find control options to auto-reload the page, refresh the resource table, and filter the entries displayed. When filtering by tag, you can select multiple tags for your filter. However, the space available to show the filter tags does not expand, which causes elements of the UI to overlap as more tags are added.

Workaround: There is no workaround. Only the first few tags used for filtering are displayed while additional tags are hidden behind other UI elements.

Bug: 35975108

Version: 3.0.2

Service API Reference Displays Wrong Path for Disaster Recovery Commands

Developers can access the Service API Reference in a browser by appending api-reference to the base URL of the Service Web UI. For example: https://adminconsole.myprivatecloud.example.com/api-reference.

For commands related to disaster recovery configuration, the reference pages are split between the ComputeInstance and DrConfig nodes. This presentation is incorrect: all disaster recovery commands belong to the DrConfig object. As a result, the paths shown on the pages in the ComputeInstance section are wrong. Those commands will result in an HTTP 404 - not found error.

Workaround: While the request and response information is correct in the pages of the ComputeInstance section, you must replace that part of the object path with DrConfig. The following commands are affected:

Documented Path	Correct Path
`/admin/wsapi/rest/3.0.1/ComputeInstance/drAddComputeInstance`	`/admin/wsapi/rest/3.0.1/DrConfig/drAddComputeInstance`
`/admin/wsapi/rest/3.0.1/ComputeInstance/drGetComputeInstance`	`/admin/wsapi/rest/3.0.1/DrConfig/drGetComputeInstance`
`/admin/wsapi/rest/3.0.1/ComputeInstance/drGetComputeInstances`	`/admin/wsapi/rest/3.0.1/DrConfig/drGetComputeInstances`
`/admin/wsapi/rest/3.0.1/ComputeInstance/drRemoveComputeInstance`	`/admin/wsapi/rest/3.0.1/DrConfig/drRemoveComputeInstance`

Bug: 36267982

Version: 3.0.2

Details of an Instance Import Are Displayed as an Instance Export

When you import an instance from an object storage backup, it appears in the Instance Imports table of the Compute service. You click the instance import to display its detail page. However, the instance import detail page has the layout and data fields of an instance export detail page, including an Import button.

Workaround: Do not click the Import button; the instance has already been imported. The "export" text labels are wrong and can be ignored. The OCIDs and properties displayed on the page are correct for the instance import you selected.

Bug: 36310038

Version: 3.0.2

When Canceling an Instance Import the Operation Does Start

When you export an existing instance to object storage, it appears in the Instance Exports table of the Compute service. You can directly import an existing instance export using the Import button on the Instance Export detail page. A pop-up window appears where you either confirm or cancel the import operation. However, even if you click Cancel the import operations is started.

Workaround: Considering the import operation is started regardless of the option you select, either try to cancel the import in progress or delete it after it has finished.

Bug: 36311280

Version: 3.0.2