| SC-2113 |
Running state changes to Creating state during application deployment. |
After deploying an application, then refreshing the Applications page, the deployed application state is reported as Running, and then seconds later is reported as Creating. |
After a few minutes, the application correctly reports as Running. |
| SC-2409 |
When deleting, you cannot switch to other tabs without closing the detail window. |
An error message may display if you try to switch to other tabs when a detail window displays. |
Close all detail windows before switching tabs. |
| SC-2456 |
Applications may need to be deleted, updated, and redeployed to pick up environmental changes. |
If you add new connections, update images, or make other environmental changes, currently deployed applications cannot make use of these changes. |
You may have to delete the running application, go back to the service catalog, and update the deployment package to pick up the changes; then redeploy the application. |
|
SC-2482
|
When creating a volume snapshot, the displayed status may remain as Creating. |
After clicking Create when creating a snapshot of a volume, the displayed snapshot status may remain as Creating for an extended period of time. |
Click Close in the Volume Snapshot panel. You are returned to the Volumes page, and the new snapshot is successfully created. You can also refresh the browser (F5) to update. |
| SC-2502 |
No indicator on Volumes page for undiscovered volumes. |
Undiscovered volumes display in the list view of the Volumes page, but there may be no indicator stating that they are undiscovered, other than missing option buttons. |
Undiscovered volumes can only be managed through OpenStack. |
| SC-2554 |
Changes to hypervisor status may not be immediately updated. |
Changes made to the hypervisor status may require several minutes to show as updated. |
None. |
| SC-2591 |
Tenant used by flavor or image can be deleted. |
A tenant in use by a flavor or image can be deleted. Tenants in use should not be open to deletion. |
None. |
| SC-2603 |
Paused or Suspended instances are not evacuated. |
OpenStack does not allow Paused or Suspended instances to be evacuated. |
OpenStack only supports evacuation of instances that are Active or Stopped. Paused or Suspended instances cannot be evacuated. |
|
SC-2623
|
Email warning error when creating new user. |
When creating a new user in an environment not set up to automatically send the email password to the new user, a warning message may appear stating that the user was created successfully, but the password email failed. Closing this warning message displays the Create New User panel, with the option to Create or Cancel. Clicking Create displays a User already exists warning. |
If the email warning error appears, then in the Create New User panel, click Cancel. The new user is created, and the Create New User panel closes. |
| SC-2713 |
A failed or incomplete install may result in an incorrect status message during re-install. |
When doing a clean installation of Workload Services, the installation script may fail due to being unable to verify OpenStack. If you attempt to run the install script again it reports the installation is already complete. |
Uninstall Workload Services, address any OpenStack issues, then re-install Workload Services. For more information, see Uninstalling Stratus Cloud Workload Services. |
|
SC-2733
|
The admin_pass property may not work as expected. |
After creating a deployment package with admin_pass set to 'password,' and successfully deploying the new deployment package, logging in with the administrator password (root/password) may fail. |
Currently, this is a limitation in the IceHouse version of OpenStack. Use SSH keys, and note that the cloud installation requires a cloud-init capable installation for the OS. This allows setting of the administrator password. |
| SC-2765 |
Instance status may display incorrectly when live migrating a Windows 2012 R-2 instance with a Cinder boot volume. |
Due to a limitation in OpenStack Ice House: when live migrating a Stratus-managed Windows 2012-R2 instance with a 30GB Cinder boot volume, the instance may remain in the "Paused" power state and may not be recovered. |
Create a backup of the boot volume and redeploy. |
| SC-2770 |
An Unhandled Exception advisory message may appear after adding a tenant to a flavor. |
After clicking Retry in certain advisories involving adding a tenant to a flavor, an unhandled exception error may display. |
Create all flavors as public, without assigning tenants to flavors. |
|
SC-2819
|
When an instance is evacuated to another compute node, OpenStack Nova does not preserve data on NFS shared storage. |
Nova preserves local storage, but if the /var/lib/nova/instances directory is mounted from an NFS share, Nova deletes and rebuilds the disk file for the instance and customer data is lost. |
A patch is available to correct this issue in the OpenStack IceHouse release. Contact your support representative for more information. |
| SC-2846 |
Possible "Too many connections" services error. |
If MySQL connections remain at the default installed specified number of connections, Heat and Keystone errors of "Too many connections" may occur. |
Stratus recommends setting the number of MySQL connections to 300. For more information, refer to MySQL Connections. |
|
SC-2877
|
If an instance fails to evacuate from a compute node, clicking Retry or Cancel in the Evacuation failed advisory does not recover the instance or cancel the operation. |
The Retry and Cancel buttons in the advisory are ineffective. |
After the compute node has recovered, go to the Deployed Applications page. For each instance that failed to evacuate, click  to display the Resource Usage panel and click  to reset the instance.
If the failed instances persist, restart the openstack-nova-compute service, which automatically cleans up the instances and allows further operations to proceed.
|
| SC-2883 |
Prior to deleting an application in a Failed state, all instances that are part of the application must be reset. |
When an application is in the Failed state, then one or more instances that are part of the application are an Error status in OpenStack, and those instances have a cinder volume attached. The heat stack cannot be deleted, as it cannot detach cinder volumes when the instance status in Horizon is in the Error state. When a Reset for the instance is issued, it changes the status from Error to Active. |
Reset any instances that are part of the application prior to deleting the application. |
|
SC-3005
|
After a YUM update on compute nodes, manager.py must be repatched. |
After a YUM update on compute nodes, the manager.py is also updated. The YUM update overwrites all manager.py updates. |
Repatch manager.py after a YUM update. |
| SC-3065 |
The number of threads in the Max Enabled Hypervisor report may be incorrect. |
The number of threads in the Max Enabled Hypervisor report may be reported as one when it is actually two or more. |
None. |
| SC-3290 |
Restarting libvirtd on a KVM-AX compute node leaves KVM-FT hypervisors in the Disabled:Up state. |
When you restart the libvirtd service, the openstack-nova-compute service may be left in an inconsistent state. The KVM-FT hypervisor requires both services to be running properly. |
If you restart the libvirtd service on a compute node, you must also restart the openstack-nova-compute service. |
|
SC-3441
|
You may be unable to upload images in Workload Services. |
If you upload an image in Workload Services, the upload may fail and the user interface may appear to hang until the failure is detected. After a few minutes, the Add Image page is displayed, but no image is uploaded. |
To prevent problems, upload images only in OpenStack Horizon. |
| SC-3685 |
If you delete an instance that is part of a failed application deployment, the deletion may fail. |
When an application fails, the Heat orchestrator may be unable to change the state of the instances in the application. You need to delete the application from the application level. |
If an application fails upon deployment, delete the whole application instead of the individual instances. Use the advisories to determine the cause of the failure, correct any problems in the deployment package, and redeploy the application. |
|
SC-3754
|
On the Networks page in the Connections tab, when you create or edit a connection and add a Network Definition, the setting of the Is All Instances field has no effect. |
The setting of the Is All Instances field has no effect. |
You must select either YES or NO because the field is required, but your choice will not affect the network connection. |
| SC-3812 |
KVM-FT instances with Cinder boot volumes may inadvertently reboot when you deploy another KVM-FT instance with a Cinder boot volume if there is not enough network bandwidth in the storage network. |
Although a 1Gb storage network is supported, a 10Gb storage network is recommended for better performance and reliability. |
If you need to deploy multiple KVM-FT instances that rely on Cinder storage, implement a 10Gb storage network. |
|
SC-3839
|
When an application has two networks, and you enable a second network interface in one of the application instances, the second interface takes over the default route and the floating IP address for the application stops working. |
Typically, only one network interface in the system can use the default gateway. If you enable a second network interface, DHCP may automatically assign an IP address with the same network and gateway, which causes connectivity issues for the first interface. |
Use static network settings for the second network interface. Assign an IP address on a different subnet to prevent the second interface from overriding the default route. |
| SC-3853 |
Enabling checksum offload for network interfaces on your KVM or KVM-FT compute nodes can cause data corruption in your instances. |
Checksum offload (for example, the TCP offload engine (TOE)) improves performance by offloading checksum verification from the system processors to the network interfaces; however, it can also lead to data corruption. |
The Availability Services installation script automatically disables checksum offload on KVM-FT compute nodes. Stratus also recommends that you manually disable checksum offload on other KVM compute nodes and OpenStack component nodes, as described in Disabling Checksum Offload on OpenStack Component Nodes.
|
| SC-4012 |
Workload Services reports an Unable to connect to nova service advisory with no apparent cause.
|
Workload Services may temporarily lose connectivity with the OpenStack Nova service because of changes in your cloud environment, for example, if a managed compute node goes down. Workload Services typically restores the Nova connection within a few minutes, and the advisory is dismissed from the Open advisory list. |
If Workload Services cannot connect to the Nova service, wait a few minutes for the connection to be restored. If the problem persists, ensure that your OpenStack controller is running, that the Nova service is running on the controller, and that the network is functioning properly. |
|
SC-4048
|
When an instance is evacuated to another compute node, the instance fails to recover and Workload Services reports that the instance is in a Failed state. |
The instance fails to recover and the OpenStack Nova compute log shows the error, VirtualInterfaceCreateException: Virtual Interface Creation Failed. |
A patch is available to correct this issue in the OpenStack IceHouse release. Contact your service representative for more information. |
|
SC-4061
|
The Deployed Applications page of Workload Services may not list all of the deployed applications, and may fail to display the Show More button. |
The list may be incomplete because the elastic search database that stores information about deployed applications may be in an inconsistent state.
Database maintenance occurs on a nightly basis, but you can manually initiate the cleanup and re-indexing process to resolve the issue sooner.
|
Manually initiate database maintenance by loading the following URL:
http://appliance/api/system/syncronizesearch
After the web page finishes loading and displays Syncronized All Search Indexes, reload the Deployed Applications page.
|
|
SC-4093
|
You cannot change the flavor of an instance that has already been deployed in Workload Services. |
Changing the flavor of a deployed instance (for example, to select a flavor with a larger disk or more memory) may result in an instance migrating to a compute node with an incompatible Availability Level. |
If you need to change the flavor of an instance, edit the image in the original deployment package, select a new flavor, and then redeploy the application. |
|
SC-4212
|
When you deploy an application in Workload Services, the application appears on the Deployed Applications page, but then the instances switch to the Deleting state and the deployment stalls.
|
The application orchestrator cannot successfully create the application stack because of a transient error in OpenStack. OpenStack attempts to recover by automatically deleting the instances, but this fails because the application is in an inconsistent state. The instances remain in the Deleting state.
|
In OpenStack Horizon, under the Default Tenant, select Project, click Orchestration, and click Stacks. Delete the Failed stack associated with the stalled deployment, and then go back to Workload Services to redeploy the application. |
|
SC-4213
|
When you create a new user account in Workload Services, and the user clicks the initialize password link in email, he or she cannot set a password on the Reset Password page.
|
Workload Services sends a first login email to the new user. If he or she clicks the link in the email to set a password, the Reset Password page may appear with an error message. The user is unable to set a password. |
The user must close the Reset Password page and click the initialize password link from the email to reload the page. |
|
SC-4216
|
When you create a Network Topology and open the Create New Environment dialog box to specify Connection rules, you cannot add rules with the Protocol set to ANY.
|
OpenStack does not support setting the Protocol to ANY. |
In the Create Environment dialog box, if you need to specify Connection rules, you must create a separate rule for each Protocol needed (TCP, UDP, ICMP). |
|
SC-4221
|
If you deploy an application that includes a network Environment with Internal Connection rules where the Protocol is set to ANY, the application stalls in the Creating state. |
OpenStack does not support setting the Protocol to ANY. |
In OpenStack Horizon, under the Default Tenant, select Project, click Orchestration, and click Stacks. Delete the Application stack and then the Environment stacks associated with the stalled application.
In Workload Services, delete the application (and ignore warnings that will be displayed about the instances missing). Edit the network Environment to replace Connection rules that contain ANY with separate rules for each Protocol needed, and then redeploy the application.
|
| KVMFT-12 |
If a KVM-FT compute node crashes or restarts, one of the KVM-FT instances restarts in the Error state while Workload Services reports an advisory indicating that an instance was not found on Ax n (where Ax n is a KVM-FT compute node).
|
If Availability Services cannot rebuild a KVM-FT instance, the instance is unable to start normally; however, the remaining KVM-FT instance in the fault-tolerant pair continues to start and run your application. |
On the Deployed Applications page, carefully locate and delete the only the KVM-FT instance that was not found. Availability Services automatically rebuilds the deleted instance. Verify that the paired KVM-FT instances synchronize and return to the Running and Paired states.
|
| KVMFT-38 |
A Network File System (NFS) share is not supported for the /var/lib/nova/instances directory on KVM-FT compute nodes. |
KVM-FT instances cannot run on NFS shares because KVM-FT requires higher I/O throughput and reliability than NFS can provide. |
None. |
|
KVMFT-67
|
If you disassociate or change the floating IP address of a KVM-FT instance in OpenStack Horizon, the original floating IP address remains listed for the instance in Workload Services. |
Workload Services may be unable to discover changes to the floating IP address of a KVM-FT instance in OpenStack Horizon. |
If you change the floating IP address of a KVM-FT instance in OpenStack Horizon, you can view the changes only in Horizon. |
|
KVMFT-68
|
If you try to create a file system on a Cinder data volume connected to a Linux-based instance, the mkfs command may display a warning message about overwriting existing data on the volume even if the volume is new. |
Unless you configure a Cinder node to securely wipe data when you delete volumes, some deleted data remains on the physical disk(s) until it is overwritten. If you allocate a new Cinder volume that uses some or all of the same physical disk space, the Linux mkfs command may detect the deleted data and warn you about overwriting it. |
If you are certain that you specified the correct Cinder volume, and you are not overwriting production data, you can proceed to create the file system. |