Known issues and limitations¶
Version 23.08¶
Upgrade notes¶
See zCompute Software Life-Cycle for the Zadara life-cycle policy, supported upgrade paths and details for scheduling an upgrade.
Key pairs - limitations¶
For Windows password recovery (Get Windows Password dialog), the UI only accepts keys in RSA (PEM) format.
In version 23.08.0 and earlier, zCompute does not accept keys for VMs in SSH formats, but only in RSA (PEM) format.
Supported DNS services¶
zCompute supports DNS services at VPC and Private Hosted Zone levels.
zCompute does not provide public-facing DNS hosted zones.
Custom instance types¶
New custom instance types can only be created after receiving Zadara approval, and after getting pricing.
The available instance types, sizes and pricing are calculated for optimal use of Zadara’s hardware resources.
V2Z custom instance types - known limitation¶
In the V2Z utility, custom instance types are not allowed for most users.
NK-5234 Extented volume size not updated on instance volume¶
Protection groups snapshots and start date¶
Protection group configuration may start creating snapshots only 24 hours after creation.
The protection group Start Date is set to the following day.
GRE via NAT gateways¶
GRE via Network Address Translation (NAT) gateways is not supported.
OpenOTP users don’t appear in the UI¶
Although it is possible to sign-on with an OpenOTP user, the OpenOTP users are not displayed in the UI (Users tab).
Version 22.09¶
Upgrade notes¶
This version upgrades from the official zCompute releases:
V22.02.2 (22.02-sp2)
V22.02.3 (22.02-sp3)
Customers that have a zCompute version prior to the two upgradable versions will be upgraded in two steps, which may require a longer maintenance window. Relevant customers will be contacted by Zadara Operations.
During the maintenance window, customers may experience sporadic API request failures or brief instances of network disruptions.
NK-6930 UI allows creation of instance types which aren’t available on the cloud¶
In zCompute v22.09.x, the UI might display instance types that require a CPU model which is not available on that cloud. For example, Premium instance types on a non-premium cluster, such as zp4.* instance types on a cluster with standard nodes only.
It is possible to create a VM with such an instance type, but the VM would fail to start.
NK-4852 Identity manager hangs because of too many warning messages¶
On deletion of an account, the identity manager service was hanging during the user update process, while issuing warning messages for each of the users in the deleted account. The identity manager could not restart successfully. As an interim workaround, in zCompute release 22.09.2, these warning messages are converted to info-level messages. This issue will be fixed in a future release.
NK-3335 UI exception when attaching security groups to a load balancer that is not ready to use¶
The load balancer service may take time to be initiated. Users will experience a UI exception while trying to attach a security group on a load balancer in the spawning state. Users can attach the security group after the load balancer is in the available state.
NK-4637: Only the ‘Reset password’ email is sent successfully¶
It is impossible to receive SMTP for events from zCompute, with one exception: Only the ‘Reset password’ email is sent successfully. This means that setting custom alerts will not send emails, but will only display in the GUI.
DVS - known issues¶
NK-3898: On DVS projects, users are not able to set project level tagging that later propagates to virtual machines. Workaround: Once the VM has been created, users can navigate to the VM and add the tag.
NK-4598: The DVS instance tab shows the key pair option while restoring the snapshot backup, even though key pairs are not supported in this version.
NK-4606: The DVS networks overview displays a warning message to clear the filters, even though no filter is applied. Click
Clear Filtersto display the overview.
Volumes that are not connected to any VM - known issue¶
NK-4242: ‘Default volumes’ are impossible to delete. We are experiencing difficulties in deleting what appear to be “non attached volumes”. Although the UI does not show the attachment, the volumes are still attached to the host.
Enhanced remote DNS status validation - known limitations¶
The DNS status for VPCs without a valid route to an IGW will always be “Degraded” when the default DHCP option with the google DNS is used.
If tenants set their own name server in the VPC’s DHCP options, and their name server is unreachable via the VPC gateway or it can’t resolve google.com, the VPC state will degrade.
Version 22.02¶
NK-3883 Cannot modify ASG max capacity via UI¶
When attempting to modify the ASG max capacity or desired capacity using the zCompute UI, validation begins on the first entered digit and can throw an error. For example, on changing the value from 20 to 30, on entering the “3” the UI displays an error indicating that the new max capacity value is lower than the current value (3 is less than 20). For example:
This issue occurs only when modifying the max or desired capacity in the UI, but the symp CLI is not affected. If the max or desired capacity is updated via the CLI, then the UI will display the correct updated value. For example:
NK-2557 Data on NVMe disks is not deleted before VM start¶
Currently these instance types are only suitable for private clouds.
NK-1106 Data volume should be deleted explicitly by users¶
zCompute does not offer a checkbox to delete VM data volumes. The VM delete dialog clearly states that data volumes must be deleted by users.
NK-3835 zCompute Toolbox Fedora shows the old version for members in the Marketplace¶
This means that tenants will still continue to download the older version if using the Marketplace. Instead look for a pre-downloaded image in the image list. If it’s not there, please ask Zadara ops or your cloud administrator to download it for you and make it publicly available in the system.
NK-4055 zCompute Toolbox Fedora is missing the Symp offline pip dependencies¶
This means that the python dependencies installation requires you to be online (internet) when running symp-update.
NK-2921 Getting Action Failed status creating listener in ALB¶
Getting an Action Failed status when trying to create a listener in the ALB on an cluster upgraded to version 22.02-SP1, but with the previous LBAAS engine.
Workaround: Upgrade the LBAAS engine before creating a new load balancer.
NK-3247 zCompute load balancer service officially supports Terraform 2.47¶
Terraform AWS Provider versions higher than 3.33 fail on LBAAS ARN validation.
Important
The new action types are applicable only after a cluster is upgraded to version 22.02.1, and then the load balancer engine is upgraded.
After your zCompute account has been upgraded, all new load balancer instances will be created using the new engine image. The upgrade of already running load balancer instances is opt in. Users should upgrade their existing load balancer in order to receive the new functionality.
The new load balancer target-group types (fixed-response, redirect) in older images are not supported and require upgrade.
NK-3786: Terraform not supported for custom policies¶
Terraform is not supported at the moment, since it uses a the AWS CreatePolicyVersion API, which is not supported.
Creation of new policies via APIs only (no GUI)¶
Creation of a new policy is only available via APIs and not from the GUI. Once the new policy has been created, the admin can apply it to users using the GUI.
Creation of direct subnet with desired MTU not possible in GUI¶
It is not possible to create a direct subnet in the GUI with desired MTU. However, the MTU can be updated post-creation.
Version 21.10¶
NK-2476 Download image from URL with Let’s Encrypt certificate will fail¶
The trusted certificate store holds a recently expired Let’s Encrypt root certificate. This may be amended per cloud, per customer request, and will be fixed on the next release for all the clouds.
NK-2169 When using Cloud Base Init, user can’t retrieve the Windows Admin password for remote connection¶
This is caused by an internal race in the update of the metadata service.
Workaround:
Either don’t use Cloud Base Init to create images, or add a startup script that will accept the initial admin password as user data and set it for the user.
NK-1873 DR: In order to restore a VM the user has to have admin role on the destination project¶
A user must have an admin or an MSP admin role in order to create a local snapshot from a remote snapshot. A bug has been created where the user needs to have an admin role in the destination project in order to restore a VM to that project. Ideally, the admin should be able to restore into a selected project without having a role on the project.
NK-1903 DR: No indication to user when a remote snapshot is ready¶
The DR metadata is written immediately to the NFS share while the actual snapshot mirror depends on when the VPSA has finished the transfer. The result is that the user may see a snapshot as ready on the remote site while a restore is not possible yet. Trying to restore the VM will lead to volume error which can be force deleted.
Workaround: Restore from snapshot N-1 or retry later if the original system is still active.
NK-1788 Some attributes are missing when creating an AMI image from a restored VM¶
When creating an image from a VM that was restored from a VM snapshot, the new image is missing some attributes. Specifically, it is missing the bus_type and disk type. The implication is that any VM created from the image VM won’t boot.
Workaround: Do not create images from a restored VM. Create the image from the volume snapshots.
NK-753 Tenant admin unable to see resources of all projects in the account¶
Tenant admins can create projects and assign users, but are unable to see the resources of these projects unless they give themselves an admin role on the project. This makes the account summary inaccurate, since only the resources from the current project are counted. In this image, you can see an empty project:
While in this image, at the account level you can only see the default VPC resources from the first project:
NK-1751 DR: Raise an error event when failed to create vm-snapshot¶
Currently if the system fails to create a VM-snapshot during the asynchronous part of the flow, no indication is given to the user.
Known issues and limitations introduced in R7¶
NK- 1582 - Placing a VM and it’s volumes into the same protection group causes some snapshots to go into error state
NK-1285, NK-1282 - Mirror is not cleaned when detaching a volume from a VM or removing an entity from a protection group
Known issues and limitations introduced in R5¶
NK-1076 - Reset password for predefined users in disabled accounts succeeds but the message claims failure
NK-1082 - There is no export to CSV on Monitoring → Events
NK-1115 - Windows VM may get data corruption if rebooted from the system
Known issues and limitations introduced in R4¶
NK-817 - Delete a NAT gateway fails with an error that resources could not be deleted due to dependencies/network associations/routes/etc. User must first delete the route to the NAT GW in the route table.
NK-831 - Deletion of ENI is not possible from the GUI
Known issues and limitations introduced in R3¶
NK-648 - Load Balancer creation menu allows selection of elastic IP that belong to an instance without a warning
NK-661 - Basic event filters do not work. Use the advanced mode instead (Click more filters)
Known issues and limitations introduced in previous releases¶
NK-234 - Direct-network is not listed as route in route-table. When creating a direct subnet, a route is added for local traffic but it is not displayed in the route table.
NK-502 - Restoring snapshot of VM always done to the default pool, even if the VM disks are not on that pool or the user explicitly selects a different pool.