Troubleshooting Ultra High Performance Volume Attachments
This topic covers troubleshooting steps you can take as well as prerequisites to verify for volumes configured for the Ultra High Performance level where either the volume fails to attach or the volume attachment is not multipath-enabled.
Troubleshooting Volume Attachment Failures
The Block Volume Management plugin is required for volumes configured for Ultra High Performance , attached using the iSCSI attachment type. If the volume fails to attach to the instance, the issue is likely caused by incorrect configuration for the Block Volume Management plugin. See the troubleshooting suggestions in this section for these issues.
If you have not configured permissions correctly for the Block Volume Management plugin, the volume will fail to attach to the instance.
Details
The volume will not show up as attached in the Console and you will see a
NotAuthorizedOrNotFound error message in the Block Volume Management plugin log.
The Block Volume Management plugin log is located in:
"/var/log/oracle-cloud-agent/plugins/oci-blockautoconfig/oci-blockautoconfig.logThe following is an sample error log entry for this issue:
2021/08/13 09:14:25.864932 compute_client_command.go:255: Updating volume attachment to the state LOGIN_SUCCEEDED ...
2021/08/13 09:14:26.155473 compute_client_command.go:260: Service error:NotAuthorizedOrNotFound.
volume attachment ocid1.volumeattachment.oc1.iad.<volume-attachment_ID> not found.
http status code: 404. Opc request id: <request_ID>Cause
The Block Volume Management plugin does not have sufficent permissions to send the iSCSI login status notification to the service.
Resolution
To configure permissions for the Block Volume Management plugin:
-
Create Dynamic Group: Create a dynamic group with the matching rules in the following code sample, to include all instances in the specified compartments:
ANY {instance.compartment.id = 'ocid1.tenancy.oc1..<tenancy_ID>', instance.compartment.id = 'ocid1.compartment.oc1..<compartment_OCID>' -
Configure Policy for Dynamic Group: Configure a policy granting permissions to the dynamic group created in the previous step to enable the instance agent access to call the Block Volume service to retrieve the attachment configuration.:
Allow dynamic-group InstantAgent to use instances in tenancy Allow dynamic-group InstantAgent to use volume-attachments in tenancy
Resources
The compute instance must have either a public IP address or a service gateway to be able to connect to Oracle services or the volume will fail to attach.
Details
The volume will not show up as attached in the Console and you will see a
user agent can not be blank error message in the Block Volume Management plugin log.
The Block Volume Management plugin log is located in:
"/var/log/oracle-cloud-agent/plugins/oci-blockautoconfig/oci-blockautoconfig.logThe following is an sample error log entry for this issue:
2021/10/15 22:16:07.881953 compute_client_command.go:255: Updating volume attachment to the state LOGIN_SUCCEEDED ...
2021/10/15 22:16:07.882185 compute_client_command.go:260: user agent can not be blank
2021/10/15 22:16:07.882204 iscsi_commands_helper.go:302: user agent can not be blank
2021/10/15 22:16:07.882212 iscsi_commands_helper.go:310: user agent can not be blankCause
The Block Volume Management plugin cannot send the iSCSI login status notification to the service due to the network configuration.
Resolution
If the instance does not have a public IP address, set up a service gateway on the virtual cloud network (VCN). The service gateway lets your instance privately access Oracle services without exposing the data to the public internet. Here are special notes for setting up the service gateway for the Block Volume Management plugin:
- When creating the service gateway, enable the service label called All <region> Services in Oracle Services Network.
- When setting up routing for the subnet that contains the instance, set up a route rule with Target Type set to Service Gateway, and the Destination Service set to All <region> Services in Oracle Services Network.
For detailed instructions, see Access to Oracle Services: Service Gateway.
Resources
Volume Attachment is not Multipath-Enabled
When you attach a volume configured for the Ultra High Performance level, to achieve the optimal performance, the volume attachment must be multipath-enabled. The Block Volume service attempts to enable the attachment for multipath when the volume is being attached. If not all of the prerequisites have been addressed, the volume attachment will not be multipath-enabled.
To check whether a volume attachment is multipath-enabled in the Console from the Volume Details page:
- Open the navigation menu and click Storage. Under Block Storage, click Block Volumes.
-
Click the block volume that you want to check the volume attachment for.
- Click Attached Instances in the Resources section.
-
Check the value displayed in the Multipath column.
-
Yes: The volume is configured for the Ultra High Performance level and the volume attachment is multipath-enabled. No further action is required.
- No: The volume is not configured for the Ultra High Performance level, the volume does not need to be multipath-enabled. No further action is required.
- No with a warning icon: The volume is configured for the Ultra High Performance level, but the volume attachment is not multipath-enabled. To achieve optimal performance, you need to ensure the volume is attached to a supported instance shape, and that the required prerequisites are configured.
-
If the volume is configured for the Ultra High Performance level, but not configured for multipath as required, the Multipath column will contain No with a warning triangle, as shown for the first row in the following screenshot:
For additional procedures to verify whether a volume is multipath enabled, including using the CLI or API, see Checking if a Volume Attachment is Multipath-Enabled.
If your volume is not configured for multipath, to troubleshoot the issue, review the information covered in this section and address any issues raised.
You must attach a volume configured for the Ultra High Performance level to an instance based on a supported shape, configured for at least 16 cores.
Supported Shapes
All current bare metal shapes support multipath-enabled iSCSI attachments. See for more information Bare Metal Shapes for performance characteristics of block volumes attached to bare metal instances.
Current VM shapes configured for 16 cores or more support multipath-enabled attachments. See VM Shapes for iSCSI and Paravirtualized Attached Volumes for performance characteristics of volumes attached to VMs.
Resolution
If the volume isn't attached to an instance with a supported shape configuration, you need to detach the volume and attach it to an instance with a supported shape configuration.
You can also edit the existing instance so that it has the correct shape configuration, but you need to ensure that you detach and reattach the volume after you edit the instance.
If the instance has less than 8 OCPUs, you might see an issue, where after you edit the instance to support multipath-enabled attachments, the volume attachment is still not multipath-enabled, even after you detach and reattach the volume. In this scenario, you need to re-create the instance from a supported shape, and then attach the volume to the new instance. For more information, see Paravirtualized volume attachment not multipath-enabled after instance is resized.
Resources
You must attach a volume configured for the Ultra High Performance level to an instance running an image that supports multipath attachments. This includes custom images.
Supported Images for iSCI Attachments
Only platform images running Oracle Linux or custom images based on an Oracle Linux image support multipath attachments.
Use one of the latest platform Oracle Linux images, with an Unbreakable Enterprise Kernel (UEK) version of UEK6U1 or higher.
For custom images, the Unbreakable Enterprise Kernel (UEK) version must also be UEK6U1 or higher. The UEK6U1 UEK is associated with the kernel major release version 5.4.17-2036, released in November, 2020. You also need to update the Storage.Iscsi.MultipathDeviceSupported property for the custom image to true and launch the instance again. For more information, see Configuring Image Capabilities for Custom Images
Supported Images for Paravirtualized Attachments
For multipath-enabled attachments, the attached instance must be running one of the following images or a custom image based on one of these images:
- Oracle Linux
- Ubuntu
- CentOS
- Windows
Multipath-enabled attachments are not supported for Oracle Autonomous Linux instances.
Use one of the latest platform Oracle Linux images, with an Unbreakable Enterprise Kernel (UEK) version of UEK6U1 or higher.
Resources
If you have updated the shape configuration or the image for an instance to one that supports multipath attachments, you need to detach the volume from the instance and then attach it again to the instance.
Resources
For iSCSI attachments, if you have taken all the steps described in this topic and you are still encountering an issue with the volume attachment, use the steps described in Step 4: Generate a Diagnostic File for Oracle Cloud Agent to generate a diagnostic file and contact Oracle support. This step does not apply to paravirtualized attachments.
Resources
In certain scenarios, a volume attachment shows as multipath-enabled in the Console, however the attachment isn't actually multipath-enabled, and the volume doesn't achieve the performance expected for the Ultra High Performance level. This issue can occur when you use both the oci-utils and oci-iscsi-config tools at the same time to configure a volume.
Use one of the following methods to check to see if you're encountering this issue.
multipath command to confirm whether a volume attachment is actually multipath-enabled on a Linux instance. Log in to the instance and run the multipath command with the ll tag, as follows:# multipath -ll/var/lib/iscsi/nodes/{IQN} for node.startup, as follows:#cd /var/lib/iscsi/nodes/{IQN}
#grep -Hrn 'node.startup' node.startup=automatic, the volume attachment isn't multipath-enabled. They must all show node.startup=manual.Resolution
/etc/fstab file. Update the /etc/fstab file to tell the systemd service to wait until the multipathd service is running before mounting the file system. To do this, add x-systemd.requires=multipathd.service for the volume. For example:UUID={$AFFECTED_VOLUME_UUID} /test ext4 defaults,_netdev,nofail,x-systemd.requires=multipathd.service 0 2 Reboot the instance after you update the /etc/fstab file.
For more information about the /etc/fstab file, see Traditional fstab Options and fstab Options for Block Volumes Using Consistent Device Paths.