Search <book_title>...

Veritas Flex Appliance Getting Started and Administration Guide

Last Published: 2021-06-11

Product(s): Appliances (2.0.1)

Platform: Flex Appliance OS

Operational notes

This topic explains important aspects of Flex Appliance 2.0.1 operations that may not be documented elsewhere in the documentation.

Software operational notes

The following list contains the notes and the known issues that apply for the Flex Appliance 2.0.1 software:

After an update to this release, an issue can occur that causes multiple email alerts to be sent for an unsupported Fibre Channel connection. If you see this issue, you can disregard the emails. The email alerts should stop after 24 hours.
For this release, while an update is in progress, do not perform any other operations in the Flex Appliance Shell or the Flex Appliance Console.
If you change the BIOS date and time on a Veritas 5250 Appliance after initial configuration, you can no longer access the Flex Appliance Console. To restore access, change the BIOS date and time back to what it was before.
When you create a bond, do not give it a bond name that is all numbers. If you do so, you can no longer create any additional bonds on the appliance.
If you were not aware of this limitation and have already created a bond name that is all numbers, contact Technical Support for assistance. Ask your representative to reference article 100050464.
Flex Appliance 2.0.1 has been tested for up to 6,000 storage devices or paths in the case of multipathing. If you zone more than the tested number of devices or paths, you may have to take additional steps to perform the following operations and avoid the associated issues:
- Initial configuration, adding a node, and Flex Appliance upgrades
  These operations may take a long time or fail.
- Factory reset
  During a factory reset, the Veritas Remote Management Interface may go blank and may not accept input.
You should consult with Veritas Technical Support before you implement a Fibre Channel configuration of that size.
Accelerator backups on Flex Appliance application instances fail if there is not enough space for the logs. For example, the logs for NDMP backups could require up to a Terabyte of storage space.
On a master server instance, the logs are stored in the same location as the NetBackup catalogs. To resolve this issue on a master server instance, resize the NetBackup catalog and make sure to allocate enough storage for the Accelerator logs as well.
On a media server instance, the logs are stored in the /mnt/nblogs directory, which has a default size of 250 GB. This directory cannot currently be resized through the Flex Appliance Console. To resolve this issue on a media server instance, contact Veritas Support and ask your representative to reference article 100049971 to resize the directory.
Accelerator backups on Flex Appliance application instances fail if there is not enough space for the logs. For example, the logs for NDMP backups could require up to a Terabyte of storage space.
On a master server instance, the logs are stored in the same location as the NetBackup catalogs. To resolve this issue on a master server instance, resize the NetBackup catalog and make sure to allocate enough storage for the Accelerator logs as well.
On a media server instance, the logs are stored in the /mnt/nblogs directory, which has a default size of 250 GB. This directory cannot currently be resized through the Flex Appliance Console. To resolve this issue on a media server instance, contact Veritas Support and ask your representative to reference article 100049971 to resize the directory.
Due to an underlying performance issue, the monitoring and alerting feature for the ROC Temperature (RAID on Chip for RAID Adapters) is disabled in this release for the Veritas 5250 Appliance. This feature will be reinstated and available for 5250 models in a future software release.
If you update the appliance and then roll back, the Activity Monitor shows an "Update appliance and reboot" task after the rollback. The task shows 90% progress and eventually fails with a "timed out" error. This task shows in error and can be ignored.
During a rollback after an instance upgrade, the Activity Monitor may erroneously show Clear state tasks with the following error message:
"Failed to clear the fault on <instance_ID>_<version>_appVols on <node_name>. Check /log/nodeworker/worker.log and VCS logs for more details."
If you see this error during a rollback, it can be safely ignored. Continue to monitor the task to make sure that the rollback completes successfully.
When you install application add-ons on an instance, the Flex Appliance Console lets you select different versions of the same OST plug-in. However, this configuration is not supported, and if you select more than one version of the same plug-in, the Apply add-ons page shows duplicate entries. Only install one version of each OST plug-in on an instance. If you need to change the version of an OST plug-in that is already installed, first uninstall it, and then install the new version.
In certain scenarios, the Fibre Channel page may briefly appear in the Flex Appliance Console on the Veritas 5150 Appliance. Fibre Channel is not currently supported on the 5150 appliance. If you see the page appear and it does not go away on its own, clear cookies on your browser, then sign out and sign back in.
When you power cycle or restart an appliance, the Flex Appliance Shell becomes available when the restart is complete. The Flex Appliance Console requires an additional three minutes before you can try to sign in.
If Fibre Channel devices are connected to the appliance but you do not have any application instances, "Abort command" messages display in the /var/log/messages log. The messages are similar to the following:
Oct 8 11:13:56 <hostname> kernel: qla2xxx [0000:af:00.0]-8804:9: Abort command mbx success cmd=ffff890ecc0a47c0.
If you see these messages, they can be ignored. After you create an application instance and assign the Fibre Channel ports, they should no longer display.
When you add a file to the repository on the Flex Appliance Console, a delay occurs after the progress bar completes before you are redirected to the Activity Monitor. Wait to be redirected, and then you can monitor the status of the task.
If you upgrade both nodes of a multi-node appliance and then roll back one of them, the Flex Appliance Console still shows the upgraded version on both nodes. The console does not show the correct version until you roll back the second node.
If the host0 or host1 port is not connected to the appliance node during initial configuration, the following error message appears that does not provide complete information:
"Network card for <interface name> is missing. Make sure that all network interfaces are connected to the appliance."
If you encounter this message, verify that all ports are connected according to the initial configuration guidelines. See Initial configuration guidelines and checklist. Then restart the node to continue the configuration.
When you create a new application instance, the Application instances section of the System topology page may show the instance status as Partially Deleted while the creation is in progress. The Partially Deleted status displays in error and can be safely ignored. You can track the instance creation progress from the Activity Monitor, and the instance status changes to Online when the instance creation has completed successfully.
If an appliance node is powered off unexpectedly and then turned back on, an issue can occur that causes the Flex Appliance Console and all application instances to stop working. The Flex Appliance Shell displays the following error when you attempt to log in:
Installation status: failed
If you encounter this issue, contact Veritas Technical Support for assistance. Ask your representative to reference article 100046737.
The following error message may display in the Flex Appliance Shell during an upgrade or a factory reset:
```
dracut:Failed to install module bnx2
```
This message displays in error and can be safely disregarded.
If you upgrade a node and elect not to restart it when prompted, the node must be restarted with the 'restart' command to complete the upgrade. If the node is restarted incorrectly due to a power outage, pressing the power button, or a similar scenario, the remaining upgrade steps fail.
If you encounter this issue, roll back the node to the previous version and then re-run the upgrade.
If you have a multi-node appliance and one of the nodes is turned off for any reason, do not restart the other node until they are both online. If a node is restarted while the other node is off, the Flex Appliance Console fails to load.
If you encounter this scenario, contact Veritas Technical Support and ask your representative to reference article 100046118.
When you create an instance, if you enter the IP address before you select a network interface, the IP address field displays the following error message:
"IP address does not belong to the selected network's subnet."
This message still displays after you select the network interface that corresponds to the IP address. To clear the message, click inside the IP address field and then click or tab outside of it.
After some operations in the Flex Appliance Shell, an "Operation successfully" message may display even if a failure occurred. Read all of the messages that display at the end of each task to make sure that no further action is required.
It is not currently possible to edit a network bond after it has been created. If you need to edit a bond, you must delete it and then create a new bond with the new settings.
See Deleting a network bond.
See Creating a network bond.
If you upgrade an application instance that has an OpenStorage (OST) plug-in installed on it, you must reinstall the plug-in after the upgrade is committed successfully.
When you add a large file to the repository, the following issues can occur:
- If the upload takes longer than 15 minutes due to slow internet speeds, the upload may fail. If you experience this issue, improve your network connection and try again. If the issue persists, you can contact Veritas Technical Support to have a representative manually copy the file to the appliance repository. Ask your representative to reference article 100048789.
- A "Failed to unmarshal JSON" error message may appear when you sign in to the Flex Appliance Console after the upload. If you experience this issue, close the current tab and open the console in a new window or tab.
After a warning message displays to let you know that your password is about to expire, you can no longer sign out of the Flex Appliance Console because an "invalid token" error occurs. In addition, if you refresh the page, you are redirected to the sign-in page. To avoid this issue, change your password as soon as the warning message displays.
If you upgrade an instance that supports rollback, the rollback may lead to inconsistencies between the NetBackup catalog and the media servers for all jobs that ran after the upgrade. These inconsistencies can affect future backups.
See Warnings and considerations for instance rollbacks.
If the VLAN tags do not match between the nodes of a multi-node appliance, the following error appears when you attempt to relocate an instance:
Failed to clear the fault on <network_interface> on <node_hostname>. Check /log/hostagent.log and VCS logs for more details.
However, these logs do not include any additional details. If you encounter this issue, check the VLAN tags on the appliance from the Settings > Network page. Verify that all of the VLAN tags have been added on both nodes.