Troubleshooting. Dell Avamar Plug-in for Hyper-V VSS, Avamar Server


Add to my manuals
94 Pages

advertisement

Troubleshooting. Dell Avamar Plug-in for Hyper-V VSS, Avamar Server | Manualzz

D

Troubleshooting

Topics:

Backup problems and solutions

Restore problems and solutions

Backup problems and solutions

The following topics provide troubleshooting information for issues that can occur during backups with the Hyper-V VSS plug-in.

Browse request fails for cluster client

A Client refused browse request error might appear if you try to browse a cluster client for data to back up.

The error appears in the following situations:

● The Avamar cluster client is not functioning. Start the cluster client, and then continue with the backup.

● The Avamar cluster client is migrating from one cluster node to another. If this step is the case, then wait for the migration to complete. Try to browse for data again.

● Another client is using the same IP address as the Avamar cluster client. This issue can occur if you specify the incorrect IP address for the cluster client during configuration. The issue can also occur if you change the cluster client name without first correctly deleting the registration for the client on the Avamar server. To resolve this issue, delete the client with the incorrect IP address from the Avamar server.

Paused virtual machine is in a Saved State after backup

If a virtual machine is paused when a backup occurs, then the state changes to a Saved State after the backup. You can reset the state of the virtual machine after the backup.

Virtual machines are skipped during backup

If a virtual machine is in a Cannot connect to virtual machine configuration storage state, then the virtual machine is skipped during the backup. Resolve the issue that prevents the virtual machine from connecting to virtual machine configuration storage, and then retry the backup.

Improving backup performance in Windows Hyper-V Server clusters with CSV

Use the following strategies to improve performance when you use an Avamar proxy cluster client for image-level backups of

Hyper-V data on CSVs in a Windows Hyper-V Server cluster:

● Maximize network bandwidth between the cluster nodes.

● Add one or more secondary proxy nodes to perform parallel backups and balance the backup load. Then allow the backup process to automatically change CSV ownership during the backup to the proxy node that is performing the backup.

● Consider using a dedicated cluster node that does not actively host and manage virtual machines as the primary proxy node.

You can also use dedicated nodes for the secondary proxy nodes.

● When a proxy node is not the owner node of the CSV that it is backing up, the backup process must stream backup data from the owner node to the proxy node before sending the backup to the Avamar server or a Data Domain system. As a

Troubleshooting 83

result, backup performance depends on network performance between cluster nodes. Backups are faster and more efficient when a proxy node owns the CSV and backup data streams directly from the proxy node to backup storage.

If the environment includes a single proxy node and you cannot add proxy nodes, you can manually move CSV ownership to the proxy node. However, review the following considerations:

● Changing the ownership of a CSV to a single cluster node may negatively impact the performance of all virtual machines on the CSV.

● The proxy node should be a robust physical node.

● The CSV owner node is responsible for file system metadata updates to the NTFS partition. The proxy node must have the capacity to accommodate these updates for the CSVs.

● Ensure that the CSV is in a healthy state before you move ownership.

● The CSV must be online and in full access mode when you move ownership.

The Microsoft cluster documentation provides instructions for moving ownership of CSVs.

Avamar virtual client is registered to a different Avamar server and backups fail

Standard Avamar plug-ins can be re-activated to a different server without uninstalling the clients first. To re-register the virtual cluster client to a different server, uninstall the virtual client. Then rerun the Cluster Configuration Tool and select the new server name.

Backup of Windows Hyper-V Server with SMB and non-SMB target

VMs is not supported

Due to an known issue with Microsoft Hyper-V, backup of SMB and non-SMB VMs fails. Microsoft KB article 2853422 contains information about a hotfix for this problem.

Hyper-V federated backup and recovery in a multiple IP or multiple

NIC environment

When running a federated (CSV or SMB) backup or recovery in a multiple IP or multi-NIC environment, specify the node information for the Hyper-V cluster:

1. Open the avhypervvss.cmd

file for the cluster client in a text editor. The avhypervvss.cmd

file is located in the cluster client var folder that you specified in the Cluster Configuration Tool.

2. Add the following flags to the avhypervvss.cmd

file:

--clusternode= Clusternode1(IP address)

--clusternode= Clusternode2(IP address)

3. Save and close the file.

Virtual machine file path or name with special characters

● If the path of a virtual machine file contains an ampersand (&) or semi-colon (;), backup of the virtual machine fails

● If the name of a virtual machine contains a less than (< ), greater than (>), colon (:), double quote (""), forward slash (/), back slash (\), pipe (|), question mark (?), or an asterisk (*), backup of the virtual machine fails

In the case of a failure, the following message is displayed:

Virtual machine 'spl_chars?*.,^$#@!%(){}[]';<F38BEE0A-1E3C-4127-8DCE-0B1D73444E82>' has one or more unsupported characters '<>:"/\|?*' in the Virtual Machine name or

'&;' in the Virtual Machine file paths and will not be backed up with image level backup

● If the name of a virtual machine ends with ..

, GLR browse fails with the following error:

84 Troubleshooting

Failed to browse the directory '\\?\Volume{c90bf39d-984a-11e7-80fd-3c08f66088b1}\*', no file were found

Mixed cluster environments not supported

Mixed environments of clustered Hyper-V server 2016 and later and 2012 R2 are not supported.

Backups with mixed absolute and relative paths will fail

Backups that contain VMs that have both absolute and relative paths will fail.

When selecting VMs for backups, if one or more VM uses an absolute path, and one or more VM uses a relative path (for example, one VM uses the absolute path C:\VM-Data-01\Fixed-VM1 and another VM uses the relative path Fixed-VM2 ), the backup will fail with an error similar to the following: avtar Error <8012>: Error in path list: Path arguments must all be absolute or all relative (as indicated by first argument), but not both.

To workaround this issue, make sure you always use absolute paths for VMs, or always use relative paths.

Incremental backups

If you use the flag --enable-incremental=false after running few incremental backups, then you must manually merge the recovery checkpoints that were created during the previous incremental backups before starting a new backup with this flag.

Backups failing when Data Domains are switched between consecutive backups

When you switch to another Data Domain during a subsequent backup, avtar is not able to locate backup history and gives an error avtar Error <5195>: Path ".system_info\HyperVBackupHistoryInfo.xml" not found in backup. (Exception looking up '.system_info\HyperVBackupHistoryInfo.xml') . To mitigate this error add the following flag to avhypervvss.cmd

file in primary var folder:

--force-backup-history-reset=true

Once the current backup passes after addition of this flag, remove the flag and proceed normally.

Hyper-V federated backups failing with error-failed to get subprocess summary for work-orders

The federated backup fails, if primary client fails to get the summary message from the secondary nodes. If backup fails to get subprocess summary for the work order, you must add --summary-retry-count flag in the avhypervvss.cmd in cluster client's var directory. The values of the flag ranges between integer 3 to 1000, whereas the default value of the flag is 3. The flag defines the number of retries by the primary client for the summary message that is sent by the secondary nodes.

RCT backups failed or completed with exception with error-failed to get the RCT extents for virtual machine

If RCT Incremental backups failed or completed with exception with error-failed to get the RCT extents for virtual machine, then you can add --promote-to-full-on-extents-error flag to true in avhypervvss.cmd in primary var folder/ directory .

Troubleshooting 85

The RCT backup for the failed or skipped virtual machine is promoted to full if the --promote-to-full-on-extentserror flag is set. Also, the backup window is increased because of the promotion that takes place from incremental backup to full backup.

Restore problems and solutions

The following topics provide troubleshooting information for issues that can occur during image-level restores with the Hyper-V

VSS plug-in.

Browse request fails for a cluster client

A Client refused browse request error might appear if you try to browse a cluster client for data to restore.

The error appears in the following situations:

● The Avamar cluster client is not functioning. Start the cluster client, and then continue with the backup.

● The Avamar cluster client is migrating from one cluster node to another. If this is the case, then wait for the migration to complete. Try to browse for data again.

● Another client is using the same IP address as the Avamar cluster client. This issue can occur if you specify the incorrect IP address for the cluster client during configuration. This issue can also occur if you change the cluster client name without first correctly deleting the registration for the client on the Avamar server. To resolve this issue, delete the client with the incorrect IP address from the Avamar server.

Access denied error when restoring a cluster

If a cluster backup is being restored to the same or a different cluster, and an access denied error is seen in the log file, ensure that the HyperV server that is initiating the restore has been added to the Backup Operator group on all the HyperV servers of the cluster.

SMB file share requirements

contains further information about the Backup Operator group.

Virtual machine fails to start after redirected restore

If you restore a virtual machine to a different Hyper-V Server and you cannot start the virtual machine because it remains in a

Saved State, delete the Saved State files and restart the virtual machine.

This issue can occur when the Integration Components (IC) version for the backup Hyper-V Server and the guest virtual machine do not match. To prevent this issue, ensure that the IC versions on the Hyper-V Server and virtual machine match.

Virtual machine configuration file is hidden after restore

If you specify the default location when you create a virtual machine, and then you restore the virtual machine as files for manual provisioning, the configuration XML file for the virtual machine restores as a hidden file.

To view the file, select the option in Windows Explorer to show hidden files.

Restored virtual machine fails to start after live migration

After you restore a virtual machine with a snapshot to a different Hyper-V Server on a CSV, you can start the virtual machine by applying the snapshot from the CSV node that runs the master agent. The virtual machine does not start if you apply the snapshot after a live migration of the virtual machine to another CSV node.

In the virtual machine settings, update the network adapter to a valid adapter in the host, and restart the virtual machine.

If you cannot perform the live migration after restoring the virtual machine to a different Hyper-V Server on a CSV, turn off the virtual machine before you perform the live migration.

86 Troubleshooting

Conflict between Hyper-V host and guest backup application

If you install backup applications on a Hyper-V virtual machine and perform image-level backups with the Hyper-V VSS plug-in, then the two backups may conflict.

To work around this issue, exclude virtual machines with backup applications from image-level backups.

Restore takes a long time when restoring virtual machines on multiple CSVs in an environment with multiple proxies

When you restore multiple virtual machines on multiple CSVs in an environment with multiple proxies, the restore might take a long time.

To improve restore performance, ensure that the primary proxy node is the owner of the CSV to which you restore the virtual machines.

Redirected restore of VMs shown as failed in Avamar Administrator

When a backup is attempted immediately after one or more VHDs attached to the VM are removed from the VM configuration, redirected restore of VMs will be shown as failed in Avamar Administrator. However, the restore will then be successful.

When one or more VHDs are removed from the VM configuration, the VHD file list corresponding to the removed VHDs is not immediately excluded from the Hyper-V VSS Writer metadata after the removal operation. Because of this option, the VHD files for the removed disk are still being backed up by the Hyper-V plug-in. These VHD files are not part of any active VHD chain that is attached to the VM at the time of backup, resulting in subsequent restores being reported as failed.

The solution to this problem is to ignore the reported assertions, verify that the VM starts correctly after the restore operation, and restart Hyper-V VSS Writer before next backup

Troubleshooting 87

advertisement

Was this manual useful for you? Yes No
Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Related manuals

advertisement

Table of contents