This section contains known issues listed under the following categories:
Data Corruption During UDP Upgrade due to Backward Compatibility
You may notice data corruption for backward compatibility of protecting UDP v5 nodes to UDP v6 RPS, it might cause data corruption on deduplication data store.
This issue does not impact if you are using one of the following options:
This issue impacts if you have:
When do you face this data corruption:
If you are upgrading from Arcserve UDP v5 to Arcserve UDP v6, you may face this issue in the following scenarios:
Symptom
The data corruption issue might happen in the following two scenarios.
Scenario 1: The Agent node is with Arcserve UDP v5 and the RPS node is with Arcserve UDP v6. Agent job fails to back up and generates the incomplete recovery point on the deduplication data store. Deleting the incomplete recovery point from the deduplication data store may lead to data corruption.
Scenario 2: The source RPS node is with Arcserve UDP v5 and the destination RPS node is with Arcserve UDP v6. The Replication job fails and generates the incomplete recovery point destination deduplication data store. Deleting the incomplete recovery point from the destination deduplication data store may lead to data corruption.
Reason
Due to the mismatch of interface between different versions, when Arcserve UDP v6 RPS tries to delete the incomplete sessions generated by Arcserve UDP v5, it may wrongly delete additional index files which leads to data corruption.
Solution
If you are upgrading from Arcserve UDP v5 to Arcserve UDP v6:
Note: Before applying the patch, do not run any backup/replication jobs.
Note: If you opt for this option, do not upgrade to Arcserve UDP v6. You can directly upgrade from Arcserve UDP v5 to Arcserve UDP v6 update 1.
If you have already upgraded from Arcserve UDP v5 to Arcserve UDP v6:
Note: Contact Arcserve Support to get the patch.
How to handle corrupted recovery points?
You cannot recover any recovery point that is corrupted due to the backward compatibility issue as the corresponding index file is incorrectly deleted.
To make sure that the subsequent recovery points do not depend on the corrupted recovery point, we recommend to take again full backup for the impacted node.
In addition, other jobs related to merge job that need to read from the corrupted recovery point might continue failing even though new full backup is done.
To solve the above, we recommend to rename the node folder in the backup destination folder of dedupe data store. Add Backup as prefix to the original node name.
For example:
NODE_NAME[1a98528f-db3c-45de-83d6-9d729815ab7d]
to
BackupNode_Name[1a98528f-db3c-45de-83d6-9d729815ab7d]
As a result, new folder name of the source node is created that automatically converts the recovery point as full when the specific node backs up to the data store again.
For the renamed folder, you can perform the following options:
You can select to delete the renamed node through data store UI, if all the recovery points in the renamed folder are not useful anymore. The automatic retention management of the Plan only applies to the new recovery points done after the renaming of the folder.
If you are using the Windows 10 Edge web browser, you may not be able to log in to Windows or Linux Agents from the Console when you click Restore or Login Agent. Use another browser to log in.
Symptom
If you have two virtual machines, VM1 and VM2, with the same GUID in the ESXi hosts and the two VMs are managed by different vCenter, VC1 and VC2 respectively. You import VM1 into the Console (you cannot import both the VMs because the Console does not allow nodes with same GUID). But you also import VMs from the vCenter VC2. When the auto discovery runs, it firstly connects to VC1 and detects VM1 by the GUID, and the Hypervisor column is updated with information of VC1. Later, when it connects to VC2, it detects the VM2 by the GUID, and the Hypervisor column is updated with information of VC2.
Solution
It is very rare that two VMs have the same GUID. In the worst case, if it happens, Host-Based agentless backup may back up the wrong VM because Arcserve UDP uses GUID to identify a VM. To solve the problem, you can manually change the GUID of one of the VMs. For details on how to do this, refer to related topic in Solution Guide.
Symptom
Unable to log in to the Arcserve UDP Console. The Console displays the following messages even after five minutes of logging in:
Identity Service is starting
Solution
To resolve this issue:
Symptom
Jumpstart job for multiple nodes fails if one of the nodes has a replication job running in progress.
Solution
You can either wait for that replication job to complete before submitting the jumpstart job, or submit jumpstart job for other nodes.
Symptom
Copy to tape job may fail in the following cases:
Solution
For the first issue, add Arcserve UDP agent node to Arcserve UDP console UI with the non-localized user name.
For the second issue, add RPS node to Arcserve UDP console with the non-localized user name.
Symptom
You will see the NAT device setting redundant on the Replicate from a remotely-managed RPS task when you need to modify the plan in the NAT environment, if the plan contains the following tasks:
When you need to modify the plan and view Task 1 configuration, you will find the option Server is behind NAT device redundant when the option Server is behind NAT router configured. The replication job can succeed, if the NAT router information is provided correctly.
Solution
You can ignore the Server is behind NAT device field on the Replicate from a remotely-managed RPS task when you modify plan, which is redundant. Use the Server is behind NAT router field to make sure the replication job succeeds.
Symptom
The registry BLI upper filter of volume class is removed. So, the BLI driver cannot monitor volumes.
Solution
You can continue to perform an Incremental backup after reinstalling the Change Tracking Driver.
Follow these steps:
<installation path>\Arcserve\Unified Data Protection\Engine\BIN\DRIVER
<installation path>\Arcserve\Unified Data Protection\Engine\BIN\DRIVER
Symptom
The following messages can be found in the event log: "Illegal operation attempted on a registry key that has been marked for deletion" or "Windows detected your registry file is still in use by other applications or services. The file will be unloaded now. The applications or services that hold your registry file may not function properly afterwards."
Solution
For cause and solution, see the Microsoft KB article 2287297.
Symptom
Mounting a session to a driver letter fails.
This happens when the destination is a local folder and on a FAT32 volume. Mounting a driver only supports creating the cache file on NTFS volume.
Solution
Add a new registry key and customize the cache file path to a different volume.
Follow these steps:
More Information:
The cache file will be created by mounting a driver when a writable volume is mounted. For Granular Restore catalog/restore job, a writable volume is created. If the session is to back up volumes on Windows 8, Windows 2012, or a later operating system, always mount a writable volume when you mount a session to a driver letter.
Symptom
The backup submitted does successfully show, but from the Arcserve UDP Agent (Windows) UI, you cannot see any job monitor. This is because the backup job meets the maximum concurrent node count setting in the data store. The backup job is put into the waiting queue.
Solution
Open the Arcserve UDP Console, and the pending job monitor will show on the node view.
Symptom
When a Host-Based Agentless Verify backup job is running, and if compression has been enabled in the plan, the Compression Percentage displayed on the job monitor is higher than the actual percentage.
All other backup jobs do not have this problem, including Agent backup jobs and Host-Based Agentless Full/Incremental backup jobs.
Solution
The compression percentage printed in the activity log is correct. Refer to it after the Host-Based Agentless Verify backup job is completed.
Symptom
The backup job of VMware VM finishes successfully on the Arcserve UDP Console. The job is marked by the green icon and activity log shows the message “Backup job completed successfully”. However, the backend process dump file, such as AFBackend.exe.7912.00.dmp, is generated in the BIN folder under the agent installation path. For example: C:\Program Files\Arcserve\Unified Data Protection\Engine\BIN.
Solution
This happens due to a problem of VMware VDDK as occasionally VDDK crashes when Arcserve UDP is closing API calls to VDDK. You can ignore this problem as this happens at the last phase of the backup job. At that point the backup job is almost complete and recovery point has been wrapped up successfully.
"Failed to integrate the language package into the BMR ISO Image".
Symptom
This issue is caused by the third-party anti-virus software (McAfee) filter driver, but it can also happen with other third-party filters.
Solution
Disable your anti-virus software and try the boot kit creation again.
Symptom
When the source machine is an Active Directory server performing a BMR to a physical machine with different hardware or to a virtual machine on a Hyper-V server, the server does not boot and a blue screen displays with the following message:
STOP: c00002e2 Directory Services could not start because of the following error: a device attached to the system is not functioning. Error status: 0xc0000001.
Solution
Reboot the system to the BMR PE environment, rename all *.log files in the C:\Windows\NTDS folder, and restart the system. For example, rename the file edb.log to edb.log.old and restart the system.
Symptom
You may be unable to map the disk/volume during BMR, when the target machine is a VM with IDE disk on a 2008 Hyper-V server or a 2008R2 Hyper-V server.
If you restore data to a VM with IDE disk on a 2008 Hyper-V server or a 2008R2 Hyper-V server using BMR, it will not be able to map the source disk/volume to the target disk/volume, even though the sizes of both disks appear to be the same. This is because that when you create an IDE disk on a 2008 Hyper-V server or a 2008R2 Hyper-V server, the actual disk size is smaller than the size that you specified.
Solution
Create a bigger disk on the VM. For example, if you want to restore data from a 25 GB disk, it is suggested to create a 26 GB disk on the target VM.
Symptom
This is observed on VMware ESX Server. For Windows 2003 VMs the default disk controller is LSI Logic SCSI adapter and the driver for this type of SCSI adapter is not included in Windows ADK 8.1. You may also observe this on some old servers with old SCSI adapters.
Solution
To solve this problem, obtain the drivers from the hardware vendor web site and load the driver from the BMR user interface.
[VDDKLOG] SSLCheckLockingCallback: locking callback overwritten! Expected 7FEE3A836E0, saw 113C2420
As a workaround, set the VDDKLogLevel registry key to 0 in the following path:
\HKEY_LOCAL_MACHINE\Software\Arcserve\Unified Data Protection\Engine\
Failed to create Snapshot copy '{xxx}_backup' across multiple volumes of the storage system ‘xxx’after retrying for '100' times with an interval of '10' seconds.
Follow these steps to create the registry key:
HKEY_LOCAL_MACHINE\SOFTWARE\Arcserve\Unified Data Protection\Engine
DoNotUseFlexCloneLicenseToCloneLun
After switching to Arcserve UDP Recovery Point View, the volume that is mounted under a file path in the backup session cannot be mounted.
Symptom
If the volume is still mounted, Arcserve UDP Recovery Point View directly opens the volume which is mounted by the operating system, not the backup session.
If the volume is dismounted, nothing is displayed in Arcserve UDP Recovery Point View under that file path.
Solution
Use Mount Recovery Point to mount this kind of volume.
Symptom
VMware recently published a KB article, which indicates that, in ESXi 6.0 or ESXi 6.0.x, a bug has been introduced in Changed Block Tracking (CBT) and it causes the CBT to return incorrect changed sector list. Arcserve UDP is potentially affected by this issue, which results inconsistent virtual machine backups (both full and incremental). For more details, refer to the VMware KB.
If the problem occurs, the catalog job may fail and Check Recovery Point may report errors.
Solution
VMware has released a patch for this issue. Apply this patch on your ESXi hosts. For more details, see the VMware KB article, or Arcserve KB article.
Symptom
This problem is observed if the Hyper-V host is Windows 2012 R2 and VM’s guest OS is SUSE Linux Enterprise Server (SLES) 12 with runlevel 5 (with graphical interface). In this case, the agentless backup job hangs at the “taking snapshot” phase and finally fails. After that, the VM guest OS becomes unresponsive.
Solution
This is not an issue of Arcserve UDP but a compatibility problem of Windows and SUSE. The same problem occurs when creating a VSS snapshot by using diskshadow command in the Hyper-V host. Follow these steps to verify the problem:
There is no workaround so far. We suggest you to work with Microsoft to solve the root cause. Alternatively, you may try changing the runlevel of SLES 12 from 5 to 3 (without graphical interface) but this does not guarantee any resolution.
Symptom
For VMware VM, pre-flight check (PFC) reports the following warning message for the Data Consistency check even when you have already provided proper credentials.
Not verified because the application failed to access the virtual machine. Verify that the user credentials are correct and have administrative privileges.
Solution
Only PFC for VMware VM has this issue. Other features such as backup are not affected. The workaround is to install Arcserve UDP Agent on the machine in which Arcserve UDP Console is installed (Agent service does not have to be started).
Symptom
Even when a SAN transport mode is possible, backup and restore jobs still use the HotAdd, NBD, or NBDSSL transport mode.
Solution
This is a known issue of VMware VDDK 6.0.1. There is no solution from VMware for now. For details you can refer to Know Issues in VDDK 6.0.1 release notes.
Symptom
Even when a SAN transport mode is possible, backup and restore jobs still use the HotAdd, NBD, or NBDSSL transport mode when the provisioned size of VM’s virtual disk is 4 TB or a multiple of 4 TB.
Solution
This is VMware known issue which, according to VMware, has been in following patches:
• For ESXi 5.5 - Patch Release ESXi550-201504001 (2112672)
• For ESXi 6.0 - Patch Release ESXi600-201505001 (2116125)
The workaround is to avoid using provisioned size which is the multiple of 4 TB. For example, do not use 4 TB or 8 TB; instead, use 3.9 TB or 8.1 TB.
Symptom
When you quiesce a VMware VM using VMware Tools, the snapshot contains corrupted data. The backup reads data from the snapshot and the data that is backed up also becomes corrupted. For more information about this issue, see the VMware KB article.
Note: This problem occurs with all VMware ESXi versions and on a VM with guest OS Windows 2008 R2 SP1 and Windows 2012. Arcserve UDP cannot detect the data corruption problem because VMware does not return an error in this case. You may not be aware of the problem until you try to restore data.
Solution
Perform the following tasks to detect and resolve the problem:
Symptom
When using the Microsoft VSS inside VM snapshot quiescing method to back up a VMware VM, the backup may not be consistent. Especially when backing up VM with applications (such as Exchange) installed.
Solution
The workaround is to use VMware Tools snapshot quiescing method, along with disabling the VSS writers MSSearch Service Writer and Shadow Copy Optimization Writer in guest OS of VM before this problem gets fixed.
Unable to connect the network adapter <<adapter name>> to the virtual switch.
Symptom
When backing up a VM, failover happens for the clustered VM before taking a snapshot . This failover causes the Hyper-V host that is recorded in the backup session to be inconsistent with the VM configuration.
Solution
You can manually connect the network adapter to a virtual switch on the Hyper-V host or use the Restore to an alternative location option to recover the VM that lets you set the restore configuration for the VM.
Symptom
Although the backup job of one virtual machine has already finished, the virtual machine's status is still "Backup up" in the Hyper-V Manager. Therefore, if another backup job for this VM starts at this time, it will fail with error "The Hyper-V VSS writer has encountered an error when processing this virtual machine". In addition, you cannot perform some operations, such as power on/off, for the VM at that time in the Hyper-V Manager. If the VM is a Hyper-V cluster, you cannot perform live migration for that VM.
This problem happens during the following situations:
Solution
While the VM is "locked", you can still use the guest OS as normal. Therefore, this has no impact on the usage/availability of the guest OS. However, if you have concerns and want to avoid this situation, you can do either of the following:
Symptom
This is a known VMware issue where it involves Changed Block Tracking (CBT). With application level quiescing, Changed Block Tracking overstates changes.
Solution
The issue is fixed in VMware ESXi 5.5 or later, and in VMware ESX 5.1 Patch 02. If VMware vCenter Server 5.5 manages any VMware ESXi 5.1 hosts, then the patch should be applied to them. For more information on this fix, see the VMware KB.
If you still see the issue, then set the following registry on the proxy server:
[HKEY_LOCAL_MACHINE\SOFTWARE\Arcserve\Unified Data Protection\AFBackupDll\<VM instance UUID>]
"ResetCBT"=dword:00000001
Example:
[HKEY_LOCAL_MACHINE\SOFTWARE\Arcserve\Unified Data Protection\AFBackupDll\502d3c43-e3c9-9919-78f9-89082ca5e1cc]
"ResetCBT"=dword:00000001
Note: After the registry value is set, the next Incremental backup job converts to a Verify backup job and then the subsequent Incremental backup jobs continue to run with the appropriate size.
For more information, see the VMware KB Article 2055943.
Symptom
The backup source VM has VMDKs larger than 2-TB and the ESX/ESX(i) server with a version lower than 5.5 can only support a virtual disk of up to 2-TB disk in size. However, during the VM recovery, the following errors can occur:
Solution
This is a VMware limitation. The maximum size that VMware ESX/ESX(i) server version lower than 5.5 supports is 2 T-16 GB, which is equal to 2032 GB.
It is recommended to use VMware ESX(i) server 5.5 as the destination to perform conversion with a large disk.
For more information, see the VMware KB Article 1012384.
Symptom
When you run a backup job on a virtual machine containing a SCSI controller with more than 7 VMDKs, the backup job fails. The backup job fails because VMware requires a maximum number of free slots for VMDKs on a particular SCSI controller to create a snapshot. A SCSI controller can have a maximum number of 15 slots. For example, if a SCSI controller has 7 VMDKs, a snapshot can be created for each VMDK. (A total of 14 slots are used with one slot free.) If a SCSI controller has 8 VMDKs, the backup job fails because the snapshot cannot be created due to only 15 available slots.
Note: Manually creating a snapshot also fails.
Solution
For virtual machines with more than seven disks on a single SCSI controller, perform the following steps:
Snapshots can now be created for each VMDK.
This issue is a VMware limitation where Arcserve Backup can only support the number of VMDK disks for backups.
For more details, see the following VMware Knowledge Base article: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2015181
To resolve this issue, remove the old VM from Central Protection Manager and import the new VM.
Symptom
This is a known VMware issue affecting ESXi 5.0, 5.1, and 5.5 hosts and virtual machines using the E1000 and E1000e virtual network adapters.
Note: This issue is resolved in the following VMware updates.
Solution
Switch to a VMXNET3 adapter. For more information, see the VMware KB Article 2059053.
Symptom
When a VM is imported to the node view, it fails if another VM with same VM instance UUID is already added to the node view.
Symptom
When a host-based VM backup plan includes a proxy machine that has the previous release of Arcserve D2D installed (for example, r16.5), the error message “Cannot find dispatch method” appears when the plan is saved.
Solution
This problem occurs because the API of the current version is not compatible with the API of the previous version of Arcserve D2D. As a work around, you can manually upgrade Arcserve D2D to the current version of the Arcserve UDP Agent (Windows).
Symptom
The host-based VM backup job hangs for hours and cannot proceed.
Solution
End afbackend.exe according to the process ID in the activity log, remove the VM snapshot if any, and resubmit the backup job.
Symptom
The behavior of the operating system creates the offline state by default.
The SAN policy was introduced in Windows Server 2008 to protect shared disks that are accessed by multiple servers. The default SAN policy from the source VM is “Offline Shared” for all SAN disks except the boot disk. Setting the policy to Offline enables the SAN disks to be offline during startup. After recovery, a new disk for the VM is created. The disk file from the VM appears to be a SAN disk where the operating system identifies it as being offline. When the offline disk is set to be back online, the disk remains online even after rebooting the system.
Solution
As a workaround, specify the DISKPART.exe command: SAN POLICY=OnlineAll setting for the source VM before backup. Because the disks can be shared among other servers, data corruption can occur. You must use the correct SAN policy to protect the data.
DISKPART.EXE command line
Query SAN policy:
DISKPART > san
SAN Policy: Offline Shared
Change SAN policy:
DISKPART > san policy=OnlineAll
DISKPART successfully changes the SAN policy for the current operating system.
Symptom
After backing up a Hyper-V VM, the volumes on iSCSI devices are not listed in the restore UI.
Solution
Create an agent-based backup plan in Arcserve UDP or use Arcserve UDP Agent (Windows) to back up the virtual machine.
Symptom
For host-based agentless backup jobs for a Hyper-V VM, if the guest operating system is Windows Server 2003, the Pre/Post commands cannot be executed. The activity log prints the warning "The virtual machine name is not expected. Pre/Post commands cannot be executed".
Solution
Windows Server 2008, Windows Vista, or a later version of operating system do not have this problem and should be used.
Symptom
When Windows 2003 R2 64-bit machine is used as the backup proxy server to protect VMware VM, at times the backup job may crash. You can see error messages as following in the backup job debug log file:
[2016/01/21 10:18:11:316 00 03820 03336 ] [VDDKLOG] VixDiskLib: VixDiskLib_OpenEx: Open a disk. {AFBackend.exe::AFBackupVirtual.dll(1746.0)}
[2016/01/21 10:18:11:316 00 03820 03336 ] [VDDKLOG] VixDiskLibVim: VixDiskLibVim_GetNfcTicket: Get NFC ticket for [datastore1 (3)] VMname/VMware_1.vmdk. {AFBackend.exe::AFBackupVirtual.dll(1746.0)}
[2016/01/21 10:19:11:691 00 03820 03336 ] [VDDKLOG] VixDiskLibVim: Error 18000 (listener error GVmomiFaultInvalidResponse). {AFBackend.exe::AFBackupVirtual.dll(1746.0)}
[2016/01/21 10:19:11:691 00 03820 03336 ] [VDDKLOG] VixDiskLibVim: Login failure. Callback error 18000 at 2439. {AFBackend.exe::AFBackupVirtual.dll(1746.0)}
[2016/01/21 10:19:11:691 00 03820 03336 ] [VDDKLOG] VixDiskLibVim: Failed to find the VM. Error 18000 at 2511. {AFBackend.exe::AFBackupVirtual.dll(1746.0)}
Solution
In Arcserve UDP Version 6.0, VMWare VDDK 6.0.1 is built in. But VDDK 6.0.1 does not officially support Windows 2003 R2. As a workaround, you can use one of the following options:
How to apply different version of VDDK other than the built-in Version (6.0.1) in Arcserve UDP
Symptom
When you use the ESX information to log in and select a resource pool as the location of the Instant VM, the Instant VM is located under the ESX host, and not in the resource pool. If you delete the Instant VM from Console and recreate an Instant VM with the same name, it will add a suffix (1) to the name of the Instant VM and not to the timestamp as expected.
Solution
This is because the ESX server is managed by a vCenter. Use the vCenter login information to create the Instant VM.
Symptom
When I start the Instant VM, and then restart the Hyper-V recovery server, the Instant VM may fail to boot.
Solution
To resolve this boot failure, restart the Instant VM.
Symptom
I have assigned a Windows volume as an NFS shared folder and provided this shared volume path as the Instant VM file path. When I format the Windows volume and then try to create the Instant VM, the Instant VM fails to get created. The Instant VM fails to get created because the ESXi host fails to add the NFS data store. VMware displays the following error messages:
VMware create VM failed.
The following error log is displayed in the vmkernal.log file:
No underlying device for major, minor
Solution
To resolve this issue, perform these steps on the Recovery Server.
The Administrative Tools dialog opens.
The Services dialog opens.
The Instant VM gets created.
The log filter does not work when the NodeName is VM (host name) and the log view is loaded by clicking the View Logs link.
Symptom
In the log view, if a protected VM does not have a hostname, then its NodeName value is displayed as "VM(node name)". In this case, other filters do not work.
Solution
You can manually modify the NodeName from VM (host name) to host name, and then all the filters will work fine. For example, modify NodeName "VM(xxxxx01-AB)" to "xxxxx01-AB".
Install mounting driver failed with error code 1460 in the debug log.
Symptom
The Windows setup API reports error 1460, which means that the timeout period expired. The default timeout value is 300 seconds for updating the device driver.
Solution
To adjust the timeout value, follow these steps:
Exchange database restore fails with error: "failed to mount it".
Symptom
This occurs when the Exchange server is installed inside a VM, and is running Arcserve UDP Agent (Windows) to protect this VM. After the backup runs for some time, if the VM is reverted to a previously saved VM snapshot, and you try to recover the Exchange database to the original location, the restore fails with error "Exchange storage group/database [DB_Name] has been restored to its original location, but failed to mount it."
The root cause is still being investigated, but currently seems to be related to the time stamp recorded inside the Exchange database and the latest transaction log files.
Solution
Try to recover the Exchange database to an alternative directory, or dismount the database and remove all files in the destination folder before performing the restore.
The Arcserve UDP Console may respond slowly if Microsoft SQL Server cannot allocate more memory.
Symptom
Microsoft SQL Server may need to allocate more memory to process the query, especially when there is extensive data in the Arcserve UDP database. But if Microsoft SQL Server cannot acquire memory due to the non-availability or maximum limit configured, then query processing will become very slow and therefore impact the Arcserve UDP Console responsiveness.
Solution
Delete some logs from the database using Log/Delete and restart the SQL Service.
Symptom
When you submit a merge job manually, the job does not run and the following message is displayed in the Activity Log:
Cannot run merge job for <node name>, another job is running
Solution
The reason for this error could be that there are other jobs running for the node in the same Recovery Point Server but different data store. As a workaround, wait for those jobs to complete and then try again.
Symptom
Backup or replication job fails with error “system cannot find the file specified”.
Check Windows Event log. McAfee detects the data store file (for example, P0000000042.data ) as Exploit-ScriptNull Trojan virus and deletes it.
Solution
Configure the antivirus setting to set the Recovery Point Server (RPS) Data Store location in the exclusion list.
Note: Some antivirus software requires you to set the exclusion list on the server side.
Symptom
This happens when you set the network throttling to a pretty low bandwidth or the network throughput to the destination RPS server is slow. Therefore, it may take several minutes to wait for the queued data to be sent.
Solution
Wait for the replication job to exit gracefully.
Symptom
When you try to add a node by IP address or node name on Windows operating systems that support User Account Control (UAC), (Windows Vista or later versions), or try to import virtual machines from a Hyper-V server that support UAC, and you use a new Windows user account that is a local account in the administrators group but it is not the built-in administrator, the following message displays:
"Administrator privilege is required."
Solution
Use a built-in administrator or a domain administrator. Or, you can disable Remote UAC.
This is the default Windows behavior called UAC remote restrictions. If you still want to use this account to add the node, disable Remote UAC by performing the following steps:
The Windows Registry Editor opens.
Note: You may need to provide administrative credentials to open Windows Registry Editor.
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\System
For more detailed information about Windows behaviors, see the article http://support.microsoft.com/kb/951016.
Symptom
Removing the hypervisor function for a node is not provided in this release.
Solution
Remove the node and add it again.
Symptom
With the all-in-one UI in Arcserve UDP, it is not expected that users will have multiple Arcserve UDP Consoles and therefore the scenario of moving the agent from one server to another will only happen when the old server is retired.
Solution
Moving the node from one console to another is a very rare scenario.
If the backup destination is a deduplicated-enabled NTFS volume on Windows Server 2012 and Arcserve UDP Agent (Windows) is installed on Windows 2003, restore, merge, or catalog jobs can fail. The job fails with Windows error code 50 and message "The request is not supported".
Symptom
McAfee or another third party software set EnableECP=1 in the registry when they are installed on Windows Server 2012.
Solution
In the registry, change the value of EnableECP from 1 to 0 under the following key:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\services\LanManServer\Parameters
For more information, see http://support.microsoft.com/kb/2817216.
Symptom
The second replication job will hang in the Prepare status and fail after 10 minutes.
Solution
You do not need to perform any specific operation. After the second replication job fails, the makeup job will be triggered.
Symptom
Replication (in) job fails and activity log displays the message “Failed to lock session on <path>. Session is already locked by File Copy job, computer name: <name>, Process ID <ID>”. It happens when backup with the File System Catalog job is enabled, and you configure File Copy job to run from replication target data store.
Solution
The File Copy job locks the session thus replication job cannot run. To resolve, you can use either of the following options:
Symptom
After a restore job, the following item properties may not be properly restored:
Solution
Use Export to PST option to restore the missing items.
Symptom
This problem can occur under the following conditions:
Solution
To remedy these conditions, perform one of the following resolutions to help you address the issue:
Symptom
Arcserve Cloud: Bucket Region is not Populated and Backed up to Amazon S3 when the Bucket Region is Ireland.
Solution
For solution, see the KB article.
Symptom
When using Update 4 or lower version agents, the following errors occur:
Exchange storage group/database has been restored to its original location, but failed to mount it.”
Solution
Upgrade the agents to Arcserve UDP Agent Version 6.0.
Symptom
When the restore destination is a remote share folder such as \\FileServer\ShareFolder\RestDest, and the backup destination is a remote share folder with the same path, such as \\FileServer\ShareFolder\RestDest, the error occurs.
If the user account which is used to connect to the root share folder is not in the permission list, the restore job will fail no matter what account is used for the restore destination folder.
Solution
Add the user account which is assigned for the backup destination to build the connection into the permission list of the root share folder and ensure the user account has proper permissions to restore the file.
Add the user account into the Backup Operators group and ensure it has the permission to override security restrictions.
Symptom
If encryption is enabled for the recovery point, you do not have to enter the password for recovery points backed up from the current server. However, if you upgrade the Windows OS (For example, from Windows 2008 to Windows 2008 R2), the passwords will not be automatically filled in on the Arcserve UDP Agent (Windows) user interface and you must re-enter the password.
Solution
Record the recovery point encryption password or session password and keep it in a safe place for retrieval.
Symptom
While restoring a VMware virtual machine, I get the following error:
You do not have access rights to this file, for more information, see the restore debug log. If necessary contact Arcserve support.
You can see the following log entries in the restore debug logs:
[VDDKLOG] CnxAuthdConnect: Returning false because SSL verification requested and target authd does not support SSL
[VDDKLOG] CnxConnectAuthd: Returning false because CnxAuthdConnect failed
[VDDKLOG] Cnx_Connect: Returning false because CnxConnectAuthd failed
[VDDKLOG] Cnx_Connect: Error message: SSL required
Solution
The reason for this error is that the SSL authentication is disabled on the VMware ESX host. To resolve this error, use one of the following methods:
config.defaults.security.host.ruissl
/etc/vmware/config;
The SSL authentication is enabled.
Server connection error is displayed when browsing the logs.
Symptom
When you browse the activity logs from the Agent or the Console, the following error might be displayed:
Failed to connect to server 'arcservedocs.com'.
Solution
You can just ignore this message.
Symptom
A node is backed up to an RPS data store, replicated to another RPS data store, and then created a Virtual Standby machine with the source of replication. After the plan deployment, you can resume Virtual Standby directly if there are older session for the Virtual Standby job. But, after you submit the job, the Virtual Standby job does not start.
Solution
To resolve this issue, perform a manual replication first. Right-click the node and click Replicate now. After the replication is complete, resume the Virtual Standby job.
Symptom
The virtual standby job fails with the following error:
Fail to get disk signature (signature is empty)
Solution
In Arcserve UDP Version 6.0, VMWare VDDK lib is upgraded from version 5.5 to version 6.0. In VDDK v6, initializing the VDDK lib needs a new field named thumbprint. The call from the VSB converter is based on Arcserve UDP Version 5.0, where the thumbprint is missing. This causes the call from the Converter to Monitor to fail
To resolve this problem, update the Converter to Arcserve UDP Version 6.0.
Symptom
The Arcserve UDP Agent (Windows) is backed up to Arcserve UDP RPS data store and the RPS is upgraded to Arcserve UDP 6.0. Also, Arcserve UDP Agent (Windows) is backed up to a shared folder and the Agent is upgraded to Arcserve UDP 6.0. The virtual standby VSB job to Hyper-V fail and the following errors are displayed:
Cannot find dispatch method for {http://webservice.arcflash.com}IsVmFileExist.
or
Cannot find dispatch method for {https://webservice.arcflash.com}IsVmFileExist.
Solution
Upgrade the Arcserve UDP Agent to Arcserve UDP 6.0 on the Hyper-V Server.
Another cause for this behavior relates powering on virtual machines that contain "read only" volumes. To correct this condition, place the volumes on the disk in a writable state.
Failed to get Recovery Point information.
This behavior occurs when you perform a V2P recovery using the latest snapshot and a conversion job for the node did not complete after the virtual standby task was redeployed to the node.
Solution
Solution
Perform the V2P recovery using Arcserve UDP Agent Bare Metal Recovery.
Note: This limitation applies to only virtual standby jobs running on Hyper-V servers.
Symptom
The Arcserve UDP Agent (Windows) has a disk of 2-TB and the ESX/ESX(i) server with a version lower than 5.5 can only support a virtual disk of up to 2-TB disk in size. However, during the conversion, the following errors can occur:
Solution
This is a VMware limitation. The maximum size that VMware ESX/ESX(i) server version lower than 5.5 supports is 2 T-16 GB, which is equal to 2032 GB.
It is recommended to use VMware ESX(i) server 5.5 as the destination to perform conversion with a large disk.
For more information, see the VMware KB Article 1012384.
Symptom
Deploying the plan failed and the following error message displays: "Unable to apply 'Virtual Standby settings' to node 'xxx'. (Failed to connect from xxx to the monitor: xxx. Invalid user credentials)".
Solution
Edit the Virtual Standby task in the plan, enter the correct password for the monitor, and save the plan.
Catalog Generation Fails for Agentless Backup of VMware VMs
Symptom
Due to a VSS snapshot problem caused by VMware or Microsoft inside the VM guest OS, catalog generation for agentless backup of VMware VMs may fail with the following message:
"Failed to map an index block to the correct volume block. An IndexAlloc Error has occurred."
Solution
Note: This problem occurs rarely. For more information, see the VMware KB Article 2006849 or the Microsoft KB Article 2853247.
File copy /archive jobs are not launched.
Symptom
File copy /archive jobs are not launched.
Solution
In some rare cases the file used to indicate a recovery point needs to be kept for file copy / archive job and cannot be removed after completion of the file copy / archive job. Such files are named as *.alck under the backup destination folder of a specific node, and file size is zero. As a workaround, you can find out these files and manually delete them.
Symptom
Download “Arcserve Backup” or “Arcserve High Availability” components from Single Installer (using ASDownloader.exe) fails when the destination volume is FAT32 filesystem because the package exceeds the 4GB file size limitation that FAT32 file system supports.
Solution
As a workaround, you can download to NTFS volume.
After upgrade to Arcserve UDP v6 on the Arcserve UDP Appliance, you will see Factory Reset in the Settings tab of the Arcserve UDP Console. If you try to perform factory reset by clicking Perform Factory Reset, and click Reset on the Confirm Factory Reset dialog, following error message is displayed:
Appliance Factory Reset failed. Manually reset the appliance factory by using the following command: "powershell.exe .arcserve_factoryreset.ps1 –perserve_data –auto_reboot " in cmd under path C:\Program Files\Arcserve\Unified Data Protection\Management\Appliance.
Note: The error message is not correct. For Appliance user who upgrades to UDP v6, Factory Reset is not supported because there is no Arcserve UDP recovery partition on the Appliance machine.
The instant VM for Linux recovery point is not supported on the appliance if you use the Linux Backup Server on the appliance.
Symptom
The Linux instant VM job fails and displays the following error message:
Failed to get the IP address from the VM $vmname. Please verify if the VM and the backup server are in the same network.
Solution
The failure is caused by the standby VM created in the instant VM job. It tries to connect to the Linux Backup Server through Appliance_hostname:8014 although you added Linux Backup Server to Console by Appliance_hostname:8018. As on the appliance machine the port 8014 is monitored by UDP Windows agent service, the instant VM job will fail.
You can use a workaround to resolve for Linux instant VM job with static IP.
Follow these steps:
You can use "netstat -aon|findstr "port" " to verify if the port is occupied.
Note: If you are upgrading Arcserve UDP v6, port 8018 on the UDP Appliance machine is configured to redirect to Linux Backup Server’s 8014. Use the following command to release 8018 on the UDP Appliance machine:
netsh interface portproxy delete v4tov4 listenport=8018
Note: If the file does not exist, create the file. Run the following command to restart Linux Backup Server:
/opt/Arcserve/d2dserver/bin/d2dserver restart
#iptables -A INPUT -p tcp --dport 8018 -j ACCEPT
#iptables -A INPUT -p tcp --dport 8035 -j ACCEPT
#/etc/init.d/iptables save
netsh interface portproxy add v4tov4 listenport=8018 connectaddress=192.168.10.2 connectport=8018 protocol=tcp
netsh interface portproxy add v4tov4 listenport=8035 connectaddress=192.168.10.2 connectport=8035 protocol=tcp
C:\Program Files\Arcserve\Unified Data Protection\Engine\BIN\Appliance
netsh interface portproxy add v4tov4 listenport=8018 connectaddress=$VMIp connectport=8018 protocol=tcp
netsh interface portproxy add v4tov4 listenport=8035 connectaddress=$VMIp connectport=8035 protocol=tcp
Important! The workaround does not apply to the following options:
Copyright © 2016 |
|