VMware : [FAQ] Frequently Asked Questions

[Gostev]

Navigation

General FAQ
This topic below covers general information about the product, and core features which are not hypervisor-specific.
READ THIS FIRST : [FAQ] FREQUENTLY ASKED QUESTIONS

Microsoft Hyper-V FAQ
The following FAQ covers Microsoft Hyper-V specific questions.
Hyper-V : [FAQ] FREQUENTLY ASKED QUESTIONS (pending content collection)

VMware vSphere FAQ
This topic covers VMware vSphere specific questions.

[Gostev]

Table of Content

General
System Requirements
Upgrades

Technology
VM Data Access Overview
Changed Block Tracking (CBT)

Processing Modes
Direct Storage Access Mode
Virtual Appliance Mode (Hot Add)
Network Mode (NBD)

Architecture
VMware Backup Proxies

Jobs
Replication
Migration
VM Copy
Continuous Data Protection (CDP)

Restore
Full VM Restore
VM Files Restore
vPower

[Gostev]

System Requirements

For complete list of the system requirements and supported configurations for all product components and protected VMs, please refer to the System Requirements section of the Release Notes document that comes with your download, and available on product page under Resources tab.

VMware Versions

Q: How can I know which Veeam versions are compatible with which VMware version?
A: You can confirm which VMware versions are supported by the current Veeam version in our User Guide.

You can check the full Veeam and VMware compatibility matrix in this KB article.

Q: Is vCenter Server presence required?
A: No, standalone ESXi hosts are also supported.

Q: Is ESXi fully supported?
A: Yes, licensed (paid-for) ESXi is fully supported.

VMware Licensing

Q: Is free ESXi (also known as vSphere Hypervisor) supported?
A: Free ESXi is not supported, because it has vStorage API for Data Protection and other management APIs locked down specifically to prevent ISVs from being able to backup/manage such hosts.

Q: What is the minimum VMware licensing required?
A: vSphere Essentials.

VMware Infrastructure

Q: Is processing of VMware vCloud Director provisioned virtual machines supported?
A: Yes.

Q: Is DVS (distributed virtual switch) supported?
A: Yes.

Q: What are the minimum privileges for the service account?
A: The minimum cumulative permissions can be found in our User Guide. Each operation has different permissions, so for full utilization of all operations involving VMware, use the cumulative minimum permissions.

Veeam Backup Server

Q: Is running Veeam B&R components in a VM supported?
A: Yes, we fully support this, moreover we provide “Virtual Appliance” processing mode that is specifically optimized for virtual backup proxy.

Q: How typical is it to run Veeam in a VM among existing customers?
A: This is typical for non-24/7, or low consolidation ratio environments, which have host resources mostly unused during the night. 24/7 shops with high consolidation ratio prefer using physical servers and direct-from-SAN backups to completely offload backup processing activities from production hosts.

Q: Where can the best performance be gained? With Veeam Backup installed as a physical or virtual server?
A: About same. Performance mostly depends on actual Veeam Backup server performance (number of CPU cores, and memory throughput). Veeam Backup running in VM on modern ESX host hardware is often faster than one running on older physical server.

Q: If the Veeam Backup server is running as a virtual machine, is there a 2TB limit backup target for each backup server?
A: No, there is no such limit in the product. Limit solely depends on your chosen backup target.

Storage

Q: Does the product support backup directly from storage (FC, iSCSI and NFS)?
A: Yes, via direct storage access mode.

Q: Does the product support restore directly to FC, iSCSI and NFS?
A: Yes. For FC and iSCSI keep in mind that the result will be thick provisioned disks. As an alternative you can use hot add mode for FC and iSCSI.


Networking

Q: Are there any TCP/IP ports required to communicate between the Veeam B&R components to the VMs in order to successfully back up the VM?
A: No.

Q: Do you encrypt network traffic?
A: You can optionally encrypt all network traffic.

Virtual Machines

Q: Is Veeam compatible with old guest operating systems?
A: Yes, we support any OS supported by VMware.

Q: What types of virtual disks are supported for VM?
A: All disks except RDM in physical mode, and disks mounted via in-guest iSCSI initiator. Such disks are skipped during the backup. Additionally, independent disk are skipped, because they are excluded from VMware snapshot. See the Virtual Hardware section here in the User Guide for information on supported disk types

Q: Is VMware snapshot a requirement?
A: Yes, this is a requirement of vStorage API for Data Protection. This in turn means that Veeam B&R cannot protect VMs for which VMware snapshot cannot be created (for examples, VMware does not support snapshots of VMs with disks engaged in SCSI bus sharing).

[Gostev]

Upgrades

Q: Do I need to perform clean install and re-create all jobs, or can I upgrade my existing Veeam Backup install seamless?
A: We support in-place upgrades according to the release notes. If you are running older versions, please contact our support for assistance with performing an upgrade.

Q: Can different versions be installed on the same server?
A: No. You can, however, run different versions in parallel on different servers while processing the same VMs without any issues (as long as jobs do not overlap). This is the approach most of our customers choose for POC testing of new versions before upgrading their production deployment.

Q: Can Enterprise Manager collect data from older Veeam Backup & Replication servers?
A: Yes. Enterprise Manager can collect data from older versions according to the release notes (“Upgrading Veeam Backup & Replication section”).

Q: Can I restore backups made with older versions?
A: Yes. You can restore backups made with any version of Veeam Backup & Replication starting from v1.

[Gostev]

Technology

Changed Block Tracking (CBT)

Q: Will the changed block tracking "break" if I backup and replicate the same VM? Or, if I backup the same VM with different jobs, or from different backup servers, or with different backup products?
A: There will be no issues, as changed block tracking is designed to handle such usage scenarios.

Q: Does changed block tracking work with NFS?
A: Yes, changed block tracking supports all types of storage (except RDM disk in physical mode which are not supported for snapshot-based VM backup).

[Gostev]

Direct Storage Access Mode

Q: How does it work?
A: Backup proxy server retrieves protected VM's virtual disks directly from production storage over SAN fabric or NFS, thus providing LAN-free data retrieval mechanism that does not affect production ESX(i) hosts or management network.

Q: How to configure the new backup proxy server for direct storage access?
A: Disable automount (“offline shared” is default since Server 2012 and it is also done automatically when backup proxy server components are installed). Install and configure SAN HBA card (or iSCSI Software Initiator). Configure fibre channel zoning and give the backup proxy server machine access and read permission on VMFS LUNs (if applicable to your SAN). Ensure that you are seeing your VMFS volumes in the Windows Disk Management snap-in as "offline".

Q: I added new LUNs and can see them in the Windows Disk Management snap-in, but the backup proxy cannot seem to "see" them?
A: Right-click the Backup Infrastructure node in the management tree, and choose the Rescan Storage command to force the storage detection on all backup proxy servers. This operation also happens automatically once every 24 hours.

Q: Can I use Direct Storage Access mode with NFS storage?
A: Yes. While natively VDDK SAN mode supports only block storage (FC and iSCSI), starting v9 Veeam provides proprietary NFS client for direct access to NFS storage.

Q: I have FC SAN and running Veeam Backup in a VM. Can I use direct SAN access mode via NPIV?
A: No. VMware supports NPIV for RDM disks in physical mode only.

Q: I have iSCSI SAN and running Veeam Backup in a VM. Can I use direct SAN access mode?
A: Yes, you need to configure direct SAN access using Microsoft iSCSI Software Initiator bundled with Windows.

Q: I have never configured Microsoft iSCSI Software Initiator, settings seem too complex.
A: See the following step-by-step guide: The Mission Manual Part 1: Veeam B & R Direct SAN Backups (courtesy of Justin Paul)

Q: Any tips on tuning of Software iSCSI initiator built into Windows?
A: Based on my experiments, the currently shipping version works great out of box. Be sure that the backup proxy server is not starving on CPU (I found that Software iSCSI Initiator is quite hungry on CPU resources). Jumbo frames are worth testing.

Q: What is the expected backup performance for direct SAN mode?
A: Depending on configuration full backup performance goes up to link speed. We have seen 4GByte/s with one proxy, but it depends highly on your environment.

[Gostev]

Virtual Appliance Mode (Hot Add)

Q: How does it work?
A: Backup proxy server uses SCSI Hot Add capability of VMware to attach disks of backed up VM directly to itself, and thus get direct access to data stored inside. As a result, data is retrieved directly from storage via ESX(i) storage stack (bypassing network stack), which allows for significantly better performance.

Q: How do I configure a backup proxy VM for Virtual Appliance mode? What host should I run the VM on?
A: No special configuration is needed for VM, however we recommended that you add additional virtual SCSI controller to the VM (edit its virtual hardware settings). We also recommend up to 8 vCPUs to avoid virtual NUMA performance issues. The host running backup proxy server VM must have all datastores where protected VMs' disks reside connected to it, otherwise hot add will not be possible.

Q: What is the expected backup performance for Virtual Appliance mode?
A: It depends primarily on production & backup storage performance and the connection. As the number of tasks on a virtual proxy is usually lower than on physical proxies, we have seen 700-1000 Mbyte/s maximum speed

Q: Can I use Virtual Appliance mode with local storage?
A: Yes. Any storage is supported, as long as it is connected to the ESX host running Veeam Backup VM. Which in turn mean means that if you only have local storage, you have to dedicate at least one VM on each host to be your hot add backup proxy (you can use your existing VMs for this).

Q: Can I use Virtual Appliance mode with NFS storage?
A: It is not recommended. Please use Direct-NFS mode (direct SAN access).

Q: Virtual Appliance mode fails to process some of my VMs and fails over to network mode, any idea why?
A: Most common reason is that the datastore is not mounted on the ESX host where the hot-add proxy is running. Also vStorage API limits could be the reason. See requirements and limitations in the User Guide.

[Gostev]

Network Mode (NBD)

Q: How does it work?
A: Backup proxy server retrieves source VM data from production storage through ESX(i) network stack over management network.

Q: When it is recommended to use Network processing mode?
A: To enable a single backup proxy server to process VMs from multiple hosts with local storage. Unless you have 10Gb Ethernet or faster, expect very slow performance (so with large amounts of data to process, consider using Virtual Appliance mode with one backup proxy VM per host instead).

[Gostev]

Architecture

Backup Proxy

Q: How does the product interact with VM data?
A: Veeam B&R uses proxy architecture. Both reading and writing VM data is done via backup proxy server (this includes restore and replication).

Q: Do backup proxy servers ever talk with vCenter?
A: Yes, they communicate with vCenter e.g. for snapshot coordination.

Q: What happens if backup server cannot contact backup proxies in the remote site, will they still run jobs?
A: No. For each VM, backup server assign proxy is assigned dynamically in the run-time. If some proxies are unreachable, backup server will consider them offline and will not assign tasks to them.

Q: Do I have to set up multiple backup proxies?
A: No, it is not required - but recommended for performance and redundancy

Q: How is the backup proxy selected in cases when there are multiple proxies available?
A: For each VM in the job, backup server classifies all backup proxies available to the job into 3 groups by access type they have to the VM's disks: SAN (1), hot add (2) and network (3). Proxy with least tasks currently assigned is picked from group 1, and if responding, VM processing task is assigned to this proxy. Otherwise, the next least busy proxy is selected from group 1. If all available proxies group 1 are already running max number of concurrent tasks, selection process switches to group 2 and repeats. Group 3 is only used if groups 1 and 2 have no responding proxies. For group 3, proxy subnet is also considered when picking the best proxy.

Q: What if no proxy is available to process the specific VM at the moment?
A: If no proxies are available in either group, selection process "sleeps" for 60 seconds - or less, if some proxy reports to backup server that it has finished processing a task - and then retries. Proxy selection algorithm will wait for backup proxy to become available for up to 23 hours by default, and then fail the job processing.

Q: What is considered a "task" when you say "concurrent tasks".
A: A task is processing a virtual disk. Keep in mind that the same proxy can lose task slots to replication (acting as a target proxy), as well as to restore operations. Moreover, restore tasks have the highest priority, and are first to be assigned to proxies as soon as they become available.

Q: Which transport mode is “best” for me?
Q: This depends entirely on your infrastructure; testing is essential as there are many factors that contribute to the overall performance, most of which are unrelated to your Veeam configuration. It is essential to perform real testing in your environment across the transport modes; there is not a “strict” hierarchy for which transport mode is the best; it is entirely normal for hotadd to outperform DirectSAN, for example, as DirectSAN is always synchronous by design while hotadd is asynchronous.

Depending on your environment, different transport modes will have different performance, so it’s important to do real world tests with your environment as differences between environments can lead to incorrect understandings of expected performance.

Q: What are the benefits and considerations for each transport mode?
A: Please see the outline below:

DirectSAN
Benefits:

• Potentially fastest mode as DirectSAN reads directly from the production storage
• Less impact on VMware environment

Considerations:

• Can be more complex to configure depending on the connection type (Fibrechannel, iscsi, SAS, etc)
• More potential points of failure (HBA, cabling, etc)
• DirectSAN is “sensitive” to VMs running on snapshots and backups may fail with unclear errors
• DirectSAN works best with Thick Provisioned disks, especially for restores. Thin Provisioned disks will experience performance penalties during restore operations, so always restore as Thick (different storages work better with Lazy/Eager Zeroed disks, so testing is required to find the best result)
• No vSAN or vVOL backups with DirectSAN, use Hotadd or NBD

Hotadd
Benefits:

• Easy to deploy/configure in Veeam
• Competitive speeds gated only by the OS’ ability to read/write to the attached VMDKs

Considerations:

• Hotadding disks may take some time (15-90 seconds), which adds “dead time” during processing; mitigate by scaling horizontally (more hotadd proxies)
• More load on VMware environment (mitigate by limiting concurrent tasks per proxy)

NBD

Benefits:

• Easy to deploy/configure proxies (if Veeam/Proxies can reach the vCenter/ESXi host(s), likely NBD will work)

Considerations:

• Throughput is capped at approximately 1.3 Gbit/s according to reports from user testing and from VMware docs (Veeam will use async I/O by default, but real world testing from users has shown that 120 MB/s is a more realistic max for most environments)
• NBD has a max number of connections allowed as it uses NFC connection which is also used by other VMware operations. VMware recommends no more than 50 concurrent disk backups, but 40 is a much “safer” max estimate
• NBD can be very sensitive to network interruptions (unreliable connections, Firewall interference (IPS/IDS), Antivirus on Backup Proxy, etc) and will break the connection without an option to reconnect. Ensure your Backup Proxies are as close to the Vmware environment as possible on the network (or on same network) and ensure that the ports from proxy to Virtualization Servers are whitelisted on your firewall.

[Gostev]

Replication

Q: What is the maximum amount of restore points supported?
A: You can have up to 28 replica restore points (4 weeks of daily replicas). Keep in mind that a large number of snapshots can lead to performance issues after failover. Also keep in mind that only replicas without additional backup do not meet the 3-2-1 rule.

Q: What happens if replication job fails in the middle of the incremental run.
A: Existing restore points remain intact, and new (unfinished) restore point will be automatically removed by the following job run.

Q: I want to test replica failover. Can I simply power on any replica restore point with vSphere Client?
A: While you can manually power-on replicas created by Veeam, it is best to perform all failover/failback operations within Veeam, as doing so outside of Veeam may disrupt replica operations and potentially cause data-loss.

Q: Does failback transfer changes only?
A: Yes.

Q: Why using replication when there is Instant VM Recovery feature for backups?
A: Unlike instantly recovered VMs, replicas are created in native format and can this can run at full I/O speed, and without any dependencies on vPower NFS server. These capabilities are very important in case of major disaster with multiple VMs affected, as well as for high I/O VMs. Replication also support advanced replica-specific functionality such as Re-IP addressing on failover, failback and so on.

Q: How does Veeam replication compare with SAN-based replication?
A: Our solution is storage-agnostic (any storage to any storage), granular (per VM versus per LUN), uses much less network bandwidth, and is application-aware to ensure successful recovery.

Q: If we have a production site doing Veeam backups, can we replicate to a remote site without having Veeam at the remote site?
A: Yes, you can install Veeam Backup server in both source, and target site, or even in both - there is no extra charge.

Q: Can I backup or replicate a replica VM?
A: Yes. Keep in mind that this scenario has some limitations and side effects such as no file-indexing, no SQL point in time recovery, no CBT for backup, more complex restore mechanisms and most likely requires additional licenses.

[Gostev]

Migration

Q: Does Veeam quick migration require shared storage?
A: No.

Q: What is SmartSwitch?
A: SmartSwitch is the process of transferring VM's running state from source to target VM during migration via suspending the source VM, transferring the running state to the target VM, and resuming target VM. This allows to significantly reduce downtime by eliminating shutdown/bootup cycle. SmartSwitch requires that source and target hosts have compatible CPUs, otherwise cold switch is used.

Q: What is cold switch?
A: SmartSwitch is the process of switching from source to target VM during migration via powering off the source VM, syncing final disk delta, and powering on the target VM.

Q: What migration mode is used in what case?
A: There are 3 scenarios:
1. Migrating VM between hosts with VM sitting on the shared storage. If VMware license allows, VMware VMotion is leveraged. If VMware VMotion is not licensed, Veeam Quick Migration is used (with SmartSwitch or cold switch depending on source and target hosts CPU compatibility.
2. Migrating VM between storage within the same host. If VMware license allows, VMware Storage VMotion is leveraged. If VMware Storage VMotion is not licensed, Veeam Quick Migration with SmartSwitch is used.
3. Migrating VM between hosts and storages at the same time. Veeam Quick Migration is used (with SmartSwitch or cold switch depending on source and target hosts CPU compatibility.

Q: How does migration utilize Veeam’s vPowerNFS technology?
A: Migration jobs are specifically optimized to work with Instant VM Recovery. If the migration job detects that the VM being migrated is running off a vPower NFS datastore, it takes unchanged data directly from the backup file (rather than pulling it from the vPower NFS server) to maximize availability and performance of the vPower NFS server.

[Gostev]

VM Copy

Q: Does VM Copy support changed block tracking?
A: No. VM Copy is designed for ad-hoc full VM Copy of running VM in the native format, and is not a replacement for backups. This can be useful for copying some running VMs from production into the test environment, creating tape backup exports in native format, and so on.

[Gostev]

Full VM Restore

Q: I restore the VM, but my existing jobs processing this VM cannot see it?
A: Jobs tracks individually added VMs by unique IDs. The unique ID is only preserved in case of in-place VM restore (to original location) - any other restore option will register VM with the new unique ID. If you have to do such restore often, consider stuffing your jobs with containers (VM folder, cluster, datastore) for dynamic job scope, as opposed to adding in individual VMs.

[Gostev]

VM Files Restore

Q: I have restored a thin virtual disk, but it got "inflated" (converted to thick) instead of remaining thin.
A: Thin disks cannot exist outside of VMFS, and while not being attached to a VM. To keep the restored disks thin, or to restore thick disk as thin, use the Hard Disk Restore wizard instead of VM Files Restore wizard.

[Gostev]

vPower

Q: Do any of the vPower features depend on certain versions of VMware, vCenter, etc.?
A: No, there are no dependencies on VMware product versions. We support all supported vSphere versions.

Q: Do any of the vPower features depend on certain VMware license levels?
A: No, there are no dependencies on VMware license levels. But note that Storage vMotion is not included in some VMware editions. Having Storage vMotion makes Instant VM Recovery a smoother experience.

Q: Will any of the new capabilities require the Veeam Backup server to be run in a virtual machine?
A: No, this is not required.

Q: Do I have to store our backups on NFS datastore in order to be able to use vPower?
A: No, you can keep using whatever backup storage you are using today. vPower engine is backup storage-agnostic. If you can backup to that storage, it can be used with vPower. The only exception is Cloud Connect Backup which does not allow vPower today.

Q: How do I create the vPower NFS datastore?
A: You do not create it manually, the process is 100% automated.

Q: Where exactly does the vPower NFS server run?
A: NFS server can run on every Windows backup repository. Each Windows backup repository can use "local" vPower NFS server (recommended for better scalability), or vPower NFS server from another Windows backup repository (if having multiple vPower NFS servers is undesired). Linux, integrated deduplication storage appliances and SMB (CIFS) based repositories do not implement vPower NFS server and must be pointed to an existing one.

Q: What are the vPower NFS requirements on the infrastructure side.
A: Because VMware ESX(i) uses vmKernel interface to access NFS storage, ESX(i) host you are using for any vPower functionality must be able to communicate with the backup repository running the vPower NFS server.

Q: How does vPower deals with RDM disk? We use servers using RDMs.
A: vPower supports vRDM disks (virtual mode), but not pRDM (physical mode). Because snapshots are not possible on the latter, pRDM disks are automatically excluded from backup.

[david.domask]

Veeam Continuous Data Protection (CDP)

Q: What is CDP intended for?
A: CDP is intend to allow you to have an “up-to-date” replica of your most critical machines, providing granular “down to the second” recovery capabilities for workloads where even seconds of data loss are an issue.
By adding your most important machines to a CDP policy, you will always have a stand-by replica with granular recovery down to the second.

Q: Which license is required for CDP?
A: From our User Guide, CDP is included in the Veeam Universal License. When using a legacy socket-based license, the Enterprise Plus edition is required.

Q: How is Replication with CDP different than normal Replication in Veeam Backup and Replication?
A: Start with reviewing our User Guide on CDP; CDP is able to achieve it’s granular recovery by use of VMware’s vSphere API for I/O filtering (VAIO); Veeam installs I/O Filters into the VMware environment, and VMs assigned to the CDP Policies in Veeam will be replicated “live” based on your policy configuration in Veeam.

Q: When should I use CDP Policies and when should I use regular Replica Jobs?
A: For your most critical workloads where data loss of more than a few seconds/minutes is unacceptable, use CDP Policies in addition to regular backups to protect the workload.

For all other workloads, use regular Replica Jobs for fast recovery.

Q: Can I backup my VMs participating in CDP Policies?
A: For original production VMs: Yes, but wait for the Initial Sync of the CDP policy to complete before performing snapshot operations like a backup.

For CDP Replicas: No, backup operations will be prevented on CDP replicas.

Q: Can I migrate VMs in VMware/with Quick Migration if the VMs are participating in CDP replicas?
A: Both Host and Storage vMotion are supported on the original Virtual Machine.

Only host vMotion is supported on the CDP replica.

Q: What are the system requirements for a CDP proxy?
A: Our System Requirements are available in the User Guide

Q: Do I need a dedicated network for CDP?
A: While not a requirement, CDP shows the best performance on a 10 Gbit or faster network with MTU 9000; a dedicated network is recommended to avoid other workloads impacting CDP operations.

Q: Can I do manual snapshots on machines participating in CDP?
A: As of Veeam version 12.1.1.56, you can take snapshots of VMs participating in CDP.

Q: What Recovery operations are possible with CDP Replicas?
A: You can perform Failover and Failback operations, as well as GuestOS Item restores from CDP Replicas

Q: What is CBT Mode with CDP? Why is the CDP processing slower in CBT mode?
A: CBT Mode indicates that there is an issue impacting the transfer of the VM changes, such as network interruptions, underperforming CDP proxies, etc.

CBT Mode will typically self-resolve and return to live syncing if the interruption is temporary in nature; if the CDP Policies continually are in CBT mode check the network connectivity between the source/target VMware infrastructure and the CDP proxies, as well as the CDP proxies for any issues.