NetApp for the first time in the world implemented and patented the snapshot technology in 1993. Since then, the technology has naturally developed and debugged. After 23 years, you can safely say that NetApp snepsing works fine, without affecting performance, regardless of the protocol: block or file. With the advent of Flash technology, the IT industry has turned its attention to the use of snepshots and thin cloning due to the fact that Flash is still an expensive technology and its resources need to be saved. For many years, NetApp has single-handedly developed and carried on its shoulders
the backup paradigm built on snapshots, where snapshots were not a temporary necessity, but the basis of this whole idea, and finally received the recognition that this is an effective and correct approach to backup and development. And with the advent of Thing Propositioning and
UNMAP, snapshots have become really convenient to use and store as backup copies directly to the main, productive storage system, which allows you to quickly restore data after a logical error. And to back up and protect against the failure of the main storage system, use data replication technologies based on snapshots to replicate only the changes, not the full set of data.
An important advantage of snapshots is their ability to more quickly remove and restore without the need to drive traffic through the host and media server (proxy server). Neither the host CPU nor the network adapters suffer from overloads, the disk subsystem does not load at all in the case of recovery from snapshot, and in the case of recovery from the replica it loads much less than with Full Backup / Restore. This dramatically improves the backup picture: backups along with hardware snapshots are becoming possible to shoot more often, and right in the middle of the day.
But this whole picture is not complete without backup software. Since the captured and replicated data itself is not consistent from the point of view of the application, and without cataloging all the information, their application becomes much more complicated. NetApp works with many well-known backup software companies to integrate their products with storage capabilities.
In this article, I want to talk about possible
interaction schemes between NetApp and Veeam Backup & Replication for executing and storing backups of virtualized infrastructure.
')
Scheme: Storage of backups on the main system in the form of snapshots
This scheme provides for the launch of the product and storage of data on the ONTAP. In order for the
data cataloging function on the LUN to work, you must have a SnapRestore or FlexClone license . For the NFS protocol, there is no need for any license to work with content cataloging, and SnapRestore is desirable for data recovery on ONTAP, but if it is not there, the
built-in NDMP client will be used (for 7-Mode only). In the case of data recovery, it is highly desirable to have a SnapRestore or FlexClone license in order to instantly recover any amount of data from the snapshot. Without a license, recovery will be performed using the copy operation from the storage system to (possibly the same) storage system, through a media agent, which will take some time.
It is important to note that the bulk of data corruption is covered by this very scheme, since “logical damage” accounts for much more, compared to accidents or disasters that occur, but much less often.
The setup of this scheme is described in more detail in the
article Veeam Backup & Replication + NetApp ONTAP 8/9.
Store and restore backups on a secondary NetApp ONTAP system
The schemes described below provide for the start of production and storage of data on ONTAP and their subsequent replication to the second (third, etc.) ONTAP. As a rule, it is used together with the scheme “Storage of backup copies on the main system in the form of snapshots”. This allows you to more quickly restore data in the event of "logical damage" (hackers, viruses, etc.), since such damage is the bulk compared to disasters and accidents. The difference between archiving and DR is that, in the first case, recovery, as, for example, in the case of tape media, always occurs on the main system, to where the data was before (or to a new location, but you won’t run the tape from the tape) for example, also here). In the second case with SnapMirror, you can switch from the first (primary) to the backup (secondary) system, in case of an accident at the site, and continue working from there. To run Archiving or DR, you need a replication license, SnapVault or SnapMirror respectively.
To perform LUN cataloging on the primary or backup storage, you must have a SnapRestore or
FlexClone license on the corresponding site.
After the data has been replicated to the backup system, it is very convenient to start cataloging and testing there, in order not to burden the production with these operations. Again, from the second system, you can pull out the reserved data to put it on a tertiary system (NetApp or third-party). As with NFS and LUNs (for a SAN, you need a FlexClone or SnapRestore license), the data will be pulled from the backup system through a media agent. In the case of LUN, it will be copied using thin cloning, automatically connected to the media agent. In the case of NFS, nothing will be cloned, as Veeam, as mentioned earlier, with the help of the built-in agent, can use the NFS protocol to "climb" and read snapshots.
VSA / Public Cloud
The data can be replicated to either the physical NetApp FAS platform with ONTAP, or
Data ONTAP Edge ,
ONTAP Select, or to
ONTAP for Cloud located within the public cloud.
Scheme: Archiving (SnapVault) - restore to the main site
When using this type of replication, if the recovery is performed by means of storage, the restoration must be performed on the primary system. If it is not, then it will have to buy. In extreme cases, it is possible to convert SnapVault to SnapMirror, and only then restore the data.
Scheme: Disaster Recovery (SnapMirror) - recovering to a spare site
As a rule, it is used together with the scheme “Storage of backup copies on the main system in the form of snapshots”. This allows you to more quickly restore data in the event of "logical damage" (hackers, viruses, etc.). In the event of a catastrophe at the main site, a spare site can be converted to the main storage system. It is logical to be able to do this so that the main and spare systems are more or less equal in performance. Otherwise, it is worth asking in advance what most critical services will need to be launched and which ones will not be running. For this scenario,
SnapMirror for SVM can be used with one of the Identity Preserve or Identity Discard recovery modes. Identity Preserve saves network address settings when it is “relocated,” which imposes the need to stretch the L2 domain between sites and requires appropriate network equipment and channels. Identity Discard allows you to pre-set new network addresses (which will occur at the moment of switching to the secondary storage system) for which data will be available on the secondary storage system, this does not eliminate the need for specialized equipment, but requires reconfiguration of hosts manually or using scripts.
Data replicated using SnapMirror to a secondary (spare storage) can then be replicated to a tertiary NetApp Data ONTAP system using SnapVault.
Scheme: Filling backups on ONTAP - target archive repository with hardware dedup
In order for hardware dedup to work most efficiently, from the side of
Veeam, you need to configure “Dedup-friendly” compression policy in Job’s advanced settings, so that data blocks fit along the block-block boundary of the WAFL file system or disable produpny dedup in Veeam.
Veeam Direct NFS Client
Starting with Availability Suite v9 supported by Direct NFS, built right into the Veeam application (not to be confused with the NFS Client for Windows), it was designed specifically for ONTAP. This allows Veeam to communicate directly with the NetApp repository, allowing you to back up and restore data directly, bypassing the hypervisor, which speeds up such operations. An important useful feature is the ability to read existing snapshots from storage systems.
By the way, did you know that NFS in VMware appeared exactly the same, thanks to NetApp?
Starting with v9, the On-Demand Sandbox for Storage Snapshots product is supported, which allows you to run thin Hardware-Assistant copies of virtual machines for testing and developing them. For this functionality, you need a license that launches
FlexClone technology .
VVOL
VVOL technology will drastically change the backup approach of the virtual environment. One of the reasons for this is a fundamentally different approach of snooping and backup, since VVOL uses hardware snapshots, the issue of consolidating and deleting VMware snapshots is eliminated. Due to the elimination of problems with the consolidation of VMware snapshots, the removal of application-aware backup snapshots ceases to be a headache for administrators. NetApp is a technology partner and
VVOL design
partner, and supports this technology from ONTAP version 8.2.1 and higher . DR functionality for VP is also supported, which allows SnapMirror hardware replication to be used. Veeam also supports SnapMirror and VVOL hardware replication management (separately for each). So far, Veeam B & R does not support VVOL together with SnapMirror, but due to the promising technology, it will appear there.
Veeam: VMware's Virtual Volumes - Cannot I Deploy? .
NetApp AltaVault
Veeam perfectly knows how to interact with
AlataVault by performing backup
uploads for long-term storage on this system with all the above examples, i.e. act as a secondary or tertiary repository of archive copies. Since AlataVault acts as a file ball, Veeam always fills data there through a media agent.
findings
Backing up data using Veeam Backup & Replication together with hardware snapshots and thin replication allows you to make backups more often, protecting both from logical data damage (viruses, hackers, etc.), and from physical destruction, without loading the CPU and host network interfaces, completely bypassing them.
Veeam has specifically developed the NetApp ONTAP NFS agent to be able to read the contents of snapshots "directly", both for cataloging and for backup, which, in turn, indicates the deep integration of these products and the most optimal use of resources.
There are many replication and backup schemes, once again confirming the effectiveness and optimality of using the snapshot-based backup paradigm.
To get the greatest effect from both manufacturers, you should use the Disaster Recovery (SnapMirror) scheme - we are recovering to a spare site with the ability to test and catalog on a backup site.
Translation to English:
ONTAP with Veeam Backup & Replication
This may contain links to Habra articles that will be published later.
I ask to send messages on errors in the text to the LAN .
Comments, additions and questions on the article on the contrary, please in the comments .