I have been doing research on multiple VM (windows) backup/replication software like Veeam, Appassure, Unitrend, and Zerto.Â
They all have one issue: if you want fast recovery at a remote site during a disaster, you have to have a live backup or standby VMs  at the remote site. Â
The problem would be that the standby VMs would require storage resources. That means the person who bough the software will need to pay premium fees in order to have a fast recovery.
If you don't have standby VMs ready, it would take around 24Â hours, or more, to export all the replicated VMs, depending on how many VMs you have.
Does anyone know if there is any software that has a faster recovery time during disaster without having standby VMs?
Thanks.Â
We have recently implemented Zerto as a disaster recovery solution. Not only does it replicated, it allows you to created boot orders for the VM's. This means you can have your SQL server start and set an automatic delay for the applications servers. It can also be configured to bring a failover test up on an isolated network for DR testing. It can also replicate between different hypervisors. For example, you can replicated from VMware to Hyper-V or AWS. They are adding more hypervisors with newer released.
So does that mean you want to have a Disaster Recovery solution where data is not on site your bunker site? but yet allows for a fast recovery in case your primary site is down?
- What virtualization solution do you use?
- What is the link between the 2 (?) sites?
- What RPO and RTO are you aiming for?
- How much data do you need to recover?
If you don't have live backup? Well as per my understanding backup is always a happened at local site (DC) on VTL and or on Tape and they were offloaded to out of DC, but as mentioned correctly it can take 24hours or more depend on the Recovery site location, accessibility & final is data size. Now the correct terminology is Online Replication or Archive/log base replication, and it is completely depend on the RPO & RTO define by business. So, answare to your 1st query : No way you can do a site recovery if you don't have DR site. Many says to take a back on tape, on disk or on storage but if all these product are installed at production site i.e. DC, will not make any sense as your DC is down and not accessible. So, "must to have live back or rather Replication to DR site.
2nd question" fast recovery without VM in passive or standby mode at DR site. VMware has SRM which does the site recovery in case of disaster. Only condition is that you have to have a Storage with replication between the site. Other option as mentioned by Mr. Smith, is DR as a service model (DRaaS) from any cloud providers. Some of the Cloud service providers also offers CDP solution while not charging for DR site but conditions is DC must be hosted with them.
Tested used my own little setup for hyper V machines have an offsite server using altaro backup offsite server backup software with windows server
restored (anywhere) the Virtual machine was up and running within a 10min entire server
I would also recommend to use Vision DoubleTake at VM level dat has an CDP , continous data protection feature for filesystem replication and SQL integration also. It can be a choice of synchronous replication over DWDM lines if latency it not excceding 0,5 ms round trip, otherwise it will impact disk write ops.
If zero downtime is a must I would recommend using VPLEX,ViPER from EMC or HDS Global Active Device that will present disk LUN from SAN as a single device to more processing nodes, but thus means app is aware of SW clustering (can run in multiple nodes sharing the same filesystem ir SAN LUN).
In such approach in VMware ESXi you will present a datastore spread over DWDM like a strech cluster so you won't have to keep in mind where the app node is really running, the hypervisor will see the strech cluster as only one storage device, thus means you can move app with vMotion very fast to a second or DR site, or recover it to a DR site. More if app is SW cluster enabled then the app nodes will run seamlessly over strech cluster.
The 2 nd option I can see is to go for Hyperconverged infrastructure and application containerization just like Docker tehcnology. How to do it: for ex. Make use of technologies like VxRail appliances and OpenStack + app transformation in Docker (for Windows VM is not so complicate). Such technolgies will apply private Cloud technology for DR.
Hi there, we are talking about Recovery from DR site, now few suggestions from my side 1) what is the defined RTO & RTO. 2) Visibility of the RPO. 3) connectivity between two or three site to meet replication requirements. 4) DR for physical & virtual, both the environment. 5) how many time in a year do the DR Drill. These point need to think and perform to achieve desire & accurate recovery from DR site.
Hi you could try Arcserve UDP -> Instant VM.-
With Veeam you can do Instant VM Recovery at the DR-site. You don't need a standby VM as it will boot the VM's you require directly from the backup storage. It takes typically 2 minutes to boot a Windows Server, so you could expect a restore time of 2-3 mins in case of total loss of the VM(s).
This can be done at a Service Provider that offers this service (restore your backups to their infrastructure). They would just import your Veeam Backups to their own Veeam console and run Instant VM Recovery on them. BAM - everything up in a matter of minutes.
But I would rather look into the DRaaS possibilities in Veeam Cloud Connect - much more streamlined and smooth solution that works out of the box.
Disclaimer: I'm a Veeam Systems Engineer.
Regards,
Sean
IfI understand correctly the guy needs a fast recovery solution for the production environment to a remote site, for Windows VM under VMware ESXi (or Hyper-V).
In my understanding a DR site means an alternate location with hot or cold standby systems, the recovery plan for business continuity is depending on their RTO and RPO.Unless an RPO and RTO are defined for IT services noboby could picture o solution for such cases. In general solutions are dependent of TB of data to be assured on remote site, basically there are many practices for assuring storage space in DR in case you would need to recover:- cold backup with ESXi that sustain test and development environment physically placed in DR, in case fast recover is mandatory, they could destroy the test/develop environment and restore data from scratch with VTL replicated in DR (backup and restore with 4TB/hour or more). The single point to be assured is correct IP addressing (test/develop could be treated as untrust zone and separated with VLAN and/or firewalls). You can use data protect and snapshots for VM, backup to tape, replicate virtual tapes and restore in case of a disaster (full recovery)- hot backup means CPU and storage for backup DR purposes but can be more faster, but cost a lot of money $$$$$$- rent some storage space and CPU from Cloud vendors, use as they need, maybee the DR location can be in the Cloud provider Data Center but data confidentiality can be a showstopper.
My proposal is to investigate the 1st option with fast backup of data snapshots (space efficiency if dedupe or data compression are available at production site at storage level) and sent them to a restore solution at remote site (virtual backups), restore ops must be tested from time to time to validate business data (not only apps).For fast backups you can try VTL or NFS appliances that include replication services, the bandwidth between sites must accommodate fast delivery to remote site (to assure that RPO and RTO, including restore times are met). I would not recommend a SW solution to replicate VM because if no storage is existing in DR dedicated for this purpose it make no sense to think on such solutions.The 2nd option if to address disk space and CPU needed with Cloud providers, otherwise disk space for VM and user data must be assured always in DR.
Hello,
I suggest taking a look at VMware - Actifio, It might be an option for the
environment you are working at. The minimum data backup for Actifio is
10TB. If your environment smaller than 10TB it will not work.
Regards,
www.actifio.com
Hello,
Check the VBA module from networker legato software, it might be what you need.
Kind Regards,
John.