-
- Novice
- Posts: 4
- Liked: never
- Joined: Nov 02, 2015 11:00 pm
- Full Name: Ari Weinstein
- Contact:
Architectural Recommendations Please - Planning Advice
Relatively new Veeam user here, looking for some planning advice. We have been using B&R 8 to backup a few Hyper-V VM's in our colo, and are planning a move to backup our whole infrastructure using B&R 9. Here are some details about our environment:
1) Hyper-V 2012 R2 using failover clustering, we have 2 nodes now, adding one more in the coming months
2) 10G LAN, and 10G ISCSI SAN, powered by Datacore. Separate ports and switches for each network
3) Exclusively Windows VM's now, there might be some Linux VM's in the future, but not a key part of what we do
4) We will probably replicate via VPN to another location initially, with Azure replication in the roadmap
5) Most VM's between 100-300GB, only a few are large (>1TB data)
6) Almost 40 VM's currently, growing by 1-2 VM's at a time as we add customers
7) Total data storage around 10-12 TB or so
My initial plan was to purchase a Dell R730XD, using at least 64GB RAM, 2xXeon CPU's, 2x10G NIC, RAID1 on 10K SAS or SSD for OS and applications, and RAID5 using 7200 RPM NLSAS for the repository. What are the recommendations for maintaining high performance for this system? Should all of the roles be installed on this dedicated appliance? One big repository, our should we split into more than one RAID group?
TIA-
-Ari
1) Hyper-V 2012 R2 using failover clustering, we have 2 nodes now, adding one more in the coming months
2) 10G LAN, and 10G ISCSI SAN, powered by Datacore. Separate ports and switches for each network
3) Exclusively Windows VM's now, there might be some Linux VM's in the future, but not a key part of what we do
4) We will probably replicate via VPN to another location initially, with Azure replication in the roadmap
5) Most VM's between 100-300GB, only a few are large (>1TB data)
6) Almost 40 VM's currently, growing by 1-2 VM's at a time as we add customers
7) Total data storage around 10-12 TB or so
My initial plan was to purchase a Dell R730XD, using at least 64GB RAM, 2xXeon CPU's, 2x10G NIC, RAID1 on 10K SAS or SSD for OS and applications, and RAID5 using 7200 RPM NLSAS for the repository. What are the recommendations for maintaining high performance for this system? Should all of the roles be installed on this dedicated appliance? One big repository, our should we split into more than one RAID group?
TIA-
-Ari
-
- Product Manager
- Posts: 6431
- Liked: 729 times
- Joined: May 19, 2015 1:46 pm
- Contact:
Re: Architectural Recommendations Please - Planning Advice
Hi and welcome to the community!
Some useful links:
- RAID considerations
- Best Practices for Hyper-V High Availability
- System requirements
Thank you.
It's hard to say anything without any info about your RPO and RTO objectives. Another significant thing that should be taken into account when planning a backup infrastructure architecture is backup method that you'd like to use. So please provide some extra information and kindly use our search engine.What are the recommendations for maintaining high performance for this system? Should all of the roles be installed on this dedicated appliance? One big repository, our should we split into more than one RAID group?
Some useful links:
- RAID considerations
- Best Practices for Hyper-V High Availability
- System requirements
Thank you.
-
- Veeam Software
- Posts: 21071
- Liked: 2115 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Architectural Recommendations Please - Planning Advice
We also have a best practices document (it is for VMware, but many things apply).
-
- Novice
- Posts: 4
- Liked: never
- Joined: Nov 02, 2015 11:00 pm
- Full Name: Ari Weinstein
- Contact:
Re: Architectural Recommendations Please - Planning Advice
We have redundant hypervisors and a redundant SAN, so most of our restores are basic file/folder recoveries. Of course, we are using Veeam so that we have the ability to recover the entire environment if necessary. We would like to be able to run multiple incrementals daily, possibly using the synthetic full option overnight. Also, the ability to "roll back" a VM after a failed upgrade or software install would come second after basic data restores. We would probably use forward incremental method, depending upon the system.
Would performance be helped by having a backup proxy in a VM on each hypervisor?
Are there differences between VMware and Hyper-V backup that I should be aware of?
Thanks,
-Ari
Would performance be helped by having a backup proxy in a VM on each hypervisor?
Are there differences between VMware and Hyper-V backup that I should be aware of?
Thanks,
-Ari
-
- Product Manager
- Posts: 6431
- Liked: 729 times
- Joined: May 19, 2015 1:46 pm
- Contact:
Re: Architectural Recommendations Please - Planning Advice
That is a subject to test. Anyway it's not a recommended approachfor large environments. It'd better to either use on-host proxy or to deploy a separate physical machine. Please take a look at this article.Would performance be helped by having a backup proxy in a VM on each hypervisor?s
Hyper-V takes a snapshot in a different way than VMware. You should keep in mind that VMs that have shared vhdx can be backed up in a crash-consistent state only. Also CBT data is stored on each node of the cluster, and if it is turned off, then VM backup job will do a full disk image scan.Are there differences between VMware and Hyper-V backup that I should be aware of?
Thank you.
-
- Novice
- Posts: 4
- Liked: never
- Joined: Nov 02, 2015 11:00 pm
- Full Name: Ari Weinstein
- Contact:
Re: Architectural Recommendations Please - Planning Advice
If we utilize the synthetic full feature, does ALL processing take place on the backup server (i.e. the "appliance" we are spec'ing out)?
Thanks,
-Ari
Thanks,
-Ari
-
- Product Manager
- Posts: 6431
- Liked: 729 times
- Joined: May 19, 2015 1:46 pm
- Contact:
Re: Architectural Recommendations Please - Planning Advice
Synthetic full backup load is splitted between backup proxy and backup repository. During synthetic full backup, Veeam does not retrieve VM data from the source datastore but “synthesizes” a full backup from data you already have in the backup repository thus the most part of load goes to repository, backup proxy is used to create a new incremental backup. Please see this article for details.
Thanks
Thanks
-
- Novice
- Posts: 4
- Liked: never
- Joined: Nov 02, 2015 11:00 pm
- Full Name: Ari Weinstein
- Contact:
Re: Architectural Recommendations Please - Planning Advice
Does transforming incremental into reverse incremental also run without needing access to the source datastore? It sounds like using forward incremental + synthetic full + reverse incremental is the best methodology for us.
-Ari
-Ari
-
- Veeam Software
- Posts: 21071
- Liked: 2115 times
- Joined: Jul 11, 2011 10:22 am
- Full Name: Alexander Fogelson
- Contact:
Re: Architectural Recommendations Please - Planning Advice
Transform is also performed locally on repository server. Though keep in mind it is pretty I/O intensive.
Who is online
Users browsing this forum: No registered users and 23 guests