Some questions on VMware as Off-site recovery / backup solution

Discussion in 'Virtual and Cloud Computing' started by Mikeyboy, May 18, 2011.

  1. Mikeyboy

    Mikeyboy Kilobyte Poster

    279
    2
    32
    Hi all,

    Just thought I would quiz the knowledgable types here, regarding the likely upcoming virtualisation in work... Just a bit of background, there is no virtualisation in place at the moment, about 12 servers, and we will be looking to put in a blade centre with a few blades in a separate office as a failover / recovery type setup. I think the general idea of the people who have thought this up is to have a vmware setup in place, and if the need arises, restore from current livestates on to a VM and use this as until original is recovered, etc.
    I just wanted to check first of all, is this the best way to do things? or is there a way of sort of mirroring current setups of physical servers onto virtual copies, so there is an up-to-the-minute version?
    I have used VMWare quite a bit recently, started using it just to play with, but when I heard this might be a forthcoming project, I did my best to set up a home lab with some ESX hosts, vCenter etc. But didn't have the power to test it properly :rolleyes: so far from an expert..

    One thing I am unsure of, when I have experimented with P2V before, as soon as I powered on the new virtual machine, it complained about duplicate names on the network etc - is this only the case if the machine you cloned is still powered on? Assuming you had to give it a different computer name, presumably this would stop any name related links working, so I take it there is a way round this? Or is it not possible to have a virtual clone and physical machine on at the same time?

    My understanding is that we will put some blades in, install ESX on each, setup a vCenter server and manage it like that? And then when that is set up, P2V all our servers onto this new setup... and go from there...

    Appreciate if someone can just confirm that I am on the right path, also give me any pointers on how to implement this? I think I am likely to be given quite a major part of this project, so before we get any further I just want to make sure I have the foundations clear...

    Thanks :)
     
    Certifications: VCP,MCSA, MCP, MCDST, MCITP, MCTS, A+, N+
  2. onoski

    onoski Terabyte Poster

    3,120
    51
    154
    Hi Mike,

    Virtualisation is an interesting concept and also an interesting technology to work with in terms of flexibility and resilience. However, in your case you mentioned livestate backup? and am assuming you're referring to Symantec livestate backup recovery software.

    The physical servers can be cloned or an image taken using Symantec backup recovery software and then the image P2V to VMware vCentre etc.

    In reference to the computer complaining about a conflict with a hostname or IP address, this is as a result of the template that was originally used.

    Basically, a VM can be converted into a template and then that template can be used to deploy or create several other VM's.

    Hence, the template when deployed into a VM retains the settings from the original VM that it was cloned from.

    To be honest this should not be an issue as you can logon to the VM in question once setup and access the network card and change the static IP address to use DHCP or another static IP address that's not in use.

    All the best and keep us informed of how your virtual environment is coming on etc.
     
    Last edited by a moderator: Jun 15, 2011
    Certifications: MCSE: 2003, MCSA: 2003 Messaging, MCP, HNC BIT, ITIL Fdn V3, SDI Fdn, VCP 4 & VCP 5
    WIP: MCTS:70-236, PowerShell
  3. dales

    dales Terabyte Poster

    2,005
    51
    142
    Part of the point of virtualization is server consolidation, you would do well to point out that you will need to purchase new servers to roll out ESXi and virtualcenter (presumably with a San appliance too VNXE3100 possibly) and whats the point of keeping the expensive old tin running as the primary source on your network when the new shiny virtual infrastructure can replace all 12 servers with 2 or 3(as long as you buy a decent amount of ram for them) think of electricity, aircon, maintenance etc.

    What is your current backup proceedure for the physicals. You may well be able to keep that proceedure but more than likely there will be alternatives that could restore a borked VM (iffy windows update etc) in minutes rather than what it possibly is now!

    I take it MGMT have heard of virtualuzation but are not necessarily understanding the whole point of it!

    When you p2v a server it gives the server a whole new set of standard vHardward (included in that are network cards), windows remembers old IP address settings for cards that have been installed previously and will moan about it when you input the same details into a new card. Also as you are aware you cannot have 2 physicals with the same FQDN working reliably on the same network, the same is true for 2 virtuals or 1 physical and 1 virtual.
     
    Certifications: vExpert 2014+2015+2016,VCP-DT,CCE-V, CCE-AD, CCP-AD, CCEE, CCAA XenApp, CCA Netscaler, XenApp 6.5, XenDesktop 5 & Xenserver 6,VCP3+5,VTSP,MCSA MCDST MCP A+ ITIL F
    WIP: Nothing
  4. SimonD
    Honorary Member

    SimonD Terabyte Poster

    3,681
    440
    199
    I would have a look at Veeam Backup as a source for your backup and DR\Recovery solution, it's #1 in the VMware backup arena for a reason.
     
    Certifications: CNA | CNE | CCNA | MCP | MCP+I | MCSE NT4 | MCSA 2003 | Security+ | MCSA:S 2003 | MCSE:S 2003 | MCTS:SCCM 2007 | MCTS:Win 7 | MCITP:EDA7 | MCITP:SA | MCITP:EA | MCTS:Hyper-V | VCP 4 | ITIL v3 Foundation | VCP 5 DCV | VCP 5 Cloud | VCP6 NV | VCP6 DCV | VCAP 5.5 DCA
  5. zebulebu

    zebulebu Terabyte Poster

    3,748
    330
    187
    To do this 'properly', you really want to look at san replication and something like VMWare's Site Recovery Manager. Of course, 'properly' implies that you have a management that understands that to reap the benefits of virtualisation, you have to invest in it - rather than just chucking a bunch of old Powereges/DLs into a colo site and running up VM copies on GSX. I can't speak for MS' virtualisation offering because from what I've seen it's light years behind VMWare's in terms of manageability so I've never bothered with it.
     
    Certifications: A few
    WIP: None - f*** 'em
  6. craigie

    craigie Terabyte Poster

    3,020
    174
    155
    Just to echo what Zeb said you could look to use a couple of HP P2000 G3 MSA which are entry level and the HP Remote Snap software that will take about 64 snapshots a day off the SAN and replicate it off site. Which is about a SAN replication every 30 minutes or so.

    If you have the cash to spend you could look at HP LeftHands using the SANiQ to replicate the SAN's.

    Whatever you decide to do you need to remember that you have to have the servers in the DR site to use in case off an emergency. When planning these we would normally go for at least 3 x Physical Nodes at your Head Office with vSphere Advanced giving FT, HA and VMotion. In the DR site, depending on the size off the company, we migth use 2 x Physical Nodes and Essentials Advanced for HA and VMotion and to live without FT for the period off time we are in DR.
     
    Certifications: CCA | CCENT | CCNA | CCNA:S | HP APC | HP ASE | ITILv3 | MCP | MCDST | MCITP: EA | MCTS:Vista | MCTS:Exch '07 | MCSA 2003 | MCSA:M 2003 | MCSA 2008 | MCSE | VCP5-DT | VCP4-DCV | VCP5-DCV | VCAP5-DCA | VCAP5-DCD | VMTSP | VTSP 4 | VTSP 5
  7. craigie

    craigie Terabyte Poster

    3,020
    174
    155
    We recommend Veeam for Backup, but for DR I don't think you can go far wrong with SAN based replication or using MS technologies e.g. Exchange DAG, SQL Database Mirror/Cluster, DFS R etc.

    But there are lots of different ways to get to the same end result.
     
    Certifications: CCA | CCENT | CCNA | CCNA:S | HP APC | HP ASE | ITILv3 | MCP | MCDST | MCITP: EA | MCTS:Vista | MCTS:Exch '07 | MCSA 2003 | MCSA:M 2003 | MCSA 2008 | MCSE | VCP5-DT | VCP4-DCV | VCP5-DCV | VCAP5-DCA | VCAP5-DCD | VMTSP | VTSP 4 | VTSP 5
  8. ThomasMc

    ThomasMc Gigabyte Poster

    1,507
    49
    111
    [edited]

    Yeah just go virtual at the main site :).

    The only thing I can think of that is close to what your asking is Acronis but I don't think thats CDP and potentially messy when the main server comes back up

    p.s Veeam B&R is awesome :D
     
    Last edited: May 18, 2011
    Certifications: MCDST|FtOCC
    WIP: MCSA(70-270|70-290|70-291)
  9. Mikeyboy

    Mikeyboy Kilobyte Poster

    279
    2
    32
    Hi Guys,

    Just wanted to update on this thread really... Had a meeting today with an external supplier who demo'd and explained some VMware stuff, was quite pleased that I already knew pretty much anything they showed me, looks like I knew more than I expected!

    Just one initial question really - we are thinking about purchasing a second blade centre, and just putting a couple of blades in intially to start the virtualisation project - from what I saw today that would be more than enough, but my main question is, what sort of memory are we looking at needing, as a minimum, in each blade? We were told today that needed 32gb, but 64gb is better (obviously costs a lot more to double the RAM), but wondered if like, 16gb is enough? Or just stick with 32gb? I know it obviously depends on what we are virtualising, would initally be 4-5 servers, we have about 12 currently, but I dont want to limit our hardware spec initially. Obviously easy enough when everything is set up to just pull a blade out and bung some more memory in, but would rather just have an idea in my head first of all...

    Also, as we currently have a blade centre in place, we were planning to utilise some of these blades currently used to be used for virtualisation - but they all only have 2-4gb of RAM in each, so would it be as simple as sticking a load of RAM in them and they would be up to the job? I think they are all Intel Xeon quad cores, haven't checked if they are on the HCL yet or anything, more just asking the question first of all :)

    Thanks for all the replies so far BTW, really given me some good ideas to work with!
     
    Certifications: VCP,MCSA, MCP, MCDST, MCITP, MCTS, A+, N+
  10. Phoenix
    Honorary Member

    Phoenix 53656e696f7220 4d6f64

    5,749
    200
    246
    if they are modern blades they can likely handle considerably more ram, so the only upgrades would be RAM and mezzanine cards for I/O
    I would stick with 32 or 64 at a minium, 16 is pushing it for a few 64bit systems which we generally give at least 4gb to each
    doubling ram doesnt ALWAYS cost a ton more, i had a client that went from 48 to 96 for a small bump in price

    also make sure you are balancing ram right for the chassis, if thats a tripple channel setup you probably want memory in base3 size jumps, so 12, 24, 48, 72, 96 not the traditional 8, 16, 32, 64, and mixing dimm sizes makes very little sense just to get to '64gb'
    also make sure your mem is configured properly across processors
     
    Certifications: MCSE, MCITP, VCP
    WIP: > 0
  11. SimonD
    Honorary Member

    SimonD Terabyte Poster

    3,681
    440
    199
    Also may want to consider that if you need to carry out maintenance on a host you may want to ensure you have enough resources to vmotion between the other hosts, something you wouldn't be able to do if you limit the ram on each host. Always try and go for as much ram as possible, as Ryan says, sometimes an upgrade doesn't cost too much more.
     
    Certifications: CNA | CNE | CCNA | MCP | MCP+I | MCSE NT4 | MCSA 2003 | Security+ | MCSA:S 2003 | MCSE:S 2003 | MCTS:SCCM 2007 | MCTS:Win 7 | MCITP:EDA7 | MCITP:SA | MCITP:EA | MCTS:Hyper-V | VCP 4 | ITIL v3 Foundation | VCP 5 DCV | VCP 5 Cloud | VCP6 NV | VCP6 DCV | VCAP 5.5 DCA
  12. Mikeyboy

    Mikeyboy Kilobyte Poster

    279
    2
    32
    Yes that was my main reason for saying that I thought we need as much memory as possible - going to start putting together a plan soon i think :p probably going to get one new blade with a large amount of memory, and once its set up move over one of the servers which is on a decent spec blade, and upgrade that with max RAM, and then another of the blades once we free that up... should be enough to be going on with then :)... will report back as things progress...
     
    Certifications: VCP,MCSA, MCP, MCDST, MCITP, MCTS, A+, N+
  13. Phoenix
    Honorary Member

    Phoenix 53656e696f7220 4d6f64

    5,749
    200
    246
    another thing, are these blades connected to a SAN?
    i'm at the point where i just assume anything thinking about VMware is doing this, but i'm reminded what assumption means at least twice a week :)
    and I'm in California this month, so that may be higher ;)
     
    Certifications: MCSE, MCITP, VCP
    WIP: > 0
  14. Mikeyboy

    Mikeyboy Kilobyte Poster

    279
    2
    32
    Yeah they would be connected to a storage module in the blade center containing some SAS disks, connected over fibre channel, should be OK?
     
    Certifications: VCP,MCSA, MCP, MCDST, MCITP, MCTS, A+, N+
  15. craigie

    craigie Terabyte Poster

    3,020
    174
    155
    Don't forget that the limitation off the storage module in the Blade is the amount off space it will give you as ideally you want to mimic the drive speeds and amount off storage at your normal office.

    A SAN could be a better way to go depending on your storage needs.
     
    Certifications: CCA | CCENT | CCNA | CCNA:S | HP APC | HP ASE | ITILv3 | MCP | MCDST | MCITP: EA | MCTS:Vista | MCTS:Exch '07 | MCSA 2003 | MCSA:M 2003 | MCSA 2008 | MCSE | VCP5-DT | VCP4-DCV | VCP5-DCV | VCAP5-DCA | VCAP5-DCD | VMTSP | VTSP 4 | VTSP 5
  16. Mikeyboy

    Mikeyboy Kilobyte Poster

    279
    2
    32
    Craigie,
    I think we are just going to end up having a mirrored setup at both locations.
    Have just been doing some research on Veeam, sounds clever stuff :D

    Just trying to figure out a few things in my head first of all though, have just been thinking how I would actually kick off the whole virtualisation process... having read some stuff earlier, it seems that when I go through a P2V process, at least in the initial testing phase, I would want to put my esx hosts initially on a separate network, so when I convert my physical to virtual servers, there is no conflict with production server names / IP addresses etc, would this be something I set up in vmware, as in set up a virtual switch or something, to put it on a separate network? or would it need to be completely isolated, putting it in a DMZ or something like that?

    Also a similar question really, regarding how we would set up the backup / replication procedure, how is this tested usually? For example, lets say I have the VM infrastructure setup, virtualised my servers and they are being replicated to the DR site nightly (not a clue how this is set up, what is copied, where it is copied to exactly), is the only way to test, to do it out of hours and take the main server offline, power on the most recent replica and redirect any IP addresses etc?

    This project is likely to be kicking off soon, and I am just trying to get my head round the basics of things like Veeam. I think I could go ahead and virtualise some of our servers without too much grief at the moment but I appreciate there is (presumably) far more to it than just making a virtual machine...

    Also Phoenix, you mentioned further up "so the only upgrades would be RAM and mezzanine cards for I/O" can you elaborate on this, do I need to add extra NICs to the blades or something?
    Thanks :)

    Also edited to add, just looking at what I will need in the BladeCenter to make the SAN accessible, so if anyone knows about the bladecenter modules etc please reply as I need help!!
     
    Last edited: Jun 22, 2011
    Certifications: VCP,MCSA, MCP, MCDST, MCITP, MCTS, A+, N+

Share This Page

Loading...
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.