ML
    • Recent
    • Categories
    • Tags
    • Popular
    • Users
    • Groups
    • Register
    • Login

    Virtualization and HA, Scalability

    Scheduled Pinned Locked Moved IT Discussion
    virtualizationscalabilityhigh availability
    73 Posts 12 Posters 8.3k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • K
      kelsey @Obsolesce
      last edited by

      @tim_g 0_1516817204394_Capture12.PNG thats what we have to do but the part i am on is the report which is this - In addition to the practical task, an individual report is to be produced on the application of virtualisation technologies to an organisation of your choice. The report should contain an introduction to the problems faced by organisations implementing highly available and scalable computing resources, evaluation of potential solutions, selection of a solution for the organisation chosen and discussion and critical evaluation of implementations with particular reference to the lab systems installed.

      DustinB3403D scottalanmillerS ObsolesceO 3 Replies Last reply Reply Quote 0
      • DustinB3403D
        DustinB3403 @kelsey
        last edited by DustinB3403

        @kelsey said in Virtualization and HA, Scalability:

        0_1516817204394_Capture12.PNG

        This is a picture of an IPOD. . .

        AKA Inverted Pyramid of Doom.

        If you lose your storage server for any reason, both physical host 1 and 2 are useless.

        K 1 Reply Last reply Reply Quote 2
        • K
          kelsey @DustinB3403
          last edited by

          @dustinb3403 thanks

          DustinB3403D 1 Reply Last reply Reply Quote 0
          • DustinB3403D
            DustinB3403 @kelsey
            last edited by

            @kelsey Do you know why if you lose the storage server the physical hosts 1 and 2 are useless?

            K 1 Reply Last reply Reply Quote 0
            • K
              kelsey @DustinB3403
              last edited by

              @dustinb3403 not really

              scottalanmillerS DustinB3403D 2 Replies Last reply Reply Quote 0
              • scottalanmillerS
                scottalanmiller @kelsey
                last edited by

                @kelsey so the report is based on NOT having HA? Then you have a HUGE opportunity to teach the class why the challenge is that even the professor failed to identify that this isn't HA!

                In fact, this is what we call an LA (Low Availability) solution. We have SO much reference material on this for you.

                https://mangolassi.it/topic/8822/why-dual-controllers-is-not-a-risk-mitigation-strategy-alone
                https://mangolassi.it/topic/8743/risk-single-server-versus-the-smallest-inverted-pyramid-design

                1 Reply Last reply Reply Quote 1
                • scottalanmillerS
                  scottalanmiller @kelsey
                  last edited by

                  @kelsey said in Virtualization and HA, Scalability:

                  @dustinb3403 not really

                  Because you have this huge network with all of these pieces, but all of them depend on the most fragile piece - the storage server. The storage piece, often a SAN or NAS device, is the most likely piece to fail out of all of them, and it is a piece that is completely unnecessary, and it is the one piece that is not protected in any way.

                  1 Reply Last reply Reply Quote 1
                  • DustinB3403D
                    DustinB3403 @kelsey
                    last edited by

                    @kelsey said in Virtualization and HA, Scalability:

                    @dustinb3403 not really

                    0_1516817951240_chrome_2018-01-24_13-18-20.png

                    Look at the image and the triangle I drew on it.

                    Based on the image, we have to assume that all Virtual Machine storage is on the "storage server". The Physical server 1 and 2 are running the hypervisor, and attaching to the storage server (iscsi or DAS or some other method).

                    In this case though the method doesn't matter.

                    If that storage server goes down, your VM files are inaccessible and the hypervisors cannot load and run them.

                    IE you're up shits creek without a paddle until the storage server is repaired.

                    K scottalanmillerS 2 Replies Last reply Reply Quote 1
                    • scottalanmillerS
                      scottalanmiller
                      last edited by

                      What your professor is proposing is know as the most standard scam in the IT industry - every vendor who thinks that they can trick someone because they don't understand risk uses this model as it is the most effective way to empty the pockets of those that try to "buy" their IT rather than doing their IT themselves.

                      It creates a dependency chain, for no reason, and every layer completely depends on all other layers, with the most dangerous layer having no protection at all.

                      http://www.smbitjournal.com/2014/11/the-weakest-link-how-chained-dependencies-impact-system-risk/

                      1 Reply Last reply Reply Quote 1
                      • K
                        kelsey @DustinB3403
                        last edited by

                        @dustinb3403 i get it now thanks

                        DustinB3403D scottalanmillerS 2 Replies Last reply Reply Quote 0
                        • DustinB3403D
                          DustinB3403 @kelsey
                          last edited by

                          @kelsey said in Virtualization and HA, Scalability:

                          @dustinb3403 i get it now thanks

                          This also means, if the switch that the storage server goes down, you're in the same boat.

                          Which this leads into the conversation that @scottalanmiller just brought up which is dependency chain.

                          1 Reply Last reply Reply Quote 0
                          • scottalanmillerS
                            scottalanmiller @DustinB3403
                            last edited by

                            @dustinb3403 said in Virtualization and HA, Scalability:

                            @kelsey said in Virtualization and HA, Scalability:

                            @dustinb3403 not really

                            0_1516817951240_chrome_2018-01-24_13-18-20.png

                            Look at the image and the triangle I drew on it.

                            Based on the image, we have to assume that all Virtual Machine storage is on the "storage server". The Physical server 1 and 2 are running the hypervisor, and attaching to the storage server (iscsi or DAS or some other method).

                            In this case though the method doesn't matter.

                            If that storage server goes down, your VM files are inaccessible and the hypervisors cannot load and run them.

                            IE you're up shits creek without a paddle until the storage server is repaired.

                            Right, in this example, if the storage server fails, EVERYTHING fails.
                            Also, if the switch fails, EVERYTHING fails.
                            Also, if both hypervisors fail at the same time, EVERYTHING fails.

                            So there are three layers of potential failure. Two of them have no protection at all. Not only do they have no protection, they serve no purpose. Why is there a switch here? No idea, it's totally unnecessary. Why is there a storage server here? No idea, it is totally unnecessary. Those would never exist unless a sales person thought that they could scam the customer and really took a risk at not getting caught. (But we see this EVERY day. It's awful.)

                            Simply by removing those points of risk (there is no reason to replace them) we can take 99.99% of the risk out of the system.

                            1 Reply Last reply Reply Quote 0
                            • scottalanmillerS
                              scottalanmiller @kelsey
                              last edited by

                              @kelsey said in Virtualization and HA, Scalability:

                              @dustinb3403 i get it now thanks

                              But does your professor?

                              K 1 Reply Last reply Reply Quote 1
                              • K
                                kelsey @scottalanmiller
                                last edited by

                                @scottalanmiller dont think so

                                DustinB3403D scottalanmillerS 2 Replies Last reply Reply Quote 2
                                • DustinB3403D
                                  DustinB3403 @kelsey
                                  last edited by

                                  @kelsey said in Virtualization and HA, Scalability:

                                  @scottalanmiller dont think so

                                  This is where you get extra credit then.

                                  Take the same picture and cut it up into the triangle but in 4 pieces. Explaining that any piece breaks the chain, and thus the system as a whole.

                                  The client system is just sitting there saying "WTF I can't work" and is just a distraction.

                                  Only the servers and switch matter in that picture, and its a perfect example of what not to do when designing server systems.

                                  K 1 Reply Last reply Reply Quote 1
                                  • scottalanmillerS
                                    scottalanmiller @kelsey
                                    last edited by

                                    @kelsey said in Virtualization and HA, Scalability:

                                    @scottalanmiller dont think so

                                    This is the fear with universities. Typically the professors are pulled from the ranks of those that failed in the field.

                                    If your professor implemented something like this to one of my customers, we'd recommend legal action for having either actively been involved in a scam (on the fiddle) or being professionally negligent. At best we'd demand immediate termination and removal as unsafe. But we'd discuss legal action with a client, as it is a failure so heavily documented and exposed that no one in a position of making that recommendation today can reasonably claim to not have known, without claiming to have been so untrained and unskilled as to have knowingly put the company at risk by faking their ability to do the job.

                                    1 Reply Last reply Reply Quote 1
                                    • K
                                      kelsey @DustinB3403
                                      last edited by

                                      @dustinb3403 what do u mean 4 pieces i am dyslexic

                                      DustinB3403D 1 Reply Last reply Reply Quote 0
                                      • scottalanmillerS
                                        scottalanmiller
                                        last edited by

                                        This article is about SAN, but really applies equally to all external shared storage in the manner that your professor proposed:

                                        http://www.smbitjournal.com/2013/06/when-to-consider-a-san/

                                        1 Reply Last reply Reply Quote 1
                                        • DustinB3403D
                                          DustinB3403 @kelsey
                                          last edited by

                                          @kelsey 0_1516818797484_chrome_2018-01-24_13-18-20.png

                                          Pretty much in any case of any piece of the triangle fails, you lose services.

                                          Take any 1 away and something is gone / not functional until repaired or replaced.

                                          1 Reply Last reply Reply Quote 0
                                          • scottalanmillerS
                                            scottalanmiller
                                            last edited by

                                            Here is a video talking about why everyone is trying to take advantage of businesses, by trying to sell them a SAN that they clearly have no need for.

                                            Youtube Video

                                            1 Reply Last reply Reply Quote 1
                                            • 1
                                            • 2
                                            • 3
                                            • 4
                                            • 3 / 4
                                            • First post
                                              Last post