Three servers esx or two esx servers?

We currently have three esx (dual dual-core, 16 GB) servers that are connected to a network of Hitachi SAN. We are trying to replace these servers with replacements and we have two options to meet our requirements about 12 GHz of CPU and 64 GB. We run around 20 virtual machines that are a combination of windows servers running exchange and file and print and redhat running oracle and jboss.

Option 1: 2 x superiors servers dual QuadCore with 32 GB each

Option 2: 3 x lower servers dual QuadCore with 24 GB each

Both servers are the same model but different specifications, one is a high-end server. Three servers operate at about 20% cheaper.

We will reuse the FC HBA, we have already three ESX licenses and we manage servers using VC, so there is no additional cost to the three servers (except the maintenance cost each year on an esx server).

The advantages I see to have three servers are;

  • 12 hearts as opposed to 8 cores

  • redundancy n + 2 for base servers

  • 66% of capacity left after a failure instead of 50%

  • servers cheaper and cheaper maintenance as collateral can be reduced by one day to the next

  • more easy upgrades to new versions of esx server can be taken offline and still leave redundancy

  • bandwidth of 50% has increased to SAN

  • bandwidth increased by 50% to network (18 versus 12 NIC NIC

I really don't see any downsides to having three servers?

My boss however like the idea of two servers and I need to convince him, if anyone has information that would help in my case I'd be very happy.

On the surface, it's easy to say, come on 3 Server... as you well pointed out the technical benefits. However, it is easy to underestimate the cost of this 3rd server. Of course, you have a sunk cost in the Vmware licensing, but you again year 2 and 3 cost. It is the same with the server, support costs for 3 years, power and cooling for another 3 years etc.

This probably isn't a huge number, but your boss is probably thinking in this direction. Maybe you need to hit a quick worksheet to calculate these costs, and then compare these costs and benefits that you mentioned. Until you do this, your boss will probably have his way

I would focus on availability and additional capabilities. Capacity is easy to align costs, so start there. The availability is important, but more difficult to quantify.

Good luck.

Dave

Tags: VMware

Similar Questions

  • Three hosts ESXi and two Clusters VSA

    Is it possible / is there a problem with it?  I know a single cluster VSA is designed to work with two or three nodes, but is it possible to have two clusters VSA on three nodes (one of the three nodes would be two VSA groups on both nodes at the same time).  I know there is a limitation that requires an instance of vCenter by VSA cluster, but I was unable to find mention of this type of configuration.

    Thank you.

    When the VSA cluster is created, a single VSA appliance is installed on each host ESXi (2 or 3 ESXi hosts) of the VSA cluster.  The VSA appliance uses local storage of the ESXi server. If you try to create an another cluster of VSA, it won't use an ESXi server that is already a member of a cluster of VSA.  In other words, a single VSA device can be installed on a server ESXi.

  • Iconia S/N is not found; support can be contacted; machine worked for three months. == No, two.

    Can anyone tell me please where I can find the S/N for the Iconia Windows Tablet? Or a way to contact support. A way to contact support. The machine was a payment for a trial in collective action and worked for two months and now won't turn on. Knowledge database tells me that I can find on a label (no label) or turn it on. If I could turn it on, I wouldn't need support.

    And Acer seems to give me no way to contact them about this product. More frustrating than precious: here's your machine: it won't work within two months, and we will not give you a way to contact us. Good day.

    Sorry all... I go back to the first question: any way to get a serial number out of this thing, or otherwise contact acer support?

    [edited to conform to the guidelines]

    I think that yours is probably W4 or W3. They both have a microSD and microHDMI microUSB slot.

    I think there are three distinct possibilities. The battery may become too discharged in storage to be able to load, the charger may be bad, or the battery may be bad. I can't remember if these models will directly run on the charger or not, but it is looking at the possibility. If they're going to sink the charger then we have reduced to the charger (possibly charge connector).

    If they will not work with just the charger then you test the output of the charger. You can use a wire wrapped around the probe on you VOM to reach the center of the connector of the charger.

    In my view, there are videos from Youtube with the info on battery recharge/boost one too discharged, but I would seriously look into your security situation before trying. No need to get your fingers inside if it is replaceable in the warranty.

  • Three-node cluster lose two hosts

    In a knot three cluster is it possible to lose two hosts and keep all three data warehouses are still online?

    lol if you lose two guests you don't have redundancy and disconnected data warehouses

  • Adding more than one host ESX and the upgrade at the same time.  What is the order of preference of the operations here?

    We are budgeting for more licenses and hardware for 2014 and want to do immediately in January.  I am trying to decide what is the best order of operations for adding two vmware hosts additional to our current installation of vsphere and also upgrade our vsphere / esx 4.1 u3 for the latest and best.

    We have 3 Dell PowerEdge R710 running ESX 4.1 U3.  In this vmware cluster, there is a vcenter server 4.1 U3 virtual machine handle this.  The three servers are about 90% memory used so we add 2 additional servers.  Probably the Dell PowerEdge r.620 since we get up 2 in the space of a 710.  Storage is on a san EMC NX4 NFS which will become too but not yet, let's just deal with the vmware part.

    When adding additional r.620 2, should I install ESX 4.1 on them and join them to the existing cluster and THEN pass vcenter and each one esx?  I ask precisely because 90% of use on 3 current servers, I need 2 other servers to support the update operations take place.  They would be able to repel vmotioned vm while each host receives an upgrade at the same time.

    So which is the best way to go about adding of the hosts and the upgrade?  Would I add as 4.1, then they would be available as targets vmotion and so I have room to start the hosts one upgrade at a time?

    If I update vcenter to the latest version, it is comparable backwards and can manage the 4.1 hosts, but also hosts 5.x?  I guess version 5.5 is the last one I would get from VMWare?  It is right of ESXi, then, how is this ESX ESXi 5.5 4.1 upgrade?  How much storage space do I need for ESXi 5.5, as I'm dechatoiement out of these r.620 Dell and need to know what hard drive to put in.

    The steps that you must follow

    -Updated vCenter Server

    -Upgrade of the ESX hosts (because we do a updating of the material we will install 5.5 on our new guests and add them to the cluster and then dismantle our existing)

    -Upgrade VMWare tools

    -Update data warehouses

    ESXi 5.5 has these storage requirements:

    • Installing ESXi 5.5 requires a boot device is at least 1 GB in size. When booting from a local disk or SAN/LUN iSCSI, a 5.2 GB drive is necessary to allow for the creation of the VMFS volume and a scratch 4 GB partition on the boot device. If a smaller drive or logical unit number is used, the installation program attempts to allocate a region scratching on a separate local disk. If a local drive is not found, the scratch partition (/ scratch) is located on the disk virtual ESXi host, linked to the/tmp/scratch. You can reconfigure /scratch to use another drive or logical unit number. For best performance and memory optimization, VMware recommends that you leave no /scratch on the ESXi host ramdisk.
    • To reconfigure / scratch, see Configure the Scratch Partition of vSphere Client vSphere Installation and Installation Guide.
    • Because of the sensitivity of the I/O devices, USB and SD, the installation program does not create a partition scratch on these devices. Therefore, there is no tangible benefits to the use of the great features of USB/SD that ESXi uses only the first 1 GB. When installing on USB or SD, tent Setup devices to allocate a region scratching on a local disk or the data store. If no disk is local or data store, /scratch is placed on the virtual disk. You must reconfigure /scratch to use a persistent data store after the installation.
    • In Auto deploy facilities, Setup attempts to allocate a region scratching on a local disk or the data store. If no local disk or data store is found /scratch is placed on the virtual disk. You must reconfigure /scratch to use a persistent data store after the installation.
    • For environments that start from a San or use Auto Deploy, it is not necessary to assign a separate logical unit number for each ESXi host. You can co-locate areas scrape for several hosts ESXi on a single LUN. The number of hosts assigned to a single LUN must balance the size of the LUN and behavior I/O virtual machines.
  • Upgrading to ESX 4.0

    Thanks in advance for any advice.  I am the IT Manager at a school district and I am planning to upgrade to Vmware 5.1.  I currently have three servers ESX 4.0 and Vcenter 32 bits on a physical machine servers.  Since I have to migrate to a 64-bit platform and Vcenter ESX to ESXI, I was thinking about a facility own would be the best approach, but I don't know which would be the most effective way.  At first I thought something like this:

    1. Create a virtual machine for the new Vcenter server on one of the 4.0 hosts.
    2. Installing Vcenter 5.0 and migrate settings from the Vcenter 4.0 (is it necessary or do I just start with a fresh VM Vcenter?)
    3. Remove the three 4.0 cluster hosts and add them to a cluster of 5 Vcenter
    4. Upgrade or install clean each host to 5.0 (can I vmotion all my VM for the two other boxes and still all keep it up)
    5. To update everything to 5.1

    I am open to any suggestion.  Also, how the update affects data banks? Volumes need to be updated to 5 VMFS.

    Thanks again for any help.

    In this case, I'd probably do a clean install of all components. First release of one of the hosts and disconnect/remove it from the current Server vCenter. Then update its firmware (if applicable), install and configure ESXi 5.1 and then configure vCenter Server 5.1 (unless you need Update Manager, you can consider the vCenter Server Appliance).

    Once done turn off HA/DRS on the old server vCenter Server, disconnect/remove hosting it and add the host to the new vCenter Server server. If guests have similar hardware, you should be able to vMotion the VMs and this way a rotation of hosts resettlement.

    André

    PS: Regarding VMFS. Although it is possible to upgrade VMFS3 to VMFS5 on the fly (after that all hosts are running the new version), I usually tend to create new warehouses of data and migrate virtual machines rather than the upgrade of VMFS.

  • Licensing ESX 3.5 ESX 3.5 Standard Foundation

    Hi all, my problem is related to a license file is not ineligible three servers ESX 3.5 just install. I have just 20 days left for the evaluation period, and so I asked my colleagues to contact a store and buy the necessary licenses. So they did, and they sent me a license for ESX Server 3.5 Foundation edition file. So I launched my VMWare Infrastructure Client, by connecting to one of the three hosts without a license and try to download the license file. The tab configuration, selected licensed features and I can see the Source license evaluation and 20 remaining days and under ESX Server Edition, I can see ESX Server Standard.

    So first of all, I tried to change this information, but I don't have the choice, in the pop-up window, there are two options, without license and ESX Server Standard and the first is turn off, and the other is already selected! If no change is possible!

    OK, I clicked on change near Source license and chose the last option, use a host license file. Searched the front received file and clicked Ok: a popup me has warned on deactivation of ESX Server assessment that would make the function Consolidated Backup no longer be available, but it's OK, since this is a test environment, and we are not interested in this feature. So I clicked Ok and received an error message saying "downloaded content license was not valid.

    I'm sure that this license file is from license VMWare web site...

    Is it possible that during the installation of the ESX Server a choice made on the Standard version, whereas now I can not use a license of the Foundation?

    Hope someone could help us, as we count the days left...

    Davide

    The Foundation licenses grant you a license Virtual Center.  Install Virtual center and install the license and then direct your guests to the license server.  Establishment licenses are not allowed because they come with coveted e Virtual host-based product Center.

  • Adding 2nd NIC killed ESX Server

    All,

    Thank you for taking the time to read my post. Here are the environment and then my question.

    3 - DELL R510 - Dual Core Quad - Intel systems

    The area has "2" Integrated NIC

    as well as "2" QUAD Port - PCI Intel network interface cards

    NIC 0 - integrated NETWORK card is a system where installed- and it is plugged into a standard Switch CISCO GIG, the tag VLAN 108

    NIC 1 - integrated is configured for iSCSI connectivity

    LOCATION 0 - Port Quad PCI NIC

    Port # 3 is connected to a GROUP of CIRCUITS on a CISCO SWITCH with VLAN 108, 104, 107 and 110 on it

    Port # 1 is used for the 2nd iSCSI connection

    LOCATION 1 - Port Quad PCI NIC

    # 7 port is connected to a different GROUP of CIRCUITS on a CISCO SWITCH with VLAN 108, 104, 107 and 110 on it

    All three servers are installed with ESX4 update 1

    When I go into vCenter to one of the three servers and go to the primary Switch virtual networking, if I add #3 NIC or Port #7 NIC to the virtual switch (it currently integrated NIC 0 in) the system disconnects the network and to connect to vCenter losses and the entire network. If physically I reboot the ESX Server, it is NOT on the network. I have the tag VLAN in NIC0 defined 4095 to accept all the VLANS as well. (I tried with and without the tag of VLAN 4095 - same result)

    Here are my questions:

    1 - Why is - what happens? How it should be configured?

    2. is there a way to get NIC0 to return to the network without having to re - install ESX?

    I reinstalled the boxes several times now and tried many variations of combinations of nic, and they all end up with the ESX Server disconnected from the network and does not.

    Really looking for some common sense here... Please and thank you.

    Hmmmm, shooting a little in the dark, but I think your next 2 steps are:

    (1) get the SC (vswif0) MAC and check your MAC address table switch to fly throughout of ports

    (2) on the console, use esxcfg-vswitch for uplink associated with the Group of port SC the first list, and then delete the 2nd network card.  This should get your SC direct return and leave the new map available NETWORK the vswitch, simply not group ports SC.  (And you don't have to reinstall :-).)

    Virtualization of happy!

    JP

    Please consider awarding points for correct or helpful answers

  • Rebuild the ESX - Dell 2950 Server

    Hey,.

    We have three servers (Dell 2950) running ESX 3.5. One of them seems to act weird and must want needs to be rebuilt. All our servers are connected to a San and we VC installed on another server.

    I was wondering what is the best way to rebuild this server? Never done this before and I don't want to lose any VM etc. Also is there anything I can follow recommended doco?

    See you soon,.

    Mark Bayliss

    DRS is enabled on your cluster? If this is the case, then place the host you want to regenerate in maintenance mode. Computer virtual running on it will be vmotioned off to another ESX host in the cluster. Once the host is in maintenance mode you can turn it off. Remove the host to virtualcenter

    Disconnect the cable to the SAN from the back of the server, or make sure SAN LUNS are not presented to the host through the SAN switch (assuming that the fibers). Disconnecting cables is the safest option

    Rebuild your media host directly or through DRAC virtual media.

    Once rebuilt, return to Virtual Center and add the host; I tend to add again but not in the cluster. Reconnect the cables of your SAN and re-run the analysis for your SAN storage space.

    Once you have the SAN re-presented to the host, then move back to the cluster.

    If you found this helpful, please consider awarding points

  • 2012r2 servers not report to WSUS after installing these updates

    After the installation of the following list of updates, all three of our 2012r2 (two physical, a virtual) servers have ceased under the WSUS running on a machine 2008r2. Before going through the process of uninstalling each of these updates one at a time, I wanted to see if someone had experienced the same thing. There is no error message about the service update on any of the servers 2012r2. The notification only I see is in the interface WSUS where he said the machines were not reported in since these updates have been installed. The updates have been installed on different days on the machines, but once installed, they stop immediately notification in.  Here is the list of updates:

    The WSUS is version 3.2.

    The latest updates in 2012r2 came down the pipe to the WSUS server. After being approved, the servers 2012r2 all found new updates from the WSUS server, install them, and the WSUS server relates now that they have all checked.

    This seems to have been a temporary problem.

  • State "Unknown" to the storage controller on one of our servers, PowerEdge R730

    Hello

    I'm having a problem with one of our servers. I tried to add all our servers PRTG to monitor the State of health and the hard drives. On two of our three servers, it works absolutely perfect, on our third-party server, all the disks are red because 'the component status' is 'unknown '.

    Then I checked this server iDRAC interface and noticed, that under the "storage->-> health and properties of physical disks" the status field contains a question mark for each hard drive. Under 'Storage-> controllers' fields status for 'Health and Properties' and 'Battery controller' it shows an exclamation mark.

    On our other servers, all statuses are all green. Version of firmware and driver are exactly the same on all three servers. I don't know if it's a hardware problem or if it's just a setting error.

    Any ideas?

    Hello.

    What version of OpenManage Server Administrator, OMSA do you use on a server? You are able to view the status of controller and physical drives in the BIOS of the controller, using Ctrl + R? What version of firmware BIOS, H730P and iDRAC?

  • Essentials Plus with 3 servers more than 192 GB?

    So basically I'm in the situation where each of the three servers ESXi 5.1 has 32 GB of memory.  I wanted to go to 64 GB however after asking Dell for technical brief for this.  We bought the memory and it got here, but I can't use the modules to obtain the memory needed (24 4 sticks).  Because each server requires a pair of 2 or 3 pairs.  So basically what I've discovered, is that each server would be cheaper to spend 72 GB instead of 64 GB at this point.  However 3 x 72 = 216 GB.  We know essentials plus only supports 192GO max that I read.  Then, how they are managed in what concerns each machine with 72 GB instead of 64 GB.  Two servers will show 72GO and the third 48GB?  Knows also distributed memory 64GB each on the three server?  Anyone run into this situation and what happens?

    In 5.1, the limits of the vRAM have disappeared - by socket only licensed

  • The upgrade to vsphere 4 essentials for vsphere 4.1 essentials more with 2 servers?

    I currently have a Dell 2950 III with processors x 5460 double hearts running vsphere 4 essentials and I am looking at upgrading on two servers with vsphere 4 essentials more with vmotion. How do I know if I can continue to use my existing 2950 and add a Dell R610 with two hexagonal x 5680 core processors? I want to be able to run most of my guests on the new R610 but power move them back to the 2950 (which could be changed during the upgrade) and if necessary R610.

    Thank you.

    The only concern, you'll need to know is the processors for R710, R610, and PE 2950. You must ensure that enhanced Vmotion is supported for these inorder to three servers for vmotion (live migration) seamlessly.

    I know R610 and R710 works, but I'm not sure what that generate a PE2950 you are referring. You can see the matrix here. http://KB.VMware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalID=1003212 an another prereq is that all of these VMs must be in the shared storage and the shared storage is presented to all these 3 hosts.

    iDLE-jAM | SC 2, SC 3 & VCP 4

    If you have found this device or any other answer useful please consider useful or correct buttons using attribute points

  • Thunderbird has stopped the connections to the servers.

    I'm having a strange problem with Thunderbird. Since yesterday, I can not connect to servers or receive, send or save drafts. What is happening to all three accounts, IMAP (Gmail two, a 3 rd part), both from home or at work. Email accounts work on other devices. There must be something to this PC and Thunderbird installation. It does work in Windows Safe Mode & Thunderbird Mode without failure either. It did not help uninstalling and reinstalling. I'm using version 45.5.0 on Win7. I see the following: the status bar again : connecting to ... Receive messages does not expire, it will stay in the connection... condition. Save a project said: there was an error saving message in drafts. Try again? Said after start-up error console: could not read the chrome manifest ' queue: / / / C:/Program % program % 20Thunderbird 20(x86) /Mozilla % 20Files / chrome.manifest'.

    While the creation of the category 'Profile after change' services, could not create service entry "Sapi speech Synth", contract ID "@mozilla.org/synthsapi; 1.

    Discover the folders for the account failed with exception: [Exception... "Component returned error code: 0 x 80550016 [nsIMsgFolder.subFolders]" nsresult: "0 x 80550016 ()" location: "JS frame: resource:///modules/MailUtils.js: MailUtils_discoverFolders: line 36" [data: no] "

    Timestamp: 29/11/2016-19:40:29 Error: NS_ERROR_FAILURE: component returned error code: 0x80004005 (NS_ERROR_FAILURE) [nsIMsgFolder.isServer] Source file: resource:///modules/FeedUtils.jsm line: 230

    Proxy... It's the proxy settings.

  • LB vs Webcomponent switch servers OAM

    I have three servers of identity. I can do what follows from the Webpass

    (1) I can report two primary identity servers and the last Fail-over.
    (2) - or - I can point to three identity-servers as servers.

    To test whether the failover works for point # 2, I dropped one of the servers identity and the webpass still works. So, what is the advantage of creating a failover server? Anyway, failover is already supported when all servers are used to balance the load. Am I missing something?

    Thank you!
    Kabi

    The two techniques are valid ways to provide redundancy.

    Multiple primitive cancers is a way of distributing the payload and increase in capacity (descaling).

    Addition of failover is generally to ensure that you can maintain a capacity of minimum service in case you start to lose primary resources. I suggest you read up on the setting of 'failover threshold '.

    It is possible that you run with more capacity you need and that you do not configure failover of resources. It is also possible that, in a deployment dispersed geographically, and cross-configure the failover of the resources of the remote sites. It is a very flexible system.

    Hope this adds useful information to your thoughts.

    Mark

Maybe you are looking for

  • ThinkPad SL510 does not load

    Look's like if I had a big problem: (0 first sorry for my English) 1. when I plug the charger in to my lights SL510 lit. If the batteries are out - computer laptop is impossible to market;2. If I plug the same charger to my second (old thinkpad r60 -

  • What is AT-GPIB card Rev.E2? Search is empty!

    What is AT-GPIB card Rev.E2?  Search is empty!  Reverend E2 means TNT or plug-and-play or something?

  • Same product key of Windows7 on different computers mobile, CQ56-219WM and CQ60-615DX. Is this true?

    I bought two laptops Compaq Presario used, a CQ60-615DX CQ56-219WM.  I used the ProductKey software to get the product key of Windows7 Home Premium for each machine.  The keys for the two machines were the same.  Is this good?  Can I depend on these

  • I have problems to open a pdf file

    I need to open these files, but do not work.http://www.CIC.GC.ca/English/PDF/kits/forms/IMM1295E.PDFI did the upgrade of the new wersion but nothing. I have a windows 7.My PDF wiever is not able to display this type of document. but is urgent to open

  • How do we use the SUBSTR function in OWB

    Hello, I'm pretty new to OWB. I need to use the SUBSTR function in owb. My source table has 16 characters in length. But in my datamar,.the length is 14 years. I want to the substring from source to the destination. The mappings are already developed