View Storage Tiering

Hello, im a little new for VDI.  VDI is already in place and im picking up things.  Currently there is only one still deployed and we want to start using storage levels

What we have

5.5 vCenter

Hosts Esxi 5.5 (5 guests)

VMware View 6

Storage

Table of NetApp - there are 3 Flash SSD storage, each 1 TB LUN.  There are two 500 GB LUNS each of 15 k SAS Drives.  And three LUN 1 to each Sata storage.  We also have lots of Sata storage, we can add, if necessary, might be able to add more SAS is necessary.

Ive been going about things and he uses is not Sata or SAS storage.  Here using only the SSD storage that worries me, because we rely on the addition of several pools and SSD is at a premium.

We use linked Clones and currently, there are two pools.  Each cluster is a floating pool so no need to keep the data of the user and they refresh overnight.  A swimming pool is only 15 desktop computers and the other is 120 workstations.

The two pools were a little Proof of Concept pools and its actually affected the production now.

Moving forward the plan is to add more pools, maybe another VM 400, as of this moment the new VM would float pools so no need to keep the data of the user.

I want to Storage Tiering as there is no budget to buy more storage SSD.

I know that I can split the storage so that the replica on the SSD storage and the BONES and the other files are on a slower storage.

So now, a few questions

1. my understanding is replica should be on tier-1 (ssd), the operating system and other files can be on the slower storage, but it would take on SAS 15 k disk storage or can it be the slower storage Sata? I hope that Sata

2 lets say, I find myself with 5 swimming pools.  IM assuming that I can have multiple replicas can on the same logical unit number? or should I be split my LUN by Replica?  This means that I have to cut my LUN differently

3. There is a limit of about 128 VM per LUN, how it works during the introduction of Storage Tiering?  IM assuming that it is not serious, the replica would be on the SSD LUN and then the rest would go to the SATA or SAS LUN disks, and I could have up to 128 VMS on the Sata/SAS LUN?

4. the VM per LUN 128 means that I could use the LUN more?  I currently have 5, but do not know if we will use the SAS LUN.  So I need to add a couple more Sata Lun?

5. If best practices says 128 VM per LUN.  Currently my LUN is all the same Raid Volume.  Is how important it?  What is VMware says to split up the LUN because its assuming the LUNS will be on separate Volumes of Raid or just because of filesystem VMFS itself?

Im going to be read allot more on VDI but I thought ide goto the forums first to see what think

Thank you

You need not split by replica LUNS. I would put all of your replicas on SSD, and yes they can share a LUN through many swimming pools.

You'll just want to have multiple LUNS that helps to balance the pool (s) on. Create as many LUNS you want to spread things, but really thinking about time zones and IOPS / s. 128 is a guideline, but not a rule lasts. It's all about having enough IOPS / s to serve requests you will get during peak usage.

I'd go with SAS if you can, but SATA is acceptable - just all this affects the user experience. Consider using SATA for task workers or people that do not generate a lot of use of storage. Trial and error and acceptance by the user would be good to come through here if you do not yet have.

Tags: VMware

Similar Questions

  • Best survey practice: tiering with different types of storage drives

    Hello

    We recently received a new SAN Equallogic, and I am deciding the implementation. Our first unit (currently only one member in a swimming pool) is a PS4100 with 24x600GB 10K SAS. The new unit is a PS4210 with 12x4TB 7.2 K NL SAS.

    Should I leave units in separate classes in the same group and cut new LUNS? Should I add the PS4210 in the existing pool and use the Storage Tiering?

    I'm a little hesitant to how Storage Tiering will work, I read a little when the mixture of types of disks, storage tiering will be balanced on the ability, and most of the i/o load will hit the higher capacity of SAN.

    Any contribution is appreciated. The storage is used for the storage of data VMFS with vSphere 5.5 cluster. Run EQL firmware 8.1.0 on both units.

    Hello

    Either will work.   As a general rule, speaking of small members 10K / 15K mix with very large 7.2 K members RPM can reduce the effectiveness of the fastest disks.  As more data on the largest Member.  This is especially true if they carry the same RAID type.

    I would start with them in their own pools.  Later, you can merge two pools quite easily.  The other way is more complex.  Since you need to remove a member, leaving the data in the original pool.  Then you can move volumes of the new pool.

    Also, please check Dell TR1091.  It explains how to configure the storage w/EQL ESXi.

    en.Community.Dell.com/.../Download

    A small change that has not been added that TR is the NOOP_TIMEOUT value.  The document shows you how to change the LOGIN_TIMEOUT value to 60.  In this menu is the NOOP_TIMEOUT, which gets the value of 30 seconds.

    Finally, make sure that the virtual machines themselves are configured with their disk values set on 60 seconds Timeout.

    With the EQL Firmware download is a PDF entitled ' iSCSI initiator and optimize your SAN environment for high availability of the OS considerations "which explains how to make the necessary changes.

    Kind regards

    Don

  • Boot volumes

    I'm a few SSDS on my SAN and want to make sure that my Vm dΘmarrez since their.  Someone know where I can find documentation on this?

    Thank you

    Then really look into storage tiering in View 4.6. http://myvirtualcloud.NET/?p=1137

  • Add the new Equallogic system

    Hello

    We use 2 PS4000 systems in a group.  Now, we wanted to add additional capacity to the group, and it seems that we must add a PS6100 model, because the PS4xxx series can have 2 systems per group.

    What is the best way to develop our Equallogic system? PS4xxx systems can be added to a group of PS6100 lossless or start over?  Or does first to transfer all data in the system of PS6100 and reset and add the PS4xxx later?

    good documentation on this subject?

    Thanks in advance

    Tom

    Yes, you can add the additional PS6100, which will remove the limits of PS4000 configuration.

    You can add it as its own swimming pool and then move the volumes on the new Member, or put it in the same pool as the PS4000.

    First pass of the PS4000 at least 5.2.5 before adding in the PS6100.   Then proceed to PS6100 to make.  I think he's not coming with 5.2.5 installed.

    If you let me know what is the size / speed / RAID type of the PS4000 and what size/TPM in the PS6100 I can offer you some general tips.

    The Administrator's guide has information about the management of members and pools, there is also a report of tech Storage Tiering

    en.Community.Dell.com/.../19974554.aspx

    Kind regards

  • Overview in vSphere Web Client?

    Hi community

    In vSphere Client c# , I had the chance to see an overview of all virtual machines and their snapshot spaces: home - inventory - warehouses and store data from cluster - storage tab views.

    By selecting my data center on the left, I had a glimpse of all the VMs on a data center and size of snapshot (on all data warehouses). See my attached screenshot.

    I just upgraded my vCenter to 6.0U2, so I did not tab views storage more.

    Does anyone know how to proceed with the Web Client vSphere? I don't want to check every vm or each data store but would like to have a similar overview of instant spaces of all the virtual machines on all data warehouses.

    Best regards

    Roland

    Version 6 does not offer this view more, see this KB article: VMware KB: The Storage Monitoring Service plug-in installs in VMware vCenter Server 6.0

    In VMware vCenter Server 6.0:

    • The selection of storage reports tab monitor an object is no longer available in the Web Client vSphere 6.0.
    • Tab views storage is no longer available in the vSphere Client 6.0.

    As I said before, you have the possibility to use RVTools or something like PowerCLI and vRealize Operations Manager, see this thread another talking about the same thing: the storage vsphere6 reports

    We were using display of storage reports to audit for forgotten shots. So far, there are some alternatives to see the snapshot size

    Use of Ops 1 Manager) and a heatmap personalized, note the Foundation (for some free version) does not have the feature. As we use Foundation, it is not an option.

    (2) RVTools - GUI based (easier for some) but pretty basic

    (3) PowerCLI - yup, get used to it. (get - vm | get-snapshot |) FT VM, SizeGB, created, name, Description)

  • Aggregation period default dimension

    Hello

    I just have a question, I noticed that any member with type summary period in the period dimension such as quarters or Yeartotal or half year if necessary are grouped by default even if they are storedata. and accordingly data can be entered with these parents and is divided between the months as well using the distribution.

    So when do we need to have parents as dynamic calc? -What for the view storage or if the users are not demanding enter data into the yeartotal or quarters? or there is no other perspective?

    And is a property of the periods summary of dimension period they can aggregate although they are to store data, like other dimensions can not do it.

    Thank you.

    Hello

    as explained, "it's a feature on the data form, not in the underlying database Essbase."

    Planning knows the logic of aggregation, since it is built in the planning and then pushed to Essbase. If the logic on the data form don't logic and recorded Level0 members but also the parents calculated.

    The nice feature, you will save an aggregation (if all the data is loaded through the data forms).

    I hope that answers your question.

    Kind regards

    Philip Hulsebosch

  • Reconnecting to vCenter to address

    Hello

    I have installed my Vmware View server and database event today. Added my vmware vcenter server.

    In my follow-up-> I keep seeing events:

    Gravity: Info

    Module: Connection to the server

    Message: vCenter to the address https://VMware:443 / sdk is in place

    and

    Gravity: Info

    Module: Connection to the server

    Message: reconnecting to vCenter to the address https://VMware:443 / sdk

    all 3 to 10 seconds. Everything seems to work ok. But why so many logging?

    We use the heartbeat of vmware our vcenter server is it has something to do with it?

    Products:

    VMware ESXi 5.5

    VMware View 6.0.1

    VMware Heartbeat 6.6

    VMware vCenter 5.5

    With Greetings:

    Sander

    Okay, I think that I fixed it. He had something to do with view storage Accelerator and Global-> Bill, like the vCenter Server access rights.

    When you create a virtual machine, he must do something with the digest of virtual disk.

    Sander

  • Location of the Datastore VM on a Cluster of DS.

    Good day everyone.

    Is there a way in the client to see wich datastore within the dscluster of each virtual machine is located.  After the configuration of the cluster of the ds on the Summary tab storage information display the name of the DS cluster only.

    Thanks to advance.sto

    Yes it is possible, select your virtual machine and click the tab views storage.

  • Remove access to esxi san 5.1 when dummylun is introduced

    We have a san EMC CX4 Clarion connected to several 5.1 ESXi hosts.

    The following procedures in kb2004605 we mounted successfully, detached and physically entered the esxi host LUNS.

    There is only a single dummylun 0 b 0 connected to the esxi host 5.1 of this san for now.

    We will remove the EMC CX4 Clarion completely san.

    Has the dummylun 0 be detached from the esxi host 5.1 before you remove us from the Clarion san

    or this lun disappears when the san network access is removed by the 5.1 esxi hosts?

    I put out of service without (HP EVA) with their pseudo-LUN 0 in both directions without real issues (except error messages can be ignored that ESXi lost access to the pseudo LUN 0), it does not really matter in the end, but it's cleaner and I prefer to follow the same procedure to remove the LUN as in http://kb.vmware.com/kb/2004605.

    Then run esxcli storage device together--State = off [ID] d to detach the controller.

    After you run this command, the controller device must show upwards as not mounted in the view storage device and you can area/trash the old SAN.

    On the next rescan the device should disappear entirely.

    Then you must also remove it from the list of devices paired with:

    basis of detached esxcli storage device delete d [ID]

  • How do you know any particular LUNS attached to hosts how and what hosts?

    Hello

    any command?

    If you have a particular LUN selected, click on the tab views storage and click on the cards.   You can then customize the map to display only the data store and host or store data and VMs or any number of combinations. Also by clicking on the tab configuration will show you just the connected hosts.

  • an internal error occurred failed to serialize the answer

    I get this error message when I try to view the storage card:

    an internal error occurred failed to serialize the answer

    The following is in the release notes for vCenter 5, maybe it's your solution:

    • With the active FCoE software, attempts to view storage cards fail with an error message
      This problem only affects the ESXi hosts that have been added to vCenter Server without any configuration of FCoE software previous. After you enable the software on these hosts FCoE adapters, attempts to display the storage cards in the vSphere Client fails. The following error message appears: an internal error has occurred: Cannot serialize the answer.

      Solution: Configure the software FCoE on the host ESXi first, and then add the host to vCenter Server.

  • RAW Device Mapping (RDM) (data store)

    Hello everyone,

    I would like to ask I datastore, RDM, I use it with some VMS but when I browse with vsphere on my storage I Ididn find listested on my datastor

    could you please sometimes why not found Ididn.

    Thank you all for your support

    Used as the RDM disks do not present as data stores in the view "storage." There you will find only the VMFS LUNS data warehouses are in format. However, all LUNS - any that they are used as VMFS or RDM data warehouses - should show when you select the vmhba appropriate in "storage adapters".

    André

  • Access minimum vCenter for user registration capacity IQ

    So literature, I read about the deployment of the CAP IQ 1.5, I understand that you must create a new role with the privileges specific 6 global rights and privileges of the extension.  This may be a completely useless question, but I want to give the domain user that I created for this task only the necessary minimum access level.  What (if any) are the specific privileges of vCenter as the needs of users in to obtain the measures, or are - this completely there?  These 6 privileges will give own nine full access of the user to the metric?

    Yes, it should work. A correction though; You must assign privileges 7, not 6. The 7 privilege is "views View Storage". The release notes will soon be updated with this information.

  • credentials do not work to the web/browser interface

    I installed ORacle XE 11.2.0 on win7 64 bit professional version (using a virtual machine).
    Logs show no errors. lsnrctl server seems ok, tnsping XE ok, connect sys as sysdba using sqlplus works.
    If I try to 'get started' on my browser (firefox latest version) starts with the url:
    http://127.0.0.1:8080/apex/f? p = 4950:1:3808355349156377
    and 3 offer buttons to view storage, sessions and settings and another to begin with request.
    All of them need credentials, but each of them returns "invalid credentials" if I use 'sys' or 'system' with the same password
    I used to properly connect to the database via sql.
    Hopefully, an explanation is little work for you, but it would be a big jump for me.
    Gerd

    user2734556 wrote:
    I installed ORacle XE 11.2.0 on win7 64 bit professional version (using a virtual machine).

    A 32-bit client? Note that XE 11.2 for x 64 should be released.

    If I try to 'get started' on my browser (firefox latest version) starts with the url:

    And if you try from another browser?

    All of them need credentials, but each of them returns "invalid credentials" if I use 'sys' or 'system' with the same password

    If different/pick-up of browser does not help, you might change the system password as follows:

    sqlplus / as sysdba
    
    sql> alter user system identified by  ;
    

    Published by: orafad on August 21, 2012 17:39

  • VM with a snapshot of 900 GB!

    Hello everyone!

    Some information:

    (1) we use ESX 3.5 Update 4 with vCenter 4 Update 1 on Netapp Storage connected via NFS for the ESX servers.

    (2) I have a Win2K3 VM who had a D: drive which was a 900 + GB hard disk that I migrated all the files to a connected iSCSI volume.

    (3) before I migrated the data, I took a snapshot of the VM, has added improved iSCSI vNIC and it configured in the operating system. And an another snapshot was taken just in case.

    4) 900 + GB of data has been copied to the iSCSI volume.

    (5) VM has been turned off to remove the 900 + GB D: drive.

    (6) VM power and using the iSCSI ok disc.

    Unfortunately I forgot to commit 2 shots... and a month later, I noticed that the VM is projected in vCenter as having an instant 923 GB!

    The tab view storage for the cluster shows the size of the snapshot!

    Research in the computer virtual folder, the delta files are not nearly this big:

    Looking at the volume via the console, files do not display a snap 900 GB:

    So vCenter think that the virtual machine has a 900 GB snapshot because I removed the 900 GB D: drive hard file in the VM properties?

    I really don't want to wait for 3 years to commit this snapshot!

    So, theoretically this VM has only really Go 2 x 3 buttons pressures correct? And shouldn't take very long to engage, if she?

    Ideas as to why please VCenter think this VM has a large snapshot?

    You must be really happy to know that you do not have a 900G instant. Commit (Group) the other two instant you can clone the virtual machine that is the safest way.

Maybe you are looking for