Page 2 of 2 FirstFirst 12
Results 26 to 47 of 47
  1. #26
    Hmmm. Can someone please confirm how MS licensing works?

    Can I not grab a copy of DC 2012 and a bunch of regular CALs and RDS CALs for each server for a flat fee licensing model?

    Or must I use the SLPA to license 2012? What else would I need in addition to this?

  2. #27
    Join Date
    Aug 2011
    Location
    Dub,Lon,Dal,Chi,NY,LA
    Posts
    1,839

    Serious Cloud Hosting Foray

    Only SPLA can be used in a hosting environment. MS rules are very specific.

  3. #28
    Do you mean that I don't need to purchase DC edition?

  4. #29
    Join Date
    Aug 2011
    Location
    Dub,Lon,Dal,Chi,NY,LA
    Posts
    1,839
    Microsoft can give you more details (or your regional SPLA vendor)

    The short version is that you can buy SALs and OS licenses per VM, or you can buy DC edition per hypervisor (but still buy the SALs). All costs are monthly under SPLA.
    HTH

  5. #30
    Join Date
    Oct 2001
    Location
    Ohio
    Posts
    8,535
    Quote Originally Posted by dediserve View Post
    Microsoft can give you more details (or your regional SPLA vendor)

    The short version is that you can buy SALs and OS licenses per VM, or you can buy DC edition per hypervisor (but still buy the SALs). All costs are monthly under SPLA.
    HTH
    Yep - if you need any help with Microsoft licensing, we can put you in touch with someone over there. We have a few connections that can help you out.

  6. #31
    Join Date
    Nov 2012
    Posts
    79
    Quote Originally Posted by cartika-andrew View Post
    based on your earlier comments - overselling CPU seems to work well - even our most packed hypervisors arent coming near to saturating CPU - but, I wouldnt over-allocate ram. I mean, why bother. If you are wasting CPU (which is what you are paying for license wise anyway), suck up the one time costs and go really dense with RAM per node and dont bother over-allocating it - you really dont need to, and your customers will thank you
    Being new to colocation and Onapp, I have been quite pleased that I made a decision to go with Onapp.

    I do have a question though. How do you know if you are coming near to oversaturating your cpus? Memory usage is easy to see but trying to understand how much of the cpu is being utilized on average is eluding me. I know Onapp provides some numbers but I really do not understand what they mean.

    Thanks for any assistance.
    Brian

  7. #32
    Quote Originally Posted by brianemwd View Post

    I do have a question though. How do you know if you are coming near to oversaturating your cpus? Memory usage is easy to see but trying to understand how much of the cpu is being utilized on average is eluding me. I know Onapp provides some numbers but I really do not understand what they mean.
    Hi Brian,

    the hypervisor graphs and data I have seen dont seem to be onapp generated. looks like we are using 3rd party solutions to monitor the overall hypervisor health. its why we started going more dense (ie up to 256GB RAM per hypervisor at this point) - our CPU utilization was never over 20% on the hypervisors, and this seemed like a serious waste of resources. I have not seen the data on the newer hypervisor nodes with the increased RAM density - but, I would expect that CPU utilization metric to go up. I am hoping for 50%, which to me, would indicate a proper utilization metric for that resource (which is over allocated) when compared to RAM, which is ultimately the limiting resource. Worth mentioning, IOPs are something we track as well, and I have never seen an issue with SAS/SSD drives - especially if you have adequate number of spindles. On SATA drives, you really need to monitor shared usage properly and identify/move certain users to higher IOP spindles. Other then that, from what we have experienced - if you arent over-allocating RAM, and you are properly building out nodes, IOPs and CPU, for the most part (obviously always exceptions), are fairly maintenance free
    www.cartika.com
    www.clusterlogics.com - You simply cannot run a hosting company without this software. Backups, Disaster Recovery, Big Data, Virtualization. 20 years of building software that solves your problems

  8. #33
    Join Date
    Nov 2012
    Posts
    79
    Thanks Andrew. If you don't mind can you share that 3rd party solution? Feel free to PM it if that suits you better. I am running 64GB of ram right now and I run SAS drives in an Equallogic SAN. It just seems the CPU utilization would be my blind spot.

  9. #34
    Hi Brian,

    sure, drop me a PM and Ill check with our team to see what they are using thats generating those pretty reports I see

    I can say that we run both local disk (ie what we call VPS) and SAN storage (ie what we call cloud). CPU metrics seem pretty consistent. If you are using modern quad/hex/octa core CPU's - I cant see you having a CPU utilization issue with 64GB RAM (obviously just a wild guess without knowing your environment - but, assuming "typical" usage, you should be fine). we would typically run dual quad/hex/octa core - we obviously improved our base CPU spec, each time we increased ram density.

    anyway, drop me a PM and Ill get you some details

    cheers
    www.cartika.com
    www.clusterlogics.com - You simply cannot run a hosting company without this software. Backups, Disaster Recovery, Big Data, Virtualization. 20 years of building software that solves your problems

  10. #35
    Join Date
    Aug 2011
    Location
    Dub,Lon,Dal,Chi,NY,LA
    Posts
    1,839

    Serious Cloud Hosting Foray

    We use zabbix which also lets you distribute the monitoring proxies across multiple clouds and locations.

  11. #36
    Quote Originally Posted by dediserve View Post
    We use zabbix which also lets you distribute the monitoring proxies across multiple clouds and locations.

    I can tell we arent using zabbix - but, spent the last little while looking at it - looks like a really nice product. in addition to other things Im sure it replaces, do you use this in place of nagios?

    anyway - thanks for this
    www.cartika.com
    www.clusterlogics.com - You simply cannot run a hosting company without this software. Backups, Disaster Recovery, Big Data, Virtualization. 20 years of building software that solves your problems

  12. #37
    Quote Originally Posted by cartika-andrew View Post
    I can tell we arent using zabbix - but, spent the last little while looking at it - looks like a really nice product. in addition to other things Im sure it replaces, do you use this in place of nagios?

    anyway - thanks for this
    So Whichever monitoring software you are using is accessing the HVs via one of the private networks?
    Mike S.
    NetCloudTrue Cloud Servers
    OnApp Cloud ★ Enterprise HA SAN ★ GNAX Atlanta

  13. #38
    Quote Originally Posted by NetCloud_VM View Post
    So Whichever monitoring software you are using is accessing the HVs via one of the private networks?
    I knew we were using nagios - I just didnt know what was producing the reports - its an opsview install along with some other graphing utilities which pull data from nagios...

    our install is distributed between our nocs and data is pulled via our WAN and Management networks
    www.cartika.com
    www.clusterlogics.com - You simply cannot run a hosting company without this software. Backups, Disaster Recovery, Big Data, Virtualization. 20 years of building software that solves your problems

  14. #39
    Anyone has a good 2U SAN to recommend? Preferrably iSCSI over 10G

  15. #40
    Join Date
    Oct 2001
    Location
    Ohio
    Posts
    8,535
    Quote Originally Posted by webhost4all View Post
    Anyone has a good 2U SAN to recommend? Preferrably iSCSI over 10G
    Take a look at the Dell EqualLogic. Should fit your needs.

  16. #41
    Join Date
    Oct 2004
    Location
    Earth
    Posts
    463
    Quote Originally Posted by webhost4all View Post
    Anyone has a good 2U SAN to recommend? Preferrably iSCSI over 10G
    Consider getting a second SAN and using active/active replication. Nothing worse then having your SAN down and the phone calls start coming in.

  17. #42
    Just my two cents - if you've got deep pockets, look at VCE. They build a tightly integrated hypervisor/compute/storage stack with a great management interface to rule all of it.

    At my current position, we currently use VMware, Cisco UCS, Cisco Nexus, and EMC VNX for our setup, vCloud director for provisioning, and custom integration into our internal portals for chargeback. We have VM's that run the gamut from 2GB per VM to 128GB per VM. Storage from 40GB per VM to 5TB per VM.
    ZFS Storage Blog - http://www.zfsbuild.com/

  18. #43
    Join Date
    Nov 2012
    Posts
    79
    Quote Originally Posted by IGobyTerry View Post
    Take a look at the Dell EqualLogic. Should fit your needs.
    Second on the Dell Equallogic. I just purchased one and so far I am been very happy with it.

    Brian

  19. #44
    Quote Originally Posted by webhost4all View Post
    Anyone has a good 2U SAN to recommend? Preferrably iSCSI over 10G
    What is your budget, capacity requirement and IO requirement?
    Mike S.
    NetCloudTrue Cloud Servers
    OnApp Cloud ★ Enterprise HA SAN ★ GNAX Atlanta

  20. #45
    Quote Originally Posted by NetCloud_VM View Post
    What is your budget, capacity requirement and IO requirement?
    Something that serves 500 VMs so 10G or multiple 10Gs are preferred. Capacity is a free float, something expandable. Is there such a thing as a clustering SAN?
    Budget is flexible.

    Comparison would be against OnApp storage with 3x2TB RE4 per node (as that's the maximum each blade node can take)

  21. #46
    I would recommend you look into ZFS SAN from Nexenta. But this is a Dedicated SAN though nothing like OnApp Integrated Storage.

    Paul at Pogo Linux is a good contact that can maybe help you more on ZFS SAN depending on your requirements. http://www.pogolinux.com/support/contact.php
    Mike S.
    NetCloudTrue Cloud Servers
    OnApp Cloud ★ Enterprise HA SAN ★ GNAX Atlanta

  22. #47
    Anyone tried using Velociraptors? Entertain me for a while but this is a last ditch attempt considering my blade nodes can only proc 3xSATA drives each. All these on a OnApp integrated SAN.

Page 2 of 2 FirstFirst 12

Similar Threads

  1. Host4Geeks foray into Dedicated Servers.
    By Host4Geeks in forum Web Hosting Industry Announcements
    Replies: 0
    Last Post: 12-29-2012, 10:50 AM
  2. cloud hosting review site 1500.00 cloud-hosting-providers.com
    By affiliatelabel in forum Domain Name with Web Site Offers
    Replies: 0
    Last Post: 11-13-2012, 05:16 PM
  3. Replies: 1
    Last Post: 12-25-2011, 12:01 AM
  4. Replies: 23
    Last Post: 08-31-2011, 02:03 PM
  5. Replies: 4
    Last Post: 04-16-2011, 02:34 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •