ML
    • Recent
    • Categories
    • Tags
    • Popular
    • Users
    • Groups
    • Register
    • Login
    1. Topics
    2. GotiTServicesInc
    3. Best
    G
    • Profile
    • Following 0
    • Followers 0
    • Topics 1
    • Posts 17
    • Groups 0

    Posts

    Recent Best Controversial
    • Configuration for Open Source Operating systems with the SAM-SD Approach

      I apologize for the way this may be laid out, I'm putting my thoughts down the best I can (I'm a bit scatter brained at the moment). Assume everything is a question 🙂

      So Assuming I'm using a Linux OS for the base, dual controller SAS setup What different ways do we have to configure the system for a SAM-SD setup?

      to make things simple, lets not talk about how to make the cluster, but assume there's four server, each server has 100TB attached to each controller (10 10TB drives, yes I know they don't exist but for easy numbers we'll use it).

      Starting from the ground goind up (before clustering). There are a couple options for redundancy. Are all of these feasible? are all of these feasible for HA?

      • JBOD on each controller, md for software RAID 0 (200TB Total Storage)
      • JBOD on each controller, md for software RAID 1 between controllers (100TB Total Storage)
      • Raid 5/6 on each controller, md for software RAID 0 between controllers (160-180TB Total Storage)
      • Raid 5/6 on each controller, md for software RAID 1 between controllers (80-90TB Total Storage)
      • RAID 1 Array on each controller, md for software RAID 0 between controllers (100TB Total Storage)
      • RAID 1 Array on each controller, md for software RAID 1 between controllers (100TB Total Storage)

      I think this would related to the number of SANs you have, for example, if you have 4 SANS in the cluster, JBOD with RAID 0 might be fine for you???

      I feel like RAID 5/6 might be pointless? Sure you can support 1 to 2 failed drives in each hardware array, but considering you are supporting multiple software arrays this might just be extra overhead with no real benefit? or maybe the rebuild time on a big RAID 5/6 array is worth having this fail safe in place?

      Hardware RAID 1, with software RAID 1, in a cluster, seems super over redundant, or maybe not if the data is that important to you (which I can see it being that important).

      Also is the cluster for redundancy? or higher storage amounts? or some kind of both? (Can you do that?)

      Maybe a question should be what is the normal setup for a dual controller SAN?

      Any input is welcome!

      posted in SAM-SD
      G
      GotiTServicesInc
    • RE: If you are new drop in say hello and introduce yourself please!

      Excellent logic! Pinky up

      posted in Water Closet
      G
      GotiTServicesInc
    • RE: KMS keys now only available upon request from Microsoft

      Interesting, I bet not auto-generating KMS keys will stop no one from pirating their software, just makes your life slightly more difficult 🙂

      posted in IT Discussion
      G
      GotiTServicesInc
    • RE: Configuration for Open Source Operating systems with the SAM-SD Approach

      so why do they make boxes with dual controllers in them then ? I'm referring to your favorite, Nexsan @scottalanmiller

      posted in SAM-SD
      G
      GotiTServicesInc
    • RE: If you are new drop in say hello and introduce yourself please!

      Hello!

      posted in Water Closet
      G
      GotiTServicesInc
    • RE: Configuration for Open Source Operating systems with the SAM-SD Approach

      Thank you for all the information so far and I hope I'm not sucking you dry for information (I've read all the links you've posted already).

      so can't DRBD use FC for the replication? or does it have to use LAN? and if we're forced to do Lan, we should be able to trunk some ports together to get higher throughput no?

      posted in SAM-SD
      G
      GotiTServicesInc
    • 1 / 1