Toughts of building new storage

  • Hello All,


    I am planing building new storage based on OMV , the storage will be dedicated for MS SQL ISCSI based.
    Now since i am limited with budget and requested of building at least 24TB and to serve dozens of db's, i have two parts question one is on idea i have if its feasible and good one with OMV for achieving as much as possible IOPS to compensate of current "poor" HW
    second if current suggested hardware will be enough to cope with IOPS.


    First question
    I would like to use SSD for caching my plan is to use 4x 480GB Sandisk extreme Pro ,2 mirrored for read + 2 mirrored for write/back (We have UPS) using EnhanceIO technology ,my question is it feasible with OMV ?
    can i use it combined with ZFS as it used its ARC(MAIN MEMORY) mechanism for fisrt cache level without using the L2ARC +ZIL , would i gain something is it a good choice or if there is better filesystem to use with EnhanceIO if it operational off-course with OMV ?


    Second question
    if first one its ok obsessively, my hardware suggested is
    storage will be based on Intel SR2600URB with 48GB memory ecc
    I plan to use 6x Seagate STBD6000100 as the storage disks mounted on 2x LSI MegaRAID 9261-8i controller (one for disks second for SSD's)
    dual 10g Intel e10g42bfsr type .
    Please advice
    Thanks

    • Offizieller Beitrag

    While I think Debian Wheezy would work fine, I think this is a little beyond the scope of OMV. You would end up doing so much outside the OMV web interface that it would probably be better to use just Debian.


    EnhanceIO sounds like a good plan. Never used it before though.


    I wouldn't use zfs. I would use hardware raid with plenty of cache and battery backup for this server (9261 should work).

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • ZFS does not sound like a bad choice with deduplikation and that much ECC memory.


    Though aaron is right that you may better off using debian without OpenMediaVault. It can't help you wirh zfs in the actual state as it is not (yet) supported via GUI.


    Oh and I think its either ZFS or hardware RAID. Not sure if it's advised to do both.


    Greetings
    David

    "Well... lately this forum has become support for everything except omv" [...] "And is like someone is banning Google from their browsers"


    Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.

    Upload Logfile via WebGUI/CLI
    #openmediavault on freenode IRC | German & English | GMT+1
    Absolutely no Support via PM!

    • Offizieller Beitrag

    If it is serving mssql databases, dedup isn't necessary. So, I think hardware raid is still the way to go.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • I would've thought that dedup on such big databases could be helpfull. I also think the data integrity vould be higher with ZFS than hardware raid in that scenario. Though my opinion is based on the little theoretical ZFS knowledge that I have.


    Greetings
    David

    "Well... lately this forum has become support for everything except omv" [...] "And is like someone is banning Google from their browsers"


    Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.

    Upload Logfile via WebGUI/CLI
    #openmediavault on freenode IRC | German & English | GMT+1
    Absolutely no Support via PM!

    • Offizieller Beitrag

    I don't know much about zfs but I thought zfs dedup was file based. So, no two databases will be the same file and sql server only uses two files for each database. Data integrity might be help but sql server keeps the most recent data in memory. So, ECC is more important.

  • Okay, I did not thought of that it's probably filebased.


    @talb better get a UPS too for this system.


    Greetings
    David

    "Well... lately this forum has become support for everything except omv" [...] "And is like someone is banning Google from their browsers"


    Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.

    Upload Logfile via WebGUI/CLI
    #openmediavault on freenode IRC | German & English | GMT+1
    Absolutely no Support via PM!

  • Thanks All for the suggestion , will take them in count
    Do you have any idea if following ZFS and EnhanceIO will be implemented in some point either natively or via plug-in in OMV?
    Thanks

    • Offizieller Beitrag

    zfs plugin is coming. The backend is done but the frontend is 75% (guessing). Not sure about EnhanceIO.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!