Building OMV automatically for a bunch of different ARM dev boards

    • nasnas wrote:

      Thank you for sharing the great news
      Still interested in feedback made with original USB3 Seagate disk enclosures using this new image (since a potential kernel fix is also included -- but since I don't own Seagate disk enclosures and don't want to waste my money for such things I can't test)
      'OMV problems' with XU4 and Cloudshell 2? Nope, please read this first. 'OMV problems' with Cloudshell 1? Nope, just Ohm's law.
    • Hello everyone,

      I'm Frepke and new on the forum and waiting for my ordered Espressobin.
      I would like to run OMV on the Espressobin, is this possible already?
      And if yes, is this done trough the installer in Armbian or an already build image?

      Thanks for any help... ^^


      kr.,
      Frepke
    • Frepke wrote:

      I would like to run OMV on the Espressobin, is this possible already?
      I let a new OMV build run at the moment, will upload it later.

      Please keep in mind that this image is based on a WiP (work in progress) Armbian configuration using Marvell's legacy kernel with a few unresolved issues yet. So please consider this being a preview and not a fully stable/supported OMV image yet. At least I won't release any official OMV images for Espressobin unless mainline kernel support is there (since Marvell's 4.4 BSP kernel might not be supported that long)

      You can watch progress over there: forum.armbian.com/index.php?/t…velopment-efforts/&page=5 (please remember: it's WiP for a reason and currently only developer feedback is needed and no end-user support available)
      'OMV problems' with XU4 and Cloudshell 2? Nope, please read this first. 'OMV problems' with Cloudshell 1? Nope, just Ohm's law.
    • tkaiser wrote:

      Frepke wrote:

      I would like to run OMV on the Espressobin, is this possible already?
      I let a new OMV build run at the moment, will upload it later.
      Please keep in mind that this image is based on a WiP (work in progress) Armbian configuration using Marvell's legacy kernel with a few unresolved issues yet. So please consider this being a preview and not a fully stable/supported OMV image yet. At least I won't release any official OMV images for Espressobin unless mainline kernel support is there (since Marvell's 4.4 BSP kernel might not be supported that long)

      You can watch progress over there: forum.armbian.com/index.php?/t…velopment-efforts/&page=5 (please remember: it's WiP for a reason and currently only developer feedback is needed and no end-user support available)
      Thanks tkaiser,

      I hope there's a stable build available soon, but in the mean time I follow the Armbian forum.


      kr.,
      Frepke
    • Frepke wrote:

      I hope there's a stable build available soon, but in the mean time I follow the Armbian forum.
      Here you go: OMV_3_0_86_Espressobin_4.4.79.7z

      But I wouldn't call this stable atm -- quite the opposite to be honest. Starting with this release no SSH login as root is possible by default so you have to login through the web UI first, access the SSH service and allow root login. I already talked to @ryecoaaron about a mandatory password change on first login (that then also changes the root passwd) and still hope this gets implemented soon (re-enabling SSH login for root might then be an option after password change).

      I couldn't test any further with this image since It seems I'm somewhat unlucky with Espressobin so far (this is my 2nd board, I'm never able to access a writeable serial console, this 2nd board also overheats a lot). Won't look into this for the next three weeks since too busy with other stuff.
      'OMV problems' with XU4 and Cloudshell 2? Nope, please read this first. 'OMV problems' with Cloudshell 1? Nope, just Ohm's law.
    • tkaiser wrote:

      Frepke wrote:

      I hope there's a stable build available soon, but in the mean time I follow the Armbian forum.
      Here you go: OMV_3_0_86_Espressobin_4.4.79.7z
      But I wouldn't call this stable atm -- quite the opposite to be honest. Starting with this release no SSH login as root is possible by default so you have to login through the web UI first, access the SSH service and allow root login. I already talked to @ryecoaaron about a mandatory password change on first login (that then also changes the root passwd) and still hope this gets implemented soon (re-enabling SSH login for root might then be an option after password change).

      I couldn't test any further with this image since It seems I'm somewhat unlucky with Espressobin so far (this is my 2nd board, I'm never able to access a writeable serial console, this 2nd board also overheats a lot). Won't look into this for the next three weeks since too busy with other stuff.
      Thanks tkaiser,

      I thought I give it a try when my espressobin arrives, but I saw your message at the Armbian forum forum.armbian.com/index.php?/t…findComment&comment=36984. That made me doubt, maybe I'll wait for a more stable version of Armbian.

      kr.,
      Frepke

      The post was edited 1 time, last by Frepke ().

    • Frepke wrote:

      That made me doubt, maybe I'll wait for a more stable version of Armbian.
      Well, currently this looks like a simple hardware issue called overheating needing either hardware fixes (active cooling) or software (checking at which voltage the CPU is fed, if possible optimize this. We don't need to think about another countermeasure called throttling since currently I get these symptoms even with an idle board).

      Let's wait and see. I'll run the Linpack benchmark the next 12 hours with active fan and heatsink and if the board survives that IMO it's obvious that we've a hardware issue. See also espressobin.net/forums/topic/overheating/
      'OMV problems' with XU4 and Cloudshell 2? Nope, please read this first. 'OMV problems' with Cloudshell 1? Nope, just Ohm's law.
    • I just wanna thank @tkaiser, Armbian and OMV for providing the best OMV installation since OMV 2.x. I just switched to it on Friday and I am more than pleased with my XU4 performance and overall behaviour. My USB 3.0 HDD (4TB Seagate Hub) does not have the UAS issues, luckily.
      Riddle me this, riddle me that
      Who is afraid of the big, black bat?
      I write on a blog (Romanian mostly)
      Testing (latest) OMV 3.0.xx on an oDroid XU4
    • Building OMV automatically for a bunch of different ARM dev boards

      tkaiser wrote:

      Frepke wrote:

      I hope there's a stable build available soon, but in the mean time I follow the Armbian forum.
      Here you go: OMV_3_0_86_Espressobin_4.4.79.7z

      But I wouldn't call this stable atm -- quite the opposite to be honest. Starting with this release no SSH login as root is possible by default so you have to login through the web UI first, access the SSH service and allow root login. I already talked to @ryecoaaron about a mandatory password change on first login (that then also changes the root passwd) and still hope this gets implemented soon (re-enabling SSH login for root might then be an option after password change).

      I couldn't test any further with this image since It seems I'm somewhat unlucky with Espressobin so far (this is my 2nd board, I'm never able to access a writeable serial console, this 2nd board also overheats a lot). Won't look into this for the next three weeks since too busy with other stuff.


      Hello tkaiser,

      A few days ago I received my espressobin from @Lupus. After I installed Putty on my MacBook, I set the environment variables and try to boot an image from the sd-card.

      With Armbian_5.32_Espressobin_Ubuntu_xenial_default_4.4.78.7z it's possible to boot. But I like to run OMV on the Espressobin, so I tried OMV_3_0_86_Espressobin_4.4.79.7z

      Unfortunately it's not possible for me to boot this image from an sd-card.

      Should it be possible to run this Image?


      Kr.,
      Frepke


      Verzonden vanaf mijn iPhone met Tapatalk

      The post was edited 1 time, last by Frepke ().

      Post by Frepke ().

      This post was deleted by subzero79: Double ().
    • ryecoaaron wrote:

      I know you aren't a big xu4 fan but what do you think of the odroid-hc1?

      Looks very promising! Some more info here.



      I already suggested few pages back to avoid USB3 cable/contact hassles by putting the USB-to-SATA bridge on the PCB with RK3328/ROCK64 in mind but Hardkernel already delivered. The idea to use the metal enclosure as giant heatsink is also great and they even thought about the 3.5" HDD use case, there's a 12V connector on the HC1 board that feeds the 12V pins on the SATA power connector. There are also some rumours wrt a HC2 (with the ability to connect two disks most probably using JMS567 as on the Cloudshell 2 which I won't consider a great solution. Adding either an USB hub or a port multiplier adds more complexity).

      Hardkernel sent out 2 HC1 review samples that should arrive soon and I'll report back here my findings (already curious how SSD performance will look like since on HC1 there's no USB3 hub in between host and disk). But to be honest: Now that XU4's Achilless heel is avoided (USB3 contact problems), that voltage-drops are less likely to happen (Cloudshell 1 problem) and that a good USB-to-SATA bridge is used avoding potential UAS hassles the only remaining question is about how efficient heat dissipation will be. HC1 single disk NAS performance will be as good as any x86 box maxing out Gigabit Ethernet anyway.
      'OMV problems' with XU4 and Cloudshell 2? Nope, please read this first. 'OMV problems' with Cloudshell 1? Nope, just Ohm's law.

      The post was edited 1 time, last by tkaiser ().

    • New

      ryecoaaron wrote:

      what do you think of the odroid-hc1?

      Small update: I'm still struggling with obscure network performance problems in my lab (only affecting XU4 and HC1 but other devices not for whatever reasons. Tried now the 3rd switch after my normal test switch died but to no avail). So still no full review and just some information.

      I collected some performance data and observations with regard to heat dissipation and thermal/throttling behaviour with OMV's kernel variant today: See comments here: cnx-software.com/2017/08/21/ha…-is-now-available-for-49/

      Also today over at Armbian we did some testing wrt AES performance. ODROID's Exynos SoC doesn't support ARMv8 crypto extensions (since being a 32-bit ARMv7 SoC with A15/A7 cores) so if we're talking about encryption/decryption performance (eg LUKS) other platforms might have a clear advantage here. Currently RK3328 on ROCK64 is best performer when judging by synthetic benchmarks (openssl --speed). ROCK64 AES single and multi threaded performance are 9 to 11 times better compared to HC1: forum.armbian.com/index.php?/t…findComment&comment=37829

      We're preparing a few more realistic tests covering LUKS encryption and then local IO performance (iozone tests) as well as NAS performance (LanTest). But this might take some time...
      'OMV problems' with XU4 and Cloudshell 2? Nope, please read this first. 'OMV problems' with Cloudshell 1? Nope, just Ohm's law.
    • New

      ODROID-HC1 mini review is online: forum.armbian.com/index.php?/topic/4983-odroid-hc1/

      Really impressive NAS performance especially with Ethernet IRQs moved from little to big core (which is something some users might want to revert to slightly lower overall consumption under load). But playing energy efficient NAS with HC1 is somewhat hard anyway (especially when compared with upcoming RK3328 boards with same performance but half the consumption -- but until there are RK3328 boards available that have also an onboard JMS578 or something similar HC1 is in a league of its own)
      'OMV problems' with XU4 and Cloudshell 2? Nope, please read this first. 'OMV problems' with Cloudshell 1? Nope, just Ohm's law.
    • Users Online 3

      3 Guests