Beiträge von gregoryx

    I will add a stop button to the Files tab. If you need to stop one container in compose that has multiple containers, I think they should be split up. The Files and Containers tabs already have Up buttons.

    Much appreciated.


    I'm liking the new setup. As you have said, I was mostly just using Portainer to deploy compose files anyway. Now we have a best-of-both setup. Nice work! 🤘


    Early in the thread there was questions about "Compose" as the name. It still seems, "Docker," makes more sense to me. Clearly, not an issue... but... 🤷‍♀️

    Whew! 22 pages... just to make certain I'm up to date.


    Gotta admit I'd become pretty lazy with the updates... new update for OMV? Yeah... go ahead... who needs to read the release notes? Didn't even notice anything had changed at first - since the containers were all still running and I do most start/stop/update stuff either command-line or Portainer or through Home Assistant.


    I like where this is going, though. So... thank you.


    But I do have a reason for posting - though it's an odd one-off, I'm sure: I monitor my containers from Home Assistant and I stop/start them for various short-term usage. The way the integration works in Home Assistant goes well if a stopped container is "exited" but not "down" - as the button in Services, Compose, Files does. I can do an "stop" from command line or from Portainer (or from Home Assistant) and all works as expected. But the "down" button removes the container and breaks my monitoring and up/down control. Would be cool (for me at least) if the Services, Compose, Containers included a "stop" as well as "up" option. If not... no big deal since I have other ways to do it. But thought it might be worth mentioning that there is a meaningful reason for this function.


    Again, much thanks and appreciation for the direction this is going. The Portainer / Yacht thing always seemed kludgy relative to other integrated functions and this seems really clean.


    Oh, yeah... one other thing to share with anyone who reads this far:


    Just to move forward but not lose my Portainer stuff (yet), I created all my containers in Services, Compose, Files but didn't kill them or remove them from Portainer (yet). Obviously, they can't be started in OMV until they're removed from Portainer... but it has the unexpected benefit of them all showing up in Services, Compose, Containers with all the accurate conditions of State / Status / Image / etc.


    Potentially dangerous if one forgets why they wouldn't start this way... but seems like a nice reverse-integration option while transitioning.

    There is no option to disable this in the web interface. Most people don't run a dns server on their OMV install. And you could use macvlan so that the pihole container would use its own IP address avoiding this problem. For the few people that have this issue, I doubt Volker would add it but you can always ask - https://github.com/OpenMediaVault/openmediavault/issues


    Otherwise, it is a two line cut&paste to fix it:
    sudo sed -i "s/^#Cache=yes/Cache=no/g" /etc/systemd/resolved.conf
    sudo sed -i "s/^#DNSStubListener=yes/DNSStubListener=no/g" /etc/systemd/resolved.conf


    then either reboot or restart systemd-resolved : systemctl restart systemd-resolved

    Nailed it. Not only posted the fix but nailed the "just copy-pate this" side. Thank you.

    Oh, sorry... yeah, that wasn't clear: running a SanDisk Ultra SATA SSD plugged into the SATA interface; that was the only reason I got the HC1 - because it seemed like it should run great with that drive and all.

    The drive gets ~390MB/s read and write in a USB3 plugged directly into a computer. So... seems like it wouldn't be the drive?


    The drive is fairy full (stuffed the 4TB pretty full in testing) so I double-checked it with USB3 to my PC and I'm getting ~390MB/s read and write still.


    Just seems like I've got something wrong since the drive works in direct SATA (to USB) mode and the HC1 /should/ work Gig-E full speed read and write? 🤷‍♀️

    Tested OMV5 with RPI and a USB-SSD and it was good. Around 30MB/s read and write. And iperf3 tests confirmed that was about all I was going to get.


    Ordered and set up the HC1 with the same drive after seeing ~100MB/s read/write promoted on the Armbian and Odroid sites.


    Iperf3 gets solid gig throughput numbers (~825Mb/s single and ~950Mb/s parallel).

    Map a drive with and test with Crystal Disk Mark and I get ~100MB/s read (as expected) but only ~30MB/s write.


    I built the system twice - once from the arm script and then from the Softy setup. Both about the same. Then applied whatever tweaks I could find for SMB (thinking that's the culprit) but no significant change.


    I'm comparing performance to a QNAP NAS and a Synology NAS (both getting full ~110MB read/write) and was optimistically hoping it would be similar on Gig-E - especially after seeing the posted Windows file-transfer speed pix on Odroid and Armbian.


    Any suggestions of what I might be doing wrong? Or what they might have done to see those ~100MB writes?