Posts by Sergio

    Last night I tried upgrading a system clone in the spare disk. :sleeping: It failed miserably. So I reverted.

    And I had another issue with docker while the plugin was offline. Suddenly my containers where not starting and then they disappeared. Without me doing anything, only rebooting. But in the end I managed to diagnose it.

    I store the data from my containers in a ZFS array. Its mounted in the Z2 directory, and I refer it in my containers, for example: /Z2/docker-data/emby.
    Well, since the docker plugin failed me, the docker service was starting before ZFS during boot. As the pool was not yet imported, there where no content in Z2, so docker decided to create the referenced folders. ZFS then attempted to mount the pool on Z2, but failed because the directory wasn't empty. So I missed the pool and containers because they couldn't boot.

    I'll try to do a clean install on the spare disk, but we'll see how my arrays will be restored. Guys, that's not a nice situation. Now I'm thinking on the several installations I've done to some customers and where left alone at the datacenter. ||

    Hello subzero79.

    Mmmm... no, I think you didn't understand.

    But it seems the problem is that my docker plugin doesn't turn on. It happened before, but previously stopping the containers at command line let the plugin start.
    Now I can't start it, using the new folder or the old one. With containers off or anything I try.
    After saving, It stays for one minute and then throws the error.

    Is that same error I reported: link but now I can't find anything to make it work.
    Docker works at command line level without any problem.

    Is there any way of changing the docker working folder?
    I'm moving my docker system and containers from one array to other.
    I stopped all the containers, then the plugin and replicated the two main folders I'm using on the new disk, but when I try to change the base path in the plugin, it throws a "communication error" or "Gateway Time-out"

    I suppose that whenever I can do this I will be easy to replicate the containers using the new shared folders.

    Thanks you guys.
    I was tinking more of something running on bonjour/zeroconf/avahi protocol. The fact is that openmediavault publishes it's domain name correctly but none of the raspbian machines does. Without doing nothing I can access openmediavault.local, but not p3steel.local (my 3d printer) or domoticz.local.
    I investigated that past night and came across that installing apples bonjour on a windows machine got this machine resolving the raspis. But that left behind the androids. There's nothing for them.
    My main router runs DD-WRT with all it's flares and wris... but my problem is that my ISP supplies modems that are locked and can't be configured in bridge mode. All my professional and nice installation flows through and relies on a cheap device.
    This morning I've done a search to all my near IPS stores and got back with 5 modems. Now I'm playing with them, It must be some valid between them. ;)

    Just wanted to know how OMV was doing it's registering so well.

    Hello guys. Today I realized that my server is also reachable at myservername.local. Just out curiosity, how exactly is that accomplished?
    Sometime ago I tried it and failed. Then I thought it was only a matter of having avahi installed, but today I've tried it again on two raspis without success.
    Is something about the domainname registration?

    Oh, sorry, i did search for "stale cache" as is.
    The file is present and as content. I can see references to the disks by id.

    Anyways I executed # zpool set cachefile=/etc/zfs/zpool.cache Z2 and rebooted.
    The pool it's online now!

    I think we also improved it with the id change.

    I got yYour omvextras update and after apt-get update shows this at the end:

    Executing your command solved the issue. Thanks Aaron!

    I haven't found anythig related to "stale cache" on the link you previously sent.
    Is there a way to "debug" it?

    I don't think is anything wrong with the pool or cache. If so it would also cause issues when manually importing. I'm suspecting more on some weird init problem with init order/permissions or anything that could have changed during a system update.

    That message shown about doing a pool update, you've seen it? Could that be the reason?

    I executed the following commands:


    # zpool export Z2
    # zpool import -d /dev/disk/by-id Z2

    And the pool was correctly imported using device id's as it showed on zpool status.

    I made sure that the correct option was in the configuration file (ZPOOL_IMPORT_PATH="/dev/disk/by-id") and rebooted. But the pool isn't there after the reboot. ;(;(;(

    The default line also didn't work.
    The output from zpool status after manually import is:

    The file list on /dev/disk/by-id is:

    The disks on the Z pool are the western digital. I've tried other directories, but there's nothing on them.

    Hello nicjo and thank you.

    The ZFS_MOUNT variable is correct but the import path is commented out. What should be in there, the /Z2 folder name?

    I'm having the same problem, but the repo failing is the omv-extras one. Is anything bad in this instalation?

    I've executed the previous commands with that same result.

    I'm having problems with the ZFS plugin. After a system restart does not mount the existing ZFS pool. I have to go to the plugins interface and manually import it every time. That's a problem, because I store my images and containers there, so even aftar the pool is online, docker and virtualbox still fail and I have to go there and restart or even reconfigure them , as in case of docker.

    I've already tried reinstalling the plugin without results.

    ZFS Pools are not imported at boot. I have to manually import them each time.
    Virtualbox says: Object functionality is limited.
    Docker, after repairing it's configuration doesn't show any of my servers.

    Ahhh... I'm about to drop this thing and make a decent windows server machine.

    EDIT: It seems that all my remaining problems are from ZFS not mounting at boot.

    Started a recovery linux distribution and modified the network interfaces file to remove all IPv6 stuff and managed to continue booting.
    Then shows the following message: "chgrp: cannot access '/var/log/dmesg': no such file or directory"
    Followed by some other messages from umount, trying to unmount some /var/folders including /var/lib/openmediavault/rrd always showing "not mounted"

    Then stucks again for a quarter of an hour then resumes booting until the end.

    After that I can log into the web gui.
    Entering docker plugin shows the following error:


    Error #0:exception 'OMVModuleDockerException' with message 'Error: "couldn't connect to host" - Code: 7' in /usr/share/omvdocker/Utils.php:63Stack trace:#0 /usr/share/omvdocker/Utils.php(180): OMVModuleDockerUtil::doApiCall('http://localhos...')#1 /usr/share/openmediavault/engined/rpc/ OMVModuleDockerUtil::getContainers('42005')#2 [internal function]: OMVRpcServiceDocker->getContainers(Array, Array)#3 /usr/share/php/openmediavault/ call_user_func_array(Array, Array)#4 /usr/share/php/openmediavault/ OMVRpcServiceAbstract->callMethod('getContainers', Array, Array)#5 /usr/sbin/omv-engined(500): OMVRpc::exec('Docker', 'getContainers', Array, Array, 1)#6 {main}

    All other things work as expected, including VistualBox. Smart doesn't show any problem.
    Deactivating the flash plugin made the system boot up without any stop.

    Seems that something was messed during the docker update, that I have a failing ECC module :/ (ejemm) and that my first network port is also failing. ?(

    As I can't diagnose the problem and it's too strange to me, I'll try to describe the situation.

    Yesterday my emby installation starrted to act strange. The emby app on my Tv seemed to log into the server but it wasn't loading all the contents. Only one or two here and there (I'm talking of video lists)
    So I went to my computer and tried to log into omv web gui, without success. Also I tried to log into ipmi web management but couldn't get into it.
    I realized that if I leave those pages loading, after some long time they load partially. In case of omv, it loads the background and title (no login) and ipmp loads an imcomplete login.

    network drives seemed accesible, but as I tried to open them windows stuck loading the content

    I also tried ssh and It connected, but as long I executed some command, as top, it became unresponsive and the ssh connection resets with a timeout.

    I tought then that the network port could be damaged and plugged the second one and configured for another IP. This worked. I could enter the OMV web gui and seemed fast and completely loaded. No IPMI though.
    So I executed omv-firstaid and configured the 1st interface with some other ip (ok) and then tried to configure the 2nd with the original ip. It failed when saving, but I could access the web gui on that ip.
    I saw anywhere that something failed with LO interface configuration (i'll try to reproduce again)
    The icons con the web gui where weird and the services where not working. Docker throws error opening it's page.... network interfaces page is empty...

    I thought of disk corruption on the SSD and got the smart values but all are ok. I executed shutdown -rF now so I could check the filesystem on boot, but showed several messages about being unable to unmount ZFS filesystems, failing stopping vbox and then stuck on "Deconfiguring network interfaces..."

    I pulled the cord and restarted it. While booting it showed the root fs check without problems.

    Then stuck. ;(

    Well, it seems not to be stuck since I can insert an usb stick and it shows detection messages.