Beiträge von subzero79

Zitat von KemikalElite

Other posts for reference:

LUKS Automount no longer working on new hardware

RE: LUKS + KeyFile + AutoMount? [SOLVED]

I moved my omv setup from an intel server to a new hp small form factor pc, now 3 of my data drives will not auto mount anymore.

My current fstab is:

Code

/dev/mapper/sda3_crypt /               ext4    errors=remount-ro 0       1
# /boot was on /dev/sda1 during installation
UUID=d61a4271-e5f7-4c46-9b52-791e5ee0fc36 /boot           ext4    defaults        0       2
/dev/mapper/sda2_crypt none            swap    sw              0       0

# >>> [openmediavault]
/dev/disk/by-label/data1        /srv/dev-disk-by-label-data1    ext4    defaults,nofail,user_xattr,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl    0 2
/dev/disk/by-label/data2        /srv/dev-disk-by-label-data2    ext4    defaults,nofail,user_xattr,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl    0 2
/dev/disk/by-label/data3        /srv/dev-disk-by-label-data3    ext4    defaults,nofail,user_xattr,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl    0 2
/dev/disk/by-label/data4        /srv/dev-disk-by-label-data4    ext4    defaults,nofail,user_xattr,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl    0 2
# <<< [openmediavault]

Alles anzeigen

In addition I have crypttab setup for auto unlocking:

Code

# <target name> <source device> <key file> <options>
#System disks - DO NOT EDIT
sda2_crypt UUID=e07a2c60-9fb4-4ff6-85b1-e2ad4c767ebf none luks,swap,discard
sda3_crypt UUID=716b232a-f44a-4f46-a276-4e563856c0c6 none luks,discard
#Data disks
data1 UUID=1dfccdef-7d88-4d5e-9f42-59bd36020c37 /root/keyfile.key luks
data2 UUID=298a0d0f-52ce-48f8-817f-d75d293a4836 /root/keyfile.key luks
data3 UUID=134a1f5e-46b1-422a-b08d-ccc8476340b9 /root/keyfile.key luks
data4 UUID=b9eb750a-5c20-4673-b652-e7e7915cb470 /root/keyfile.key luks

This always worked before on the old server.

Now syslog shows:

Code

Dec 15 17:39:03 omv-server monit[996]: Filesystem '/srv/dev-disk-by-label-data1' not mounted
Dec 15 17:39:03 omv-server monit[996]: 'filesystem_srv_dev-disk-by-label-data1' unable to read filesystem '/srv/dev-disk-by-label-data1' state
Dec 15 17:39:03 omv-server monit[996]: 'filesystem_srv_dev-disk-by-label-data1' trying to restart
Dec 15 17:39:03 omv-server monit[996]: 'mountpoint_srv_dev-disk-by-label-data1' status failed (1) -- /srv/dev-disk-by-label-data1 is not a mountpoint

Repeating for each drive. Every time OMV starts it shows "a start job is running for /dev/disk/" cycling through the uuid and label of each drive for 90 seconds before it fails. the data1 partition on the system drive still works for some reason I don't understand.

Now OMV doesn't load because of an issue with trying to load mergerfs while the drives aren't mounted, so i tried to remove the unionfs from the omv config manually, now it shows a bunch of errors and the file systems and luks tabs arent loading in the web interface.

Couldn't extract an UUID from the provided path '/dev/disk/by-label/data2'.

Failed to execute command 'export PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin:/usr/local/sbin; export LANG=C.UTF-8; omv-salt deploy run --no-color quota 2>&1' with exit code '1': omv-server: Data failed to compile: ---------- Rendering SLS 'base:omv.deploy.quota.default' failed: while constructing a mapping in "<unicode string>", line 42, column 1 found conflicting ID 'quota_off_no_quotas_' in "<unicode string>", line 69, column 1

So essentially the entire omv system is now broken, and I even tried creating a new omv install on a flash drive, and it still won't auto mount the data drives. At this point I'm better off just running a media server off Windows 10 but cant convert my ext4 drives back to ntfs. This is too much troubleshooting for a simple media server with basic encryption.

I don't understand how just using a different motherboard has caused automounting to stop working.

Alles anzeigen

You should drop to terminal and start testing the disks. See how if unlocking manually works, then test the key file. Just go one by one. Maybe remove the key file from the crypttab change it to - and test to be prompted at boot for unlock n

votdev ?

At the moment from what i can found on existing plugins and core ts files, to pass data from a row selection in a table to a form is done via the URL

Exmple LVM plugin in extend we pick the row with _selected[0[

version: "1.0"
type: component
data:
  name: omv-storage-lvm-lv-datatable-page
  type: datatablePage
  config:
    autoReload: false
    stateId: 87081dac-a91b-4a5e-901e-e69290b533ee
    sorters:
      - dir: asc
        prop: devicefile
    store:
      proxy:
        service: LogicalVolumeMgmt
        get:
          method: getLogicalVolumesList
    columns:
      - name: _("Device")
        prop: devicefile
        flexGrow: 1
        sortable: true
        hidden: true
      - name: _("Name")
        prop: name
        flexGrow: 1
        sortable: true
      - name: _("Capacity")
        prop: size
        flexGrow: 1
        sortable: true
        cellTemplateName: binaryUnit
      - name: _("Volume Group")
        prop: vgname
        flexGrow: 1
        sortable: true
      - name: _("Active")
        prop: isavailable
        flexGrow: 1
        sortable: true
        cellTemplateName: checkIcon
    actions:
      - template: create
        execute:
          type: url
          url: "/storage/lvm/lvs/create"
      - template: edit
        execute:
          type: url
          url: "/storage/lvm/lvs/edit/{{ devicefile }}"
      - type: iconButton
        icon: "mdi:camera"
        tooltip: _("Snapshot")
        enabledConstraints:
          minSelected: 1
          maxSelected: 1
          constraint:
            - operator: falsy
              arg0:
                prop: "attributes.snapshot"
        execute:
          type: request
          request:
            service: LogicalVolumeMgmt
            method: createLogicalVolumeSnapshot
            params:
              devicefile: "{{ devicefile }}"
            successNotification: _("Created snapshot of logical volume '{{ devicefile }}'.")
      - type: iconButton
        icon: expand
        tooltip: _("Extend")
        enabledConstraints:
          minSelected: 1
          maxSelected: 1
        execute:
          type: url
          url: "/storage/lvm/lvs/extend/{{ _selected[0].devicefile }}"
      - type: iconButton
        icon: collapse
        tooltip: _("Reduce")
        enabledConstraints:
          minSelected: 1
          maxSelected: 1
        execute:
          type: url
          url: "/storage/lvm/lvs/reduce/{{ _selected[0].devicefile }}"
      - template: delete
        execute:
          type: request
          request:
            service: LogicalVolumeMgmt
            method: deleteLogicalVolume
            params:
              devicefile: "{{ _selected[0].devicefile }}"

version: "1.0"
type: route
data:
  url: "/storage/lvm/lvs/extend/:devicefile"
  title: _("Extend")
  editing: true
  notificationTitle: _("Extended logical volume '{{ devicefile }}'.")
  component: omv-storage-lvm-lv-extend-form-page

Then we can read in the form with _routeParams,

version: "1.0"
type: component
data:
  name: omv-storage-lvm-lv-extend-form-page
  type: formPage
  config:
    request:
      service: LogicalVolumeMgmt
      get:
        method: getLogicalVolume
        params:
          name: "{{ _routeParams.devicefile }}"
        transform:
          capacity: "{{ size | binaryunit }}"
        filter:
          mode: omit
          props:
            - size
      post:
        method: extendLogicalVolume
    fields:
      - type: hidden
        name: devicefile
      - type: textInput
        name: name
        label: _("Name")
        value: ""
        disabled: true
        submitValue: false
      - type: select
        name: vgname
        label: _("Volume group")
        placeholder: _("Select a volume group ...")
        valueField: name
        textField: description
        disabled: true
        submitValue: false
        store:
          proxy:
            service: LogicalVolumeMgmt
            get:
              method: getLogicalVolumeCandidates
          sorters:
            - dir: asc
              prop: name
      - type: textInput
        name: capacity
        label: _("Capacity")
        value: ""
        disabled: true
        submitValue: false
      - type: slider
        name: size
        label: _("Size")
        hint: _("The percentage of the total space in the volume group to use by the logical volume.")
        value: 100,
        validators:
          min: 1
          max: 100
    buttons:
      - template: submit
        execute:
          type: url
          url: "/storage/lvm/lvs"
      - template: cancel
        execute:
          type: url
          url: "/storage/lvm/lvs"

So I can pass only one parameter.?

Is it not possible to pass more data from the selection other than the URL ? a state object?

I have done nothing in Angular but just looking at a few post in stackoverflow seems like router or routerlink can pass data to the components. Looks very similar on what react-router-dom does, passing state object to the new route component, in this case the omv form-page.

https://stackoverflow.com/ques…angular-routed-components

PD: I am not in the developer section. So i just posted here.

The volume Needs to be registered in the db backend. I can see The 280gb volume already mounted. So unmount The disk in terminal, also remove the volume

line in /etc/fstab and proceed to mount the volume via the web ui. After that you’ll be able to create shared folders in the volume.

Very confusing post. Raid section doesn’t have used column. Shows the size of the array. Now then you go to a folder in Windows’s that says some other size.

Mixing a block device, with a folder. Take a screenshot maybe show what the problem is.

When people report that kind of failure is usually a bad network card causing an overflow of retries. Have a spare card to try?

If IP is static change it to DHCP, just to discard any conflicts. Like for example assigning the router ip address to the server.

Guid? I think you mean uuid. That number is assigned when the Fs is registered in the db. Which means when added and mounted. After that will remain the same number unless you remove the fs, which means taking the entry out of the db, which means un-mounting If it succeeds. Add it again it will be another number.

If is the only kernel installed it could go unbootable if you remove it.

I don't know why is trying to remove the kernel, is that current running one? otherwise if not then it doesn't matter

the loaded/running one shows with

uname -r

Is it important for you? i mean are you using a wireless card?

Zitat von kledder

Is it possible to encrypt the entire second partition of the boot drive and not the first?

I ask this because my motherboard only has 4 SATA-connectors and I need all of those for storage. I don't want USB for a boot device. If possible, my plan is to create 2 partitions on a SSD where the first partition contains the operating system and the second partition contains some storage which needs to be encypted.

Yes, but the iso omv installer doesn't give that many options. So use debian you can choose how to layout the partitioning in the system.

You can use LVM and leave a free empty logical volume for the remaining space and encrypt that using luks afterwards.

Or not use lvm, just plain partitions. the first one remains open, the other one encrypted.

Plain bottom, if you need to split the boot drive use the debian installer

In mergerfs you can access the xattrs of the file folder and they will show their origin branch. Just check the mergerfs docs. Also I not sure if these xattrs are visible via samba or nfs.

No, it won’t work.

The plugin fork did a lot of changes internally.

The old ssh method relied a lot on systemd. How it was mechanized is documented in a blog post which was mentioned in the fork. You can follow that post and organize the units manually. Or you can use the other ssh method which uses drop bear at initramfs.

I do not use it since I have omv as virtual machine so I can automate unlock something like `ssh pve qm terminal $vmid`

I have to check the plugin error, look like is missing the db section entry, that should be added at install of the plugin.

No didn’t port the plugin. I created a branch to use crypttab, simplifying the process alittle bit but I didn’t get any feedback

https://github.com/OpenMediaVa…-luksencryption/issues/39

I haven't dont an iso install in a while, do you remember if the hostname was setup autmatically or there is a prompt an you entered plethi.local?

Because if the installer is appending local automatically to the hostname in the db, that means there might be something to correct there as you can see the error triggers in a new install.

Zitat von snsnse

But should OMV keep changing the MAC it would be a little awkward doing this repeatedly due to other configuration tied to the MAC address that my set-up requires (open ports, etc).

It shouldn't change, MAC address is tied to the hardware you cannot change it, you can spoof it temporally as those changes are lost at reboot.

It shouldn’t change on upgrade, odd issue there. But why don’t you paste the new address in the dhcp list of the router.

Omv doesn’t provide in the web ui a field Mac change.

Other than that you should check netplan configuration at /etc/netplan And check netplan docs if they provide MAC address configuration.

How did you install omv?

UUID mount naming and device mount was used at the beginning.

Eventually(maybe in omv4?) changed due to users complaining not having friendly names , because of that they were unable to associate the physical disk/partition to the filesystem in an explorer or terminal.

Now it mounts first option using the fs label and if not present using the device name(id) as seen by the kernel.

This behavior is hard coded and cannot be changed using omv env variables.

Yeah the hostname should be only allowed characters for hostname.

Check in general the hostname and domain don’t put dots in the hostname.

https://github.com/openmediava…vault/json/schema.py#L414