Raid Management in the GUI remove the drive from the array, wipe it and then add it it back to the array
Ok thanks. You mean the super block is not recoverable on that one drive even though it is still working in the array?
So my only choice is to erase the whole drive and rebuild it?
run mdadm --readwrite /dev/md0
However I would run a long smart test on that /dev/sdc just in case it's breaking down
That doesn't do much, it goes into readwrite on first write anyway.
But none of this restores the superblock on that drive, it still says it's missing and it won't assemble or mount on reboot.
Self test says the drive is ok.
The are aging drives, but the are quality WD Gold and still having good smart reports.
Is there not a way to force rebuild the super block on one drive?
And if the mdadm -D /dev/md0 report is as pasted (4 working drives), I'm definitely using 4 drives and not falling back to parity right?
So after first thinking there was something wrong with SMB, I realised the RAID FS had mysteriously gone offline after a Shutdown and Powerup.
On a fresh boot, the RAID filesystem doesn't mount, and various commands tell me that /dev/sdc is missing its superblock
On fresh reboot:
So /dev/sdc is missing superblock.
But if I now run mdadm --assemble --scan -v
It seems to assemble fine, with no complaints. It says all drives are online. I can mount it and all my files are there.
Yet still,
mdadm --examine /dev/sdc
mdadm: No md superblock detected on /dev/sdc.
root@h4:~# systemctl status mdadm
● mdadm.service
Loaded: masked (Reason: Unit mdadm.service is masked.)
Active: inactive (dead)
Fdisk -l does not list /dev/sdc
Is there a way to restore the superblock on sdc and have it mount at boot as before?
Thanks.
I've setup 3 Odroid HC2 boxes now and cron always send me that error.
pam_tally2 --reset alleviates it for a while, but it always comes back.
I'd say this would be the case for anyone running OMV4 on Odroid HC1/HC2/XU4 if they have email notifications setup
I'm baffled that Apple made SMB default protocol. I mean, the added compatibility on networks with Windows machines probably clears up their feedback and support lines a lot, but when you're used to the lightning speed of AFP, SMB seems like your interfacing with a windows 98 machine over 100mb LAN.
Once a transfer gets going, the speed of SMB is only marginally slower than AFP. But the latency of directory listing is atrocious. Not to mention network discovery delays and phantom network drives hanging around in Finder than have dropped off the network. Its 2020 geez, why go backwards?
Can someone give a bit more detail on how to get this working in OMV5?
I assume SSH in and 'apt-get install Netatalk'. . What else needs to happen to place nice with OMV5?
EDIT: Bit of further research shows latest versions of SMB3 along with Catalina killing of NetBios *should* alleviate these issues, although I'm doubtful. I can still see a windows machine in Finder that hasn't been on the network since yesterday. Archaic is a word that springs to mind.
"But your docker stuff actually does not need to be on a shared folder, right? You can put that on your system drive"
please do not store plex conf on boot drive, , because plex database can grow rapid and eat all free space on boot disk, if you use data disk for it, you have enought space, but filesystem is noexec by default
So you see the conundrum.
This is an Odroid HC2, so there is only a microSD and one drive to play with.
But I guess I could partition off a small potion of the data drive for the Docker stuff that needs exec, though not that easy with an already partitioned 4TB ext4 drive when my other machines cant read ext4.
Thanks for the help anyways,
Not sure what forum this belongs in really but here goes.
I installed Plex in a docker after watching a tutorial and all was fine for months.
All of a sudden after no changes or updates on my part, Plex wouldn't play back any audio.
After looking into the transcoder failure errors in the logs, it seems the problem is because the transcoder libraries are stored in /confg/Plex... in /sharedFolders and that OMV mounts drives as noexec by default.
That all makes sense to me, and I'm aware I can modify /etc/openmediavault/config.xml to change the noexec flag to exec for the drive.
Noexec is a smart thing to do though, it seems a shame to have to change it for the whole 4TB drive just for a couple of small files.
I tried to symlink just the /Codecs folder over to the SDcard that OMV is on, but I get a 'too many symlink levels' error, presumably because /config is itself a symlink. Moving the whole Plex config fodder to the SD wont work because its a few Gb, way bigger than the space left on the SD.
I'm also totally confused as to why it was working perfectly fine up until now without needing to disable noexec.
Am I going about this all wrong? Is there a better way?
So I was trying to change the DNS on the Odroid when I realised I couldn't because) no interfaces were listed.
When I added one that started with 'et' (which I believed must be the ethernet port) and rebooted, I can now no longer access the device.
It shows up on my router as having an IP address, but it wont respond to any http or ssh requests.
If I were to pull the SD card, how could I restore the network interface section to its blank (auto?) status that appears to be the default for Odroid?
Thanks.
Solution:
For AFP to connect for a user account, at least one Shared Folder must have Read/Write or Read-Only ticked in Users->Privileges. (or Shared Folders->Privileges)
This is not the case for SMB service which makes this confusing.
Being that this effects AFP but not SMB, perhaps a note in the AFP section would be appropriate.
So I'm a OMV newb and not a big fan of SMB due to it's tendency to keep phantom network shares floating around Finder sidebars are being disconnected.
So why can't I use AFP?
I created a user fo myself with default settings. Doubled checked its part of the 'users' group.
I created a Shared folder, added it to Apple Filing shares with default permissions.
Now I can access the share fine with AFP via the Guest account.
I can access the share fine via SMB with my user account credentials.
But no matter what I try, Finder always just says 'Connection Failed' when using AFP and my user account credentials
Any ideas?
