Beiträge von tkaiser

Zitat von vko

Question: How to I set the frequency to 1.8/1.4GHz?

Just set it to 1.8 GHz and you're done. Little cores at 1.4GHz or 1.5GHz doesn't change a lot, it's the big cores that make the difference.

Zitat von vko

xhci-hcd xhci-hcd.11.auto: Ring expansion failed

Known 'USB sucks on Linux' problem but this shouldn't occur with my OMV image since https://forum.armbian.com/topi…oherent-pool-memory-size/ -- if this also occurs with the Armbian based OMV image then please report this over at Armbian (I stopped contributing there)

Zitat von TOMillr

Is the index of the file system stored on the system drive?

It's either cached in RAM (all the available RAM usually ends up being used for this sort of caches) or on the disk in question. Whether the disk spins up or not when you 'browse' the filesystem depends on too many factors to give any meaningful answer. I just had a quick look at an OMV5 test install:

/dev/sdb1 /sharedfolders/ext4 ext4 rw,noexec,relatime,jqfmt=vfsv0,usrjquota=aquota.user,grpjquota=aquota.group 0 0
/dev/sdb2 /sharedfolders/btrfs btrfs rw,noatime,nodiratime,compress=lzo,space_cache,subvolid=5,subvol=/btrfs 0 0

The ext4 formatted share has relatime set so accessing the filesystem contents will spin up the disk at lest once a day while the btrfs formatted share uses noatime,nodiratime so simple read accesses won't result in write activity. But that's just one of the many reasons why the disk needs to be accessed.

Zitat von vko

Any idea why formatting an SSD connected over USB is taking so long and crashing the board in the process

Most probably since USB(3) sucks.

Edit: there's a need to clarify: when you dealt with the SSD on the USB port did you then power the M4 via the SATA HAT or USB-C? If the latter then this is known to be M4's Achilles heel (undervoltage resulting in freezes under load). Since powering via USB-C (not USB PD compliant) is such a sh*t show FriendlyELEC now even sells an own PSU module for the board.

Zitat von MarcS

The problem was a Kernel confilct with ZFS

Quick web search suggests that it was rather a missing reboot. Anyway: if you're now on Proxmox's 4.15 that's fine.

Zitat von TOMillr

I'm curious how OMV handles spin-up an spin-down of connected hard-drives.

OMV does nothing here since this all depends on what the Linux kernel does depending on how you configured your storage. With 3 drives and RAID5 all drives will spin up, with other attempts that keep each disk independent from each other this won't happen.

You need to provide a bit more details about your setup...

Where did the idea to fiddle around in the OMV UI to set a 'static IP address' originated from? From a guide like this here New User Guide - Getting Started with Openmediavault ?

Did you already reboot after 'installing' the ZFS plugin?

Zitat von Akeem

the OMV machine always receives the same IP address 192.168.0.115 from the router

What about reverting back to DHCP then if it's always the same address (as it's almost always the case with DHCP in this century. And usually modern router allow to set a static DHCP lease at the router's side too)? At least just for testing purposes...

BTW: There is no 'OMV firewall' and there's also nothing like 'an outside network' when it's about 'OMV services'. Your router should do NAT and as such packets that originate from the outside get headers rewritten so any connection attempt from the output looks like a connection originating from your router. The packets that go back are sent from OMV to the router and there the NAT engine rewrites the headers again and send the packages to the outside.

Zitat von Nefertiti

I know Tkaiser would disagree about using beta so it is up's to you!

Tkaiser pointed out directly above that the new and different error @vecnar now experienced will be fixed only on beta images and also that 0.8 will be released soon. 0.8.0rc16 will most probably be the next stable version. Test matrix can be found here https://docs.google.com/spread…ApAyBBfc_BGv2A/edit#gid=0 and if you want to contribute then Pine64 IRC is the place to be.

@Nefertiti the problem with 'beta releases' is that stuff is supposed to break without any reason since ayufan generates those releases without testing. They are only there to collect feedback and this is something experienced users are needed for willing to give constructive feedback where it belongs to (not here in OMV forum). So when using beta images you are simply on your own if stuff doesn't work as expected

@vecnar I hope you are aware that you're reporting different issues. If not even a PCIe link between SoC and your new SATA HBA can be established then this is something different than the SATA HBA not being able to establish SATA links with disks.

Zitat von vecnar

rockchip-pcie f8000000.pcie: PCIe link training gen1 imeout!

This means that the PCIe connection between RK3399 and your PCIe card is already failing. ayufan reported recently he found an underpowering issue for PCIe which will be fixed with next release (0.8.0 is about to be released soon).

Zitat von getName()

If you need to monitor services, do it with existing frameworks (e. g. Icinga)

IMO a bit overkill especially when it's about only one single host where you want to monitor services. Also overkill but pretty straightforward: webmin. You can configure email notifications for service status and also trigger commands when a certain condition is met (e.g. service not running any more).

But I agree that diagnosing and fixing the problem is the smarter approach (logs exist for a reason).

And OMV users need to be careful to never manage OMV's services (e.g. Samba) via Webmin.

Zitat von liv

I cannot mount anymore

So it worked before? If that's the case most probably silent bit rot on the rootfs. You might want to check package integrity.

Zitat von jata1

My key question is really about handling incremental backups from my Mac to the NAS

TimeMachine (on OMV4 using Netatalk/AFP, later with OMV5 using Samba/CIFS)

• Create a new OMV user called whatever with a password (that's always the first step you need to do on an OMV install anyway, saves you so much troubles if you do NOT skip this step)
• Assuming you're on OMV4 create a new AFP share with appropriate permissions (admin and user read/write), enable the TimeMachine checkbox
• On your Mac open System Preferences --> TimeMachine --> Add Backup-Volume
• optional: enable encryption (highly recommended), let the password be stored in Keychain and if you use a password manager, store it there too (without this password you won't be able to access your backups so choose one with high entropy but easy to type since in case of disaster recovery if you have to setup your whole Mac from scratch you need to enter the passphrase)
• Choose your new volume, use the logon credentials from before (and let them be saved in Keychain of course) and wait until the first full backup has finished
• optional: now that you know how much data will be backed up set a quota for this AFP share with 120% the capacity. This will result in TimeMachine on the Mac thinning up older backups and you don't end up with too much snapshots
• optional: if you want to speed up the first backup enter this as root in Terminal.app on the Mac: sysctl debug.lowpri_throttle_enabled=0
• Test restore

The last step is the most important one since only a working backup defines a functional backup.

Zitat von R2D2

A chance to do proper debugging with the script doesn't exist?

Every bash script can be run in debug mode by executing it with /bin/bash -x in front. But it's pointless since the issue is a prior failed OMV installation attempt. You would need to clean your local apt database as the error messages already suggested and since there's an OMV4 image for your board the whole stuff makes no sense at all (same with Orange Pi PC for which you can simply use OMV_4_Orange_Pi_PC_Plus.img.xz since the only differences are (missing) eMMC and wireless).

Zitat von R2D2

So I should be on the latest and greatest version

No. I was referring to pull the latest version of the script directly from Github. You chose to use armbian-config from apt.armbian.com (Igor controls this repo and I have not the slightest idea how often he updates stuff, I also have not the slightest idea how he deals with problems in his software, usually he doesn't respond, see here or there). Anyway: what you're trying makes no sense. Just use the NanoPi M4 image from Sourceforge. If you want to continue with armbian-config be prepared to open an issue on Github if it still doesn't work as expected.

The above was only to outline how one would today try out OMV5 on ARM. Once Debian Buster is ready and then once OMV5 is ready (2020? Or maybe still in 2019?) the armbian-config variant most probably will work then. Until then this is stuff for somewhat experienced testers and developers.

Zitat von R2D2

I tried to do a fresh install now on armbian like @tkaiser suggested but the omv install fails. It's quite hard to get anything copied from the console because the gui appears quite quickly.

Which gui? Anyway:

Zitat von R2D2

/usr/sbin/softy: line 587: [: 24298: unary operator expected

This line is empty: https://github.com/armbian/con…bc4a/debian-software#L587

So no idea what you're trying to do but what I tried to explain previously was how to get an OMV5 install today (which requires using a Debian/Armbian Buster image in the first place while you seem to be using a Stretch image?)

Zitat von crashtest

Address the user, not me

'The user' hopefully now knows where to look at (fallback behavior to guest logon if authenticated logon fails in 'non domain' situations. What you've written below is incorrect as almost always when you try to deal with SMB issues):

Zitat von crashtest

In such a case, any credential provided by the work station would be accepted by the server for read only access

Nope, first authenticated try fails and 2nd guest logon (fallback) then might succeed. It's important to understand not only the details but also the basics and I was not talking to @ras07 but to YOU since it's important for you to understand that you're not knowing what you're babbling about. Unfortunately again to no avail. This level of ignorance and arrogance is really epic.

If you wouldn't refuse to learn it's as easy as increasing Samba's 'Log Level' on the OMV server and to look what's happening. But nope, instead of improving on your knowledge you prefer to agitate users as usual.

Zitat von exitcode0

Atomic Pi? ... I'm thinking about buying it and would appreciate feed back and notes from current owners.

Multi-system case

Zitat von crashtest

The Domain login "might" need an extra step (not tested):
**Before setting up a user in Credential Manager, I'd test access first. A CM user may not be necessary.**
Set up a user name and password in Credential Manager. (The same username and password used to login to Win10 would make sense, but any username and password would do.)

What an insane BS.

• This is not related to 'Domain' at all! If client and server joined the same Domain Kerberos should be used to avoid transmitting logon credentials at all. This requires network time synchronized and local DNS working (something you constantly try to sabotage with your 'guides')
• When Kerberos is not in use Windows tries to authenticate with a username and password against an SMB server. It either uses the logon credentials of the local user (say user 'user1' and password 'pwd1') or if for this server other logon credentials are needed (e.g. 'user2' and 'pwd2') they need to be stored in Credential Manager so Windows knows to authenticate with this combination instead. Therefore 'The same username and password used to login to Win10 would make sense' is utter BS

So much confusion spread again...

https://support.microsoft.com/en-us/help/4046019/

Those SKU for larger organizations fortunately disabled unauthenticated accesses at the client side by a policy (to enforce security best practices in organizations).
This is not related to joining a Domain or not (same goes for Credential Manager). So two Windows 10 SKU disallow guest logons and require an adjusted policy, the other Win10 variants still allow it and even try guest logon as fallback if the authenticated attempt failed

Windows 'strategy' is to either try the local logon credentials for authentication or those that are stored in Credential Manager for the server in question (if available), then do a fallback to guest logon if allowed. In situations where clients and servers joined the same domain Windows will try to use Kerberos (so not sending any logon credentials at all over the wire but using Kerberos tickets) unless the logon attempt is made in a stupid way (unfortunately some older gentlemen constantly try to propagate using IP addresses instead of names and encourage sabotaging local name resolution while having no idea what a 'domain' really is).

In OMV you simply increase Samba's Log Level and then watch what happens when a client connects, the above Microsoft KB article explains where/what to look for in Event Viewer.