Beiträge von iulianv

    But you can see how the command is being built. It isn't hard to change it.

    I know what grep , awk and sed are, and after the examples I've been given in this thread I sort of understand in theory what the command does. The regular expression stuff is the difficult part :)


    Code
    root@openmediavault3:~# cat /sys/devices/virtual/thermal/thermal_zone0/temp
    cat: /sys/devices/virtual/thermal/thermal_zone0/temp: No such file or directory

    The motherboard is an ASUS SABERTOOTH 990FX R2.0

    I checked the BIOS and the temperature of the cpu is listed as 45 C which seems normal since while you're in the bios there's no OS cpu power management going on.


    I ran again sensors-detect and rebooted but nothing changed.

    Yes that worked too.

    But the temperature is definitely wrong. I don't think it's the script, probably something wrong with lmsensors.



    This is an AMD FX-9590 one of the hottest cpus of its time. (and by hottest I mean literally hottest, TDP 220W)

    As a hunch maybe the temperature is probably the

    Code
    temp1: +27.0°C

    Worked like a charm too :))


    OK, last one ^^


    Sorry, I didn't mean to belittle your work in any way. I aplologize if you thought that.

    Would it not be a good idea to somehow gather the scripts for the various mainboards in this thread in a wiki or something so people can refer to it ? (rather than sifting through this thread I mean)


    Soo... that being said.


    How about this one?


    cpu-temp says:


    Code
    root@openmediavault3:~# cpu-temp
    Specified sensor(s) not found!

    Hi


    I don't know if this code fits, try it, you don't have a standard board

    With an AMD I don't use it, it seemed very small

    Code
    sensors nct6776-isa-0290 | grep 'CPUTIN' | cut -c 10-25 | grep '[0-9.]\+' | sed 's/\.//'

    The board is an Asrock 970 Pro3 R2.0

    I followed the guide and your suggestion and now I get this:


    Maybe I should not have set the DIVISOR variable?

    I can't seem to unset it now.

    If you are replacing the hard drive you can simply redirect the shared folders to the new hard drive. That way you don't need to delete folders and create them again. The rsync jobs would not need to be touched in this case, when you modify the shared folder the rsync job will already be correctly configured in the new shared folder reference.


    To redirect a shared folder you just have to edit it and modify the relative path. Use the tree by clicking the icon on the right in the GUI to avoid making mistakes.

    Yes but I don't have one partition per HDD, I have a bunch of HDDs mounted as one btrfs partition. Before btrfs, in OMV6 and OMV5 I used to use LVM but it was a lot of work, physical volumes, logical volumes and then also an ext4 or xfs on top of that (if I remember correctly). Quite a headache. Btrfs simplified all that.

    Maybe I should look into other ways to have one big partition spanning multiple HDDs? Mergefs? Unionfs? I have no experience with those.

    I have a failed disk in a btrfs linear array which I have to remove. btrfs device remove does not work becaure of read errors on that drive. I lookedt at dmesg and once the command reaches the problematic part of the drive at first it tries to go down SATA levels until it can't anymore, throws a bunch of IO errors and marks the fs read-only. I also tried simply deleting the files in the terminal but the result is the same.

    So I decided to remove the fs entirely.

    I have quite a few shared folders/samba shares/rsync targets and jobs configured.

    Is there really no way to remove and recreate the filesystem without also deleting and recreating all those above? It's really tedious work :( Or maybe a way to tell btrfs I want to remove that drive disregarding whatever data I might lose in the process?


    PS. Checkboxes to remove multiple rsync jobs/samba shares all at once would be really nice :)

    So I installed the proxmox kernel in the hope it would better support a pcie SATA adapter in my system. No it didin't :))

    Now my network graphs in the performance statistics show zero activity.

    "ip link show" now tells me my interface is no longer called "enp3s0" but "enp4s0". Otherwise the system works fine.


    In the network interfaces menu I see my nic with the old name and its static ip address.

    I also see that I can add another interface based on the new enp4s0 name. I debate internally if I should do that and if it would break something. I decide to risk it.

    I add a new interface with a different static ip. I close the page and restart the browser. I can now access the system with the new ip. Reboot. Go back to the interfaces menu and delete the old one wich was no longer showing graphs. Go to the new one and change the static ip to the previous one since I have rsync jobs which depend on ip address. Reboot.

    Now my graphs are back. Hope this helps .

    The respective columns do not appear in Storage/Disks (yes I know how to hide/show columns) as announced in the 7.3.0 changelog.


    Just out of curiosity I also checked in Storage/SMART/Devices. Not there either. I have 3 OMV systems with various hardware configurations, I updated them all to 7.3.0 and all seem to be missing this announced feature, which btw I find rather useful since I have always wondered if my disks spin down or not after a certain time of inactivity.

    After a lot of search I understand that we need a real SATA controller LSI card for a NAS.

    No you don't. I have used various PCI and PCIe to SATA adapters. Do a little research before you buy. From experience I can tell that you should steer clear of cards with more than 2 ports and that pretend to do some sort of raid in hardware. You need cards that are simple SATA adapters.


    These I have and work:

    ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02) (probably can't find it anymore it's a PCI card :) )

    Marvell Technology Group Ltd. 88SE9230 PCIe 2.0 x2 4-port SATA 6 Gb/s RAID Controller (rev 11) (works but like I said 4 ports not a good idea, I haven't populated all ports)

    Silicon Image, Inc. SiI 3132 Serial ATA Raid II Controller (rev 01) (2 port card, works).

    So I updated a backup server from 6 to 7 today (actually 2 servers but one uses BTRFS so this doesn't apply). Everything went fine. As I've read and also as the installer says at the end mdadm is now a plugin and has to be installed after the upgrade, from the plugins page.

    Funny thing though, once I rebooted there was my raid array all mounted and accessible. I looked at the plugins page to see if somehow the plugin got installed automatically. It wasn't.


    So, should I still install the plugin or would I be making thing worse?