[RESOLVED] Network config on 0.5 Sardukaar

  • Hi All,


    After upgrading from 0.4 to 0.5 I pressed the Apply config button. The system crashed completelly and no network adapter can be found anymore. I had to reinstall OMV 0.5 from scratch and I pressed the same button again after makijng a netowrk config...and again, complete crash of the system, network adapters cannot be found.

  • So we can help;


    What kind of drive is being used as the OMV system drive? Should be an HDD or SSD no USB sticks. See my signature for why.


    Are you doing an upgrade from 0.4 to 0.5 or are you installing from an ISO?


    What if any errors are shown when the system crashes?


    When the system crashes does it freeze, attempt to reboot, shutdown.

  • I'm using a 120GB ATA Maxtor disk as system drive. I also have 4x2TB WD Green drives and 2x 1TB WD Green drives. I have 2 LVM's configured. I also have 3 (3COM) PCI network adapters configured as balance-RR to a Mikrotik router. I think these network adapters could be a problem, since I had no troubles upgrading in virtual enviroment as a test, but it failed on production.
    OMV server is Intel I5 with 8GB ram.


    1. I tried to upgrade using omv-release-upgrade. Everything went OK without errors. Then I tried to recreate bond0 and this is where the troubles started. As soon as I input all the config data (in GUI) for my bonding config and pressed apply config button, the button started rolling (spinning) and nothing happened until I got an error message that OMV can no longer be contacted. I tried pinging it to no success. Went to a console and saw, that the only IP set was at lo0 with a local IP 127.0.0.1 and no other network config was setup. Then I rebooted this machine and a lot of services just couldn't get up (I can make some screenshots when I come home). I had to was cca. 20 minutes to be able to login to OMV, where there was no network configured.


    2. I removed all the raid drives, only left system drive in. Burned the 0.5 ISO and started to reinstall OMV. All went well, network adapters were found. Again I setup a bonding interface in balance-rr mode and pressed apply config button and the exact same thing happened. At this point I still had no other drives connected. Rebooted the machine again to see same errors and had to wait 20 minutes again, to see that the network config was again absent.


    3. I reinstalled 0.4 and updated it to the latest version, setup bonding and everything works as it should.


    EDIT:


    I even tried to install 0.5 again and only tried to change the IP to static on eth0 without bonding....But as soon as I pressed apply-config, same problems.

  • Ok, so I again upgraded OMV from 0.4 to 0.5. The upgrade went OK, rebooted comupter beforemaking any changes. Computer rebooted OK. Then I only changed IP address on bond interface (it could just aswell be a normal eth port - same thing happens) and same thing happened yet again. ifconfig only return lo0 interface with IP 127.0.0.1 and if I try to restart network it says" Rerconfiguring network interfaces.../etc/network/interfaces:21: duplicate option ifdown: couldn't read interfaces file "/etc/network/interfaces". The file is there and if I open it it has exactly same setting inside as the .old file I created just before upgrade. After this error I reboot again. NFS, Samba startup failed (I believe because it couldn't find a valid IP), OMV is inaccesible, only CLI works on console. Then I get to login window (it didn't take 20 mins now, but was a normally timed startup). Now all I see is spam on monitor consisting of 1 message "monit: Cannot connect to the monit daemon. Did you start it with http support?"


    Then I ran omv-firstaid after finally I could login and choose to restore backup config and rebooted. Same errors, network and network services unavailable (which is logical since the network is down).


    I cannot reproduce this problem in a virtual enviroment. The computer is almost new and last night went through a stress test and passed it with flying colors.


    Any ideas?

  • I removed all 3COM network adapters and it works. But its strange it works with same adapters without a problem on 0.4. Did anyone try, does 4-port Intel E1000 work ok with OMV 0.5?


    EDIT: Wrote too fast. As soon as I changed IP on onboard eth0 adapter from dhcp to static and pressed apply-config same problems again.... I can't change IP address over GUI on any interface without a network crash. These adapters all work 100%!


    EDIT: I just inserted a Planet low-cost gigabit card and disabled onboard adapter. I can now change IP without crash.

  • Looks like a driver issue with the 3COM's, which type?
    Intel 2-port E1000 works pretty well (One port onboard for Web-UI, the Intel is connected to a VLAN in another network with bonded links), I just updated this afternoon from 0.436 to 0.58 without any problems. Even the LVM configuration with four 1TB drives passed the update (I totally uninstalled every plugin, even the LVM one), made the update and reinstalled the LVM, everything is still in place.

    Homebox: Bitfenix Prodigy Case, ASUS E45M1-I DELUXE ITX, 8GB RAM, 5x 4TB HGST Raid-5 Data, 1x 320GB 2,5" WD Bootdrive via eSATA from the backside
    Companybox 1: Standard Midi-Tower, Intel S3420 MoBo, Xeon 3450 CPU, 16GB RAM, 5x 2TB Seagate Data, 1x 80GB Samsung Bootdrive - testing for iSCSI to ESXi-Hosts
    Companybox 2: 19" Rackservercase 4HE, Intel S975XBX2 MoBo, C2D@2200MHz, 8GB RAM, HP P212 Raidcontroller, 4x 1TB WD Raid-0 Data, 80GB Samsung Bootdrive, Intel 1000Pro DualPort (Bonded in a VLAN) - Temp-NFS-storage for ESXi-Hosts

  • Hmm... I cannot believe that only the 3COM are the problem, they are compatible with Debian since Kernel 2.4. The 3C2000-T is basically the same NIC as the 3C940 and I have been using this NIC type for years with Linux (As being a 3COM NPE I was nearly married with this company ;) ).


    IMHO there is a problem with link bonding and the Sardaukar release of OMV. As written above I upgraded from 0.436 but link bonding with exactly the same hardware does not work anymore. The upgrade itself went through without problems, but link bonding does not work.


    To check that nothing remained from the old version I did a fresh install from scratch, but to no avail. I can set up a bond, but then the problem begins. When I click on add a interface, I can choose the two interfaces of the Intel 2port E1000, but cannot enter any value, i.e. the IP address or a bond name. When I save it and click on the rotating arrow and then refresh the page, a new bond was created, automatically named as bond0. Now I can change the values, enter an IP address and so on. Then save, rotating arrow and then - nothing. Communication failure.


    The interfaces are named as eth0, 1 and 2. Eth2 was choosen as the management IF during install, eth0 and 1 are on the separate Intel NIC. These two ports are configured to be used in a VLAN connected to a VMWare hosts-only network (192.168.100.). Typing ifconfig eth2 at the console shows that it still has the original IP (192.168.10.) but I cannot connect to the Web-UI. Opening omv-firstaid and set the eth2 to the original address let me connect to the Web-UI again, but the bond than has no IP address. Setting the IP address of the bond with omv-firstaid or at the Web-UI results in a communication failure again.
    Now I can't even delete the bond at the Web-UI.

    Homebox: Bitfenix Prodigy Case, ASUS E45M1-I DELUXE ITX, 8GB RAM, 5x 4TB HGST Raid-5 Data, 1x 320GB 2,5" WD Bootdrive via eSATA from the backside
    Companybox 1: Standard Midi-Tower, Intel S3420 MoBo, Xeon 3450 CPU, 16GB RAM, 5x 2TB Seagate Data, 1x 80GB Samsung Bootdrive - testing for iSCSI to ESXi-Hosts
    Companybox 2: 19" Rackservercase 4HE, Intel S975XBX2 MoBo, C2D@2200MHz, 8GB RAM, HP P212 Raidcontroller, 4x 1TB WD Raid-0 Data, 80GB Samsung Bootdrive, Intel 1000Pro DualPort (Bonded in a VLAN) - Temp-NFS-storage for ESXi-Hosts

  • I'll raise my hand as well. As soon as I try to apply an address it messes up my entire network config and I have to omv-firstaid.


    Right now I'm trying to set the bond manually on devices eth2-5

    • Offizieller Beitrag
    Zitat von "bysard"

    So a few upgrades since I started this post, but this wasn't solved.


    File a bug report on bugtracker.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Atleast 3 people (incl. me) have posted detailed info about a problem. It can easily be reproduced. Also I see that 3 moderators/admins have posted in this thread. I don't have a bugtracker account. I will downgrade to 0.4 where everything I need works as intended. Maybe some of the developers will test the bonding problem and inmplement a solution later on, then I will upgrade again. You can post a bugtarcker now that you are 100% aware of it.

    • Offizieller Beitrag

    There is only one developer of OMV (votdev) and he doesn't see every post or thread. That is why it needs to be in the bug tracker. The moderators are not able to fix bugs.


    As for bug tracker, anyone can sign up for an account. I have enough to do and will not start submitting bugs for other people unless I am effected by the bug. While there is plenty of information in this thread, the bug poster needs to know all the step to reproduce. I don't have this bug on any system or the equipment to test. So, I would not be the right one to submit it.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • 1. Votdev commented in this thread
    2. I know moderators don't fix bugs!
    3. You are not submitting bugs for me or any other user... you are submitting an OMV bug or a Linux distro bug!
    4. Version 0.4 works for me just fine, also nicer GUI
    5. Don't fix the bug for all I care

    • Offizieller Beitrag

    Nice attitude. All I said is that you should submit the bug since I don't know enough about the bug or have the bug and don't have an abundance of time.


    And even though Volker posted earlier in the thread, maybe he hasn't read the rest. This is a very busy month for him. Maybe he has it fixed as well. Who knows.

    omv 7.0.5-1 sandworm | 64 bit | 6.8 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.13 | compose 7.1.4 | k8s 7.1.0-3 | cputemp 7.0.1 | mergerfs 7.0.4


    omv-extras.org plugins source code and issue tracker - github - changelogs


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • bysard: You have to understand that in case there are further questions someone with the problem can give better answers than we can.


    So please don't blame us for asking you to report this in the bugtracker. This is how everybody should do it. Not just us Mods, but all Users.


    Greetings
    David

    "Well... lately this forum has become support for everything except omv" [...] "And is like someone is banning Google from their browsers"


    Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.

    Upload Logfile via WebGUI/CLI
    #openmediavault on freenode IRC | German & English | GMT+1
    Absolutely no Support via PM!

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!