Beiträge von Rd65

    no hostname.. hm... do a
    #hostnamectl set-hostname host.example.com
    something like #hostnamectl set-hostname openmediavault.local
    as desribed in your hosts file. and restart.
    i think this will fix it and nginx will start too.

    ok now i found a easy way to manage alternate configs via default and env-vars.
    but why nobody hints about this if someone asked for?
    i think my problems are fixed now.. or made by myself :) maybe possible the omv4 and I still find together at the future.

    Well, I wanted to use omv to save time ... in the meantime, I'll spend more time getting it right as I need it. That bothers me. Yes I missed the parameter in the Docu, you can now make fun of it for days and laugh. If I have more time, I might really sit down and try to write something for omv - but I already know what I'll hear ... "that's not the design goal of omv".


    But if i ask a precise question and get fray Bullshit as Answer...that realy annoys me. Sorry. Simply ask ryecoaaron how long and how many posts does it takes to get a simple info about "omv-mkonf mdadm".


    Because of salt, I think this software is more suitable for configuring server farms (cluster management) than controlling local daemons with simple textfiles. It may work, but it does not make things any easier and is incomprehensible like the Windows registry in conjunction with AD servers. Something like that does not need a Nas and that is not Linux philosophy too.


    Maybe i'm wrong about Salt... but: https://en.wikipedia.org/wiki/KISS_principle


    its not my job to decide which software or scripts is used in omv, but its my decission to use it or not. So i will not update to and even not develop on omv5. You say, there is no more development on omv4.. ok.. all things told.
    And yes, maybe it was not a good idea to use omv for my needs... maybe i do it in the future by hand again.. as in the last 30 Years! I do not complain about that.

    slowly....
    logging= is invalid. they speak about logging in the docu but there was and there is no parameter "logging=" inside!
    printer admin is deprecated since samba 3.6 and it was anounced as dead earlier.
    thats what i told.
    maybe the omv 5 as officialy called "beta" may have'nt these problems - but omv4 is the stable release.
    but i don't want to talk about update and release politics or how riding a dead horse (saltstack), i want to complain about a wrong, not documented and missleading parameter "logging=" in smb.conf. And its not a theoretical discussion, it manifests on wrong logs or better 0-logs. everyone cant try on its own system! this is what i described and you and Volker can do what ever you want with this proven info... fixing this isue is one of lot's of possibilitys.
    maybe you can't fix this isue because of Volker... whyever.. maybe he is the biggest linux guru on earth.. but you can build your own Opinion about that and check/proven my arguments.


    And now again... logging= is not a valid config entry. show it in the docu.. or remove it from the /usr/share/openmediavault/mkconf/samba.d/10global!


    If you (or Volker) show me a valid reference for logging=syslog or whatever it takes as parameter, i will immediately stop complaining and ask the guys from samba, why they dont support doccumented parameters... but now it's on OMV.

    maybe inotify is your friend...or auditd. Or lsof. Or iotop.. or iostat...
    try to insert the usb disk but not mount it and disable smart. now it stops after few mins. or is it running?
    if it stops, mount it... now it stops after few mins. or is it running
    if it stops, enable smart. now it stops after few mins. or is it running?
    if its running... which damons running in the testtimeframe? take a look on /var/log/daemon.log and /var/log/syslog and var/log/kern.log and #dmesg
    You can even do a #init 1 (switching to singleusermode) on ssh to see if it depends on deamons on the runlevels. and so on...
    i woud try at first #iotop -o -d 30

    logging=syslog is simply wrong!
    please show me only one official link to samba doku which describe a config item and parameter as "logging=syslog" and i never more complain about this.
    Yes, there are parameters like syslog=0 ... but this is not logging=syslog.
    And no, i will not do discussions about this on github! Its your job to rtfm sambadocus because you build configs by your design!
    And even no, i don't know why testparm do not compalin about the config/parameter logging=, but i see testparm complaining on very old obsoletet config entrys in smb.conf made by omv4! But Testparm is not the omniscient garbage dump!
    if this all is no reason to take a look for you ... then even github will not help you.

    i try to fix some omv errors and want to read samba logs.. - as defined in smb.conf -> log file = /var/log/samba/log.%m
    but samba create files in /var/log/samba/log.ipaddress with a leght of 0
    i found that there is a buggy entry "logging=syslog" which depends on that.


    You can use syslog... or logfiles like these in /var/log/samba/ ... but if you set both, that will go wrong!
    the problem is..."logging=syslog" and the logs are hardcoded strings in /usr/share/openmediavault/mkconf/samba.d/10global
    Now i removed the buggy "logging=syslog" and all things fine. so please - on the next omv4 update, fix that.
    i don't want to fear the next update... and do it all again...

    A lot of external Disks (via USB) use controllers (USB2SATA Bridges) which hide features of the disk and may use own (firmware depended) timeouts. its logic.. you use an usblink.. not sata... only the backend between controller and disk is sata.
    emulating something is never the same as the original.

    Befor you set params with hdparm on omv, try them in the shell.
    Many Disks do not support acousticmanagement and apm as example.
    Most Disks run unconfigured at its best. Don't disable caches...
    look at smart errors... there are lot of possible error sources by managing disks..

    the dhcp data seem to be ok.
    but the warning/error on dhcp -v point to a missconfig of the resolver daemon.
    in old times /etc/resolv.conf was a file containing a nameserverip. easy and well done.
    in modern times /etc/resolv.conf is a link to a file which is handeled/manages bei the resolv deamon.
    if this link is broken, the system complains about... and resolving may gone wrong.
    the idea was to manage the resolv.conf depending on events like upcomming and ending dialup modem connections, offlinemode and so on... but on 24/7 systems like a nas this is crap.
    in the case the host file and /or hostname is even wrong too... the System can't find its own name. Sometimes this will break nginx. maybe nss-myhostname help.. but the reason for this is a bunch of missconfigured configs.


    so the way is:
    set a hostname and the domain in omv... or at the system (/etc/hostname)
    check your /etc/hosts file

    Code
    # cat /etc/hosts
    # This file is auto-generated by openmediavault (https://www.openmediavault.org)
    # WARNING: Do not edit this file, your changes will get lost.
    127.0.0.1 localhost
    127.0.1.1 nas.local nas


    check if # hostname --fqdn
    shows a well formated hostname like this -> nas.local
    (2 names with a dot)
    now do a #dpkg-reconfigure resolvconf
    this will restore the resolv.conf-symlink
    mybe you need to delete the /etc/resolv.conf at first.
    check your /etc/network/interfaces

    Code
    # The loopback network interface
    auto lo eth0
    iface lo inet loopback
    
    
    # The primary network interface
    iface eth0 inet dhcp

    shoud look like this - but replace all eth0 with your network adapter name.
    enpXsY or whatever - depends on your hardware, udev and lot's of script crap. list with #cat /proc/net/dev


    do a #omv-initsystem
    omv will re-read configs
    #reboot


    now dhcp will setup via the daemon resolvconf a working and linked resolv.conf


    read about
    https://wiki.debian.org/NetworkConfiguration
    and
    take a look at this:
    https://askubuntu.com/question…resolv-conf-must-be-a-sym

    on a working and well managed machine i would say:


    please post the result of
    #nmap --script broadcast-dhcp-discover


    but i am afraid this will go wrong now.


    maybe you can try


    #dhcpdump -i yourinterfacename
    or
    #dhclient -v

    as i told... do the steps and it will work.
    maybe you mixup dhcp wiht dns?
    you get your adress by dhcp from the router and if your router is your dns forwarder, he knows a hostname combined wit this adress (and mac). then it dosn't matter if you get static ip (as a static lease) or a well known hostname managed by your router.
    no reason for static adresses, no reason to play on nameserver entrys. the router do it for you.
    no there is no need to reinstall.. use a backup ;)

    reset your network with omav-firstaid 1. to dhcp on ipv4 and after reboot reset 2. your webcontrolinterface and 3. use your router as dns provider. there is no need to change network config on providerchange, this is managed by the router.
    Its not a good idea to use a external dns server without knowlege because in this way clients ask 1000 things about the internal home-net to external DNS Servers .. without any sense.


    OMV4 is based on Stretch. OMV5 is based on Debian. If you install OMV on top of Debian you will end up with the same system as if you were using the image. Maybe some different packages will be installed.

    OMV5 on Buster... its all debian.
    but you can even install OMV4 on Buster(backports) with netinstalls... if you dont want to be a betatester for OMV5.

    first of all.. changing network configs while login via ssh, http or rdp most times go wrong! You pull the carpet while you stand on.
    if you use a headless PI, this makes it complex, on a Nas with desktop access its easy to change network things.
    on a Nas with Desktop call http://127.0.0.1 to do configs in omv.
    But you can change ipv6 while using ipv4 and vice versa. but this is complex too.
    the easyest way is not touching default dhcp behavior after installation and do not enter/change anything about networks in omv.
    the most systems are in well working dhcp mode after installation. There is no need to play on this configs.
    don't touch netconfigs via net.

    Does that make sense? Jessie is end of life since more than one year.

    it think it make sens because it fix issues and give time to migrate - if you want migrate.
    The issue is a wakeup call... but if you don't want migrate, you can stay on jessie and live with old software. its a decission of the root, not of paket maintainers far away to migrate.
    But I would also recommend the update.

    it may be a better solution to (re)write a plugin to manage the "networkmanager" config tool which is linux standard.
    it manages all kinds of netlinks including wifi and wifi-ap, bridging, lan, vpn, wan, ipv6... and its easy to configure.
    you can use networkmanager now by disable the interface in omv but you will not see any statitics about your net-device. (depends on monit-config i think.)


    Networkmanager configures all Links not setup in /etc/network/interfaces so you need to remove all entrys excluding lo ... (thats the reason to disable netconfig in omv too) ...