Beiträge von el_viejo

    antipiot: why so rude?


    for sure asking for help requires a basic attitude of cooperation but on the other hand these days public internet is a dangerous place for private information such as ip-addresses in particular ipv6-addresses as they use to change less frequently. my opinion is hendriks concerns are quite justified.


    for me it would be helpful to know how to distinguish between internal and external ipv6-addresses so i can check it myself.



    regards


    peter

    hello all,


    i am struggling to get access to omv's webgui on it´s local ipv6 address. on the ipv4 side everything runs smooth (calling webgui by ip address or by hostname). my router is a fritzbox 6490 cable with fw 6.84 and ipv6 enabled and connected to a ds-lite cable-internet connection. on my omv-box in network settings i tried both options (auto, dhcp) for ipv6 but no success. calling webgui by the ipv6 address results in a browser error (connection refue. the strange thing i noticed: my router assigns an ipv6 address but when i try to check it i.e. using wol2 (http://oettecds.bplaced.net/download/WOL2.1_Beta1.7z) it tells me there is only an ipv4 address assigned but no ipv6.


    maybe there is something missing in my routers ipv6 configuration? any clue welcome :)



    regards


    peter

    after quite a while of being absent from omv i went back to set up a small home fileserver with omv 1.19. as i already gained some experience by using former versions everything went smooth but there is one issue. as described in the title line as soon as i activate the checkbox to force a secure connection to webgui i am no longer able to access it. in order to resolve this i have to use omv-firstaid or editing the openmediavault.xml configuration file manually. after a reboot it is accesible again. i have no idea where the problem is. any comments apreciated.


    thanks a lot


    peter

    finally i found the solution and it seems it is caused by faulty acpi settings in my motherboards bios. thanks to this post:


    https://plus.google.com/111861…5878613/posts/ad7F85TQwQL


    i was lead to this thread:


    https://bugzilla.novell.com/show_bug.cgi?id=689848#c17


    in the bios setup i change fan speed settings from auto to 50% and switched off over night my server. 12 hours later doing a cold boot it seems that finally finally kworker doesn't cause trouble any more :)


    so all users of the Intel D945GCLF2 board should be aware if there is trouble with kworker causing 100% cpu load on one processor core after a cold boot.



    greetings


    peter

    i am very sorry for having to take back my prior statement about having solved the kworker issue. i switched off the server last night and turned it on today in the afternoon and after that cold boot kworker was back again occupying one of the two cpu cores by 100% load... grrrrrr...


    anyway, i keep investigating the problem and let you know.



    greetings


    peter

    and by the way: the problem that kworker is causing 100% cpu load at cold boot seems to be fixed too. all i did is add the parameter acpi=noirq to the kernel line in /boot/grub/menu.lst. an update-grub brought an error message that there is no default boot entry selected so i did grub-set-default 0 and again update-grub and then a reboot.


    voilá kworker is behaving like it is supposed to be.



    greetings


    peter

    since i have upgraded from omnius to fedaykin it seems the script restart_anacron is not getting executed during boot. the root file system is always mounted rw right after boot and the command remountro results in a "mount / : busy". a workaround is execute /etc/init.d/restart_anacron manually after boot. both ways (starting it from the cron job menu of omv or by ssh) are working. after terminating the script the file system is mounted ro and i can do remountrw and remountro as often as i like without the mount busy error message.


    but how can i make the machine executing /etc/init.d/restart_anacron at boot time? i tried the command update-rc.d restart_anacron defaults which gives me this message:


    update-rc.d: using dependency based boot sequencing


    but it does not change anything. after next reboot file system still stays mounted rw. any clue is apreciated. by the way everything else works really smooth under this configuration: voyage linux 0.8.5 and omv 0.4.4.1. and as always: lots of thanks to volker for his great work and those who are in charge for the plugins.



    regards


    peter

    still no good news about the kworker 100% cpu load problem. what i can confirm so far is:


    kworker only behaves bad after a cold boot. the first few minutes after a cold boot kworker is doing fine but suddenly it starts eating cpu load from about 1-2% minute by minute it continues increasing until it reaches 99-100% and there it stays.


    workaround is: causing a soft boot from the webgui of omv or doing another cold boot.


    another issue i have (and for the moment i can't narrow it down whether it's because of omv or because of voyage linux): in the past i was able to start my server by sending the magic packet called wol or wake-on-lan. the mainboard itself is capable to do it and it worked pretty realiable in the past. after switching to voyage linux and omv 0.3.x i am no longer able to use wake-on-lan. the server just does not start even in the webgui wake-on-lan is enabled. i have no idea where to investigate.


    maybe someone of you guys using voyage linux and omv 0.3.x can tell me about your experience. probably you have wake-on-lan working or you can confirm it does not work for you neither.


    the thing is there is not too much information about voyage linux available it seems it has a pretty small web community. anyway i am happy using it and by this way saving the life of my flash memories :)



    greetings


    peter

    bad news. it seems the workaround mentionend one post above is not really solving the problem. i have switched off the server last night and this morning i started it again and there we are: kworker with 100% cpu load on one core. well, it seems that at least a reboot resolves the problem...


    still investigating...



    greetings


    peter

    arturmon you made my day with your clue to set remountro and remountrw as cron jobs without scheduling it. that's an awesome solution and now there is no longer need for a plugin. my happiness just gets a little reduced by process called kworker which puts permanently 100% cpu load on one cpu core. i have tried out to do the following steps taken from post #63 of this bugreport thread (you have to scroll down a lot to find the post):


    https://bugs.launchpad.net/ubuntu/+source/linux/+bug/887793


    it says:


    Code
    Putting the following line into /etc/modprobe.d/local.conf helped for me:
    
    
    options drm_kms_helper poll=N


    after rebooting the machine it seems that kworker does no longer consume 100% it is more around 0-2%. i am going to monitor this behavior and will report here if something happens.



    greetings


    peter



    ps: thank you guys of this thread who pointed me in the right direction to run omv on voyage linux and of course special thanks to volker for his great work.

    it is really strange, after several reboots it seems the problem went away. i am now able to enable/disable the nfs server from the webgui when the file system is mounted r/w. i have no idea what the problem was. anyway...


    another thing i found out today is that after boot the file system is not getting remounted r/o. it stays r/w and i have to access the system via ssh manually calling remountro. i followed all steps of the given howto in this thread but it seems something prevents execution of restart_anacron during boot up or something prevents the remountro command at boot time. still have to find out what it is.



    greetings


    peter

    while setting up omv on voyage linux i tried to enable nfs file services. after clicking ok on both pages (nfs settings page and nfs shares page) i get the following error message in the webgui:


    Code
    Failed to execute command 'export LANG=C; sudo update-rc.d 'nfs-kernel-server' enable 2>&1': update-rc.d: error: no runlevel symlinks to modify, aborting! update-rc.d: using dependency based boot sequencing


    clicking the show details button tells me:



    of course voyage system is mounted in r/w mode. after investigating i found out it has to do with starting/restarting the nfs-kernel-server. when i use the command /etc/init.d/nfs-kernel-server restart in a ssh shell on my server i can see the shares but by the webgui it does not work properly. maybe somebody has a clue what to do to make it work from the webgui and get rid of the error message.



    thanks and greetings


    peter

    hi guys,


    i am really excited about your approach to save flash memory lifetime using voyage linux. it takes a little more time to set up the system but it works. and even the newer kernel makes my server consuming less energy. great work. it would be perfect if somebody who knows how to do it could write a r/w plugin for omv to allow full control by the webgui.


    maybe the idea of using a r/o filesystem could influence the whole development of openmediavault. would be nice to know what the developer thinks about it.



    thank you very much and keep up the outstanding work


    peter

    it seems that the problem is solved. maybe something between the kernel and my motherboard. i have added the squeeze backports repo to the sources.list and installed the 3.2.0-amd64 kernel. during the build of the initramfs there were some messages claiming that there is some firmware missing for my r8169 nic so i installed the package firmware-realtek from the backports repo and rebootet. after reboot there was an error message claiming that neither initramfs nor the kernel could be found. after hitting a key i found myself in the grub boot menu again. so i selected the first boot entry (new kernel 3.2.0) and this time the machine was booting fine. kacpid behaves now and does no longer cause 100% cpu load on one of the processor cores.


    Code
    PID    USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
     8501 openmedi  20   0 19116 1256  912 R    2  0.1   0:00.02 top
        1 root      20   0  8404  796  660 S    0  0.0   0:01.21 init
        2 root      20   0     0    0    0 S    0  0.0   0:00.00 kthreadd
        3 root      20   0     0    0    0 S    0  0.0   0:00.13 ksoftirqd/0
        6 root      RT   0     0    0    0 S    0  0.0   0:00.00 migration/0
        7 root      RT   0     0    0    0 S    0  0.0   0:00.00 watchdog/0


    may this info help you in case of... but i still can't tell what the real problem was. i read a lot about erroneous dsdt acpi tables and how to patch them but that seemed to complicated to me so i decided to give it a try with a newer kernel.



    greetings


    peter

    i tried to upgrade from omv 0.2.x to 0.3.x but failed so installed 0.3.x from scratch on a 4gb usb flash-drive and everything is running fine so far (i am aware of the frequent writes problem while using omv on flash memory drives). yesterday i got noticed in the webgui overview of a cpu-load of 50% and further investigation narrowed it down to one of two cpu cores which is under 100% load by the process kacpid. i tried to kill that process but no success. i have no idea where this is coming from as i use standard hardware for that server (see my signature) and the same hardware has been running fine with omv 0.2.x.


    Code
    PID USER      PR  NI  VIRT  RES  SHR S   %CPU %MEM    TIME+  COMMAND
       34 root      20   0     0    0    0 R  100  0.0 116:37.54 kacpid
    19490 openmedi  20   0 19068 1296  912 R    4  0.1   0:00.03 top
        1 root      20   0  8356  840  708 S    0  0.0   0:01.13 init
        2 root      20   0     0    0    0 S    0  0.0   0:00.00 kthreadd
        3 root      RT   0     0    0    0 S    0  0.0   0:00.00 migration/0


    the syslog contains hundreds of those messages and i guess they are related to the kacpid cpu load problem:



    any clue of you experts is very welcome. i have already googled about kacpid and debian but haven't found any useful information for my exact case. and there is absolutely nothing which brings kacpid and omv together. so it seems i am the only one so far with this little inconvenience. nor did a forum search here any results.



    greetings


    peter



    ps: volker you do a great job, i really appreciate your hard work.
    pps: edited to add a screenshot of process list and extract from syslog