Jump to content

viking777

Members
  • Posts

    399
  • Joined

  • Last visited

Posts posted by viking777

  1. unlike viking, I like the splash screen but I just don't seem to get one...

     

    Next time you boot, when you get to the grub screen highlight the Mandriva entry and press F2. At the end of the options line add the entry:

     

    splash=nosplash

     

    Then press enter and I bet you get the splash screen back - that is how it works for me anyway.

     

    If you want it permanently edit /boot/grub/menu.lst with the same entry.

  2. UUID's are a complete pain. I always change my fstab and menu.lst to mount by label which gets around all difficulties (dev/disk/by-label/.... and root=LABEL=....... respectively). Of course this cannot be done by default as devices don't have labels by default you have to add them manually (see man e2label) but once you have done it you can read the labels easily and it overcomes the partition renumbering problem.

     

    NB if you want to label your swapfile you need mkswap -L .....

  3. I agree with you totally Ian, one of the chores I perform after every 'One' install is to go through all the installed packages and remove those I think I don't need. Different language packs are one of the more obvious things to go. Don't get me wrong I think the removal (or perhaps non-installation would be a better term) of unwanted packages tailored to every individual machine is a blinding idea - if it works properly!

  4. Well, I thought I would install it and it also asked me about packages to remove. I can only think it was comparing the LiveCD environment that was running, versus what it would install for my hardware. Therefore, nothing too major to say yes go delete all the crap packages that it listed. It had listed nvidia, and god knows what else, which was completely irrelevant for my machine (all Intel hardware).

     

    Yeah, I know that is what it is doing, but on my system it also listed Nvidia drivers and kernels to be removed - but I do have an Nvidia card, that is where the confusion comes in. As I said in the first post I agreed to its suggestions and everything seemed to work, I still have an Nvidia driver, so I guess it knows what it is doing better than I do!!

     

    Mind you that is not hard :lol2:

  5. Thanks for all these replies guys they are arriving too fast to keep up with!!

     

    daniewicz, I tried your suggestion above and I combined it with ian's suggestion of the haldaemon restart and the Ctl/Alt/Backspace. Sorry to say that all four together do not bring back the usb_storage module.

  6. Thanks for that Ian, I have already tried that route without success. It should work I am sure, but it doesn't, I really don't know why. The only way to 'reset' udev is a full reboot and then I am back to square one.

     

    As I replied to daniewicz I think this is a lost cause, even the writer of the script hasn't found a way around it. I think the answer lies somewhere deep inside the linux kernel, so I have no knowledge of it for sure, and I guess the people that do are a little to busy to worry about such a minority interest as ZTE modems!!

     

    Thanks anyway.

  7. If you delete the udev rule and reboot. then your USB port is working right?

     

    After you do this, what modules are listed when running KInfoCenter? I thought there might be an additional module listed there that you were not aware of.

     

    Yes, when I delete the rule and reboot my usb hard drive appears again as normal, however there is no discernible difference in in kinfocenter. There is most definitely a difference in lsmod though:

     

    lsmod|grep usb
    usblp				  12480  0
    usbmouse				4896  0
    usbkbd				  6368  0
    usbhid				 41440  2 usbmouse,usbkbd
    ff_memless			  5448  1 usbhid
    usb_storage			91620  0
    scsi_mod			  129836  5 sr_mod,usb_storage,sg,libata,sd_mod
    usbcore			   121932  11 usblp,usbmouse,uvcvideo,usbkbd,usbhid,usb_storage,wacom,uhci_hcd,ohci_hcd,ehci_h
    cd

     

    As you can see the usb_storage module is now listed and that is what makes the difference.

     

    I have to say that I have gone back to the site where I found this udev script and read through some of the comments posted. There is a common agreement that it makes the modem work every time, but there is also a common agreement that it does so at the expense of other usb devices. The writer of the script agrees with this and says that he is trying to find a way around it, but obviously so far he has not. There was one reply from someone who claimed to have found a workround but I tried it and it certainly didn't work for me. If those individuals can't find a way to solve it, there is not much hope that I will!

  8. The External hard drive is listed there daniewicz but the usb_storage module is not getting loaded so it is not recognisable as a storage device.

     

    Here is some more info:

     

    modprobe -v usb_storage
    insmod /lib/modules/2.6.24.7-desktop-1mnb/kernel/drivers/usb/storage/usb-storage.ko.gz

     

     

    lsmod|grep usb
    usblp				  12480  0
    usbkbd				  6368  0
    usbmouse				4896  0
    usbhid				 41440  2 usbkbd,usbmouse
    ff_memless			  5448  1 usbhid
    usbserial			  30440  0
    usbcore			   121932  11 usblp,usbkbd,usbmouse,usbhid,uvcvideo,wacom,usbserial,uhci_hcd,ohci_hcd,ehci_hcd

     

    No sign of usb_storage so no usb devices.

     

    Also

     

    service -f haldaemon restart
    Stopping HAL daemon:											[  OK  ]
    Starting HAL daemon:											[  OK  ]

     

    Makes no difference either.

  9. Well I knew it was too good to last. After my euphoria at finding a solution to my ZTE modem problem I now find out the downside.

     

    The line in the script

     

    RUN+="/sbin/rmmod usb_storage"

     

    is doing what you might expect, it is removing the usb_storage module completely meaning I can no longer communicate with any USB device such as my external hard drive. So I now have to choose between having a working modem or a working external drive, and if I want the latter I have to delete the udev rule and reboot - not exactly convenient.

     

    I have tried

     

    modprobe usb_storage

     

    but that doesn't restore it, not sure why.

     

    Is there anyone out there with an idea as to how I can make this a little easier to deal with? I basically need a command that will replace the usb_storage module when I want to use it without a reboot.

     

    Of maybe you can think of another workround?

  10. Here is an update on my previous post.

     

    1) To my surprise the method of not adding anything related to splash screens in menu.lst seems to have worked - I still have no splash screen on boot - Hooray! To test this I booted again with 'splash=nosplash' added to the boot line and the splash screen came back. Pretty conclusive I would think although not very intuitive.

     

    2)I have solved the problem with the rc5.7.1 kernel not booting. It may well be something that only affects a small minority like myself. I will attempt to explain what went wrong in case anyone else out there is affected the same way. If you never mess around with menu.lst after installation you can probably stop reading now.

     

    To my mind trying to mount or boot anything by UUID is a laughable idea as they are not human readable. On the other hand mounting by device has its problems too, especially if you mess around with partitions a lot (they get renumbered and you can't boot). I therefore always mount and boot everything by label which solves both problems. Therefore a typical menu.lst entry on my machine will look like this:

    title man2009rc1
    kernel (hd0,4)/boot/vmlinuz BOOT_IMAGE=man2009rc1 root=LABEL=mandriva2009  resume=UUID=1bf61ac4-43b6-4aa7-90a2-f136b17b2d44 vga=788
    initrd (hd0,4)/boot/initrd.img

     

    (I don't bother changing the 'resume= entry' as I never use hibernate/suspend anyway)

     

    A typical fstab entry will look like this:

     

    /dev/disk/by-label/mandriva2008.1 /media/mandriva2008.1 ext3 defaults 0 0

     

    Now the problem with the rc5.7.1 kernel was that its menu.lst entry looked like this:

     

    title 2.6.27-desktop586rc5-7.1mnb
    kernel (hd0,4)/boot/vmlinuz-2.6.27-desktop586-0.rc5.7.1mnb BOOT_IMAGE=2.6.27-desktop586rc5-7.1mnb root=/dev/disk/by-label/mandriva2009 resume=UUID=1bf61ac4-43b6-4aa7-90a2-f136b17b2d44 splash=nosplash vga=788
    initrd (hd0,4)/boot/initrd-2.6.27-desktop586-0.rc5.7.1mnb.img

     

    The entry "root=/dev/disk/by-label/mandriva2009" works in fstab, but it doesn't work in menu.lst and this is why it wouldn't boot.

     

    Note of course that the menu.lst line is automatically generated when the kernel is upgraded, it is not hand written by me I just forgot to change it to the correct format. But you have to ask how it came to write it in that format in the first place?

     

    BTW whilst I have been writing this post I have come up with another little problem as well. Middle clicking to paste from Klipper doesn't work, you have to open Klipper, highlight the entry you want, right click in the document and select paste.

  11. I just installed Cooker rc1 today and I foolishly thought that somebody out there might be interested in my experiences of it so here goes.

     

    The install process is quite slick now, but one aspect of it worries me slightly. This is the point at which the 'One' installer tells you that you really don't need the following 'X' number of packages on your system and would you like to remove them? Now don't get me wrong, this is a great idea, but the question that leaps into my mind is simply "How much can I trust this?" I first met with this aspect on the previous beta, and I must admit my answer to the installer on that occasion was - no, you cant remove 60 odd packages. This time, I reasoned that as this was only a testing installation of a testing distro I might as well answer 'yes' and see what happened. I don't think it made a great deal of difference in the end. I did run into difficulties during the install- which I will discuss shortly, but I doubt it was due to packages removed by this process. The point is though that this is a hell of a big question to ask of a newcomer to Linux, particularly as it only has a yes/no answer and no real information as to what it might be basing its decisions on. My own personal preference would be a 'yes but' answer meaning that I will allow you to remove some of the packages but not necessarily all of them.

     

    So the install proceeded and indeed completed. My first task on installing any distro is to add Krusader to it - Linux without Krusader is like a planet without air -uninhabitable. This was where the problems started. On attempting to run either MCC or urpmi I got the following error:

     

    rpmdb: Program version 4.6 doesn't match environment version 0.105
    error: db4 error(-30972) from dbenv->open: DB_VERSION_MISMATCH: Database environment version mismatch
    error: cannot open Packages index using db3 -  (-30972)
    error: cannot open Packages database in /var/lib/rpm
    unable to open rpmdb

     

    You will see that the error message mentioned var/lib/rpm, so I took a look at this directory and found that both its owner and group were shown as 'rpm'. Now I have never heard of a user or group of that name and I certainly don't have one on 2008.1 so I decided to change the user and group to 'root' to see what happened. Initially, nothing happened, but after a reboot things went back to normal and I could update and add new software again. This error was not exactly obvious to me, to a newcomer it would have been very difficult to overcome.

     

    With the rpm database problem fixed I was able to update the installation. This involved amongst other things getting hold of the rc5.7.1 kernel. This is totally useless, I tried it on my previous install (beta2) and again on this install. On every occasion it fails with the message:

     

    Switchroot: /dev does not exist in new root
    Booting has failed

     

    One other complaint, and I have voiced this one before, I thought it had gone away, but now it is back. That is the damned splash screen that obscures the boot messages whatever you type in menu.lst. However, I may have found a way around it (although I am not sure about this yet). It seems that all the usual entries in menu.lst (nosplash, verbose, quiet, etc etc) make no difference whatsoever, but deleting the entry 'splash=xxx' completely seems (for the time being at least) to have succeeded in ridding me of this bothersome splash screen. I don't know if this is coincidence or by design, but I do wish there was some way that design changes like this could be notified to users/testers so we don't have to suffer these annoyances for weeks before we stumble upon our own workrounds.

     

    The last problem with it of course is KDE4. But then Mandriva devs are not responsible for that piece of hideous nonsense.

     

    And that is about as far as I have got for now. If you discount KDE4 it is working pretty well, but you can't discount it really because it is there. It is hideously ugly and in terms of development would be better called KDE 0.4.1 rather than its present designation. If I ran a distro the default KDE would be 3.5.9 (or 10 if that appears) and KDE 4 would be the option not the other way around - I wouldn't want to lose customers becuase of someone elses miscalculation.

  12. I did read to the end before moving it. As it's a heartwarming tale rather than a question, it belongs in Everything Linux. If you'd given it a different title, I could have put in it Tips & Tricks. :)

     

    Hey - I don't mind, I am still in a good mood from yesterday.

     

    My only thought was that if anyone was looking for fixes to zte modem problems they probably wouldn't look in this part of the forum first.

     

    Still I am sure the search functions would take care of that.

  13. First a word to the mods. I haven't posted this in the wrong place it really does have to do with networking but it covers a lot of bases, so please read to the end before you decide to move it.

     

    I have become so disillusioned with my ZTE MF622 modem that today I set out to the shops with a view to buying a replacement for it (the huawei e220 as it works better with Linux). When I got to the shop I was told that they no longer sold the Huawei but they could sell me this super new model for £30 more that will make my life complete. I didn't even bother to ask how well it worked with Linux I just left the shop slightly disappointed.

     

    When I got home I found on my doorstep a cheque for a four figure sum - I had won on the premium bonds!!! Now I have had premium bonds for the best part of 50 years, and although I have won in the past, it has never been more than the mimimum amount (£50). Now I don't think that Bill Gates need worry about me knocking him off the top of the rich list, but it really does feel good to win something like that.

     

    (For the benefit those not familiar with premium bonds it is a little like a lottery but you only buy the ticket once and it is entered into the draw every month for as long as you keep it. When you decide you don't want it any more you can cash it in for the amount that you paid for it.)

     

    Anyway, as if that wasn't enough, later on I went online looking for ways to solve my ZTE modem problem (it had stopped working with 2008.1 altogether) and after very little searching (which somewhat goes to make up for the 12 months I have spent searching up to now) I actually found a fully workable solution which I will share with you now. First let me make it clear that I have nothing whatsoever to do with formulating this solution - I wish I had that knowledge, but I just don't, so full credit goes to the person on whose blog it was posted :

     

    http://blog.ufsoft.org/2007/11/30/zte-mf62...dem-under-linux

     

    In case the link doesn't work for you the fix involves using this script:

     

    ACTION!="add", GOTO="ZTE_End"
    
    # Is this the ZeroCD device?
    SUBSYSTEM=="usb", SYSFS{idProduct}=="2000",
    SYSFS{idVendor}=="19d2", GOTO="ZTE_ZeroCD"
    
    # Is this the actual modem?
    SUBSYSTEM=="usb", SYSFS{idProduct}=="0001",
    SYSFS{idVendor}=="19d2", GOTO="ZTE_Modem"
    
    LABEL="ZTE_ZeroCD"
    # This is the ZeroCD part of the card, remove
    # the usb_storage kernel module so
    # it does not get treated like a storage device
    RUN+="/sbin/rmmod usb_storage"
    
    LABEL="ZTE_Modem"
    # This is the Modem part of the card, let's
    # load usbserial with the correct vendor
    # and product ID's so we get our usb serial devices
    RUN+="/sbin/modprobe usbserial vendor=0x19d2 product=0x0001",
    # Make users belonging to the dialout group
    # able to use the usb serial devices.
    MODE="660", GROUP="dialout"
    
    LABEL="ZTE_End"

     

    Which you save in: /etc/udev/rules.d/15-zte-mf622.

     

    I have tried it in four distros now and it works in every one (it takes about 20 seconds to happen after you plug the modem in, but previously the only way to get this modem to work was to cold boot with it plugged in beforehand and as i mentioned before it had stopped working in 2008.1 completely). My ZTE modem now has better functionality than my existing Huawei.

     

    So not only have I won a four figure sum, but I have saved myself another £50 on a new modem and both on the same day.

     

    They don't get much better than that, at least not in my experience :D :D

  14. Hi web, thanks for reminding me of that program. I used to use VMC when it first came out and it does work well with the Huawei. I didn't know it had made it to the Mandriva repos though. Unfortunately for me it doesn't work with my other modem - a ZTE model. It is recognised, but won't connect (in fact the connect button is greyed out, perhaps because this modem is used on another network I am not sure). The point is though that drakconnect handles these modems pretty well on its own once you overcome the problem I mentioned in my opening post, that is the main reason I switched to using Mandriva in the first place (although using VMC does give you a few 'extras' like sms messaging I grant you).

     

    Lately though Mandriva 2008.1 has actually lost the ability to detect my ZTE modem at all(except as a storage device). The last time I used it (in July) it worked and now it doesn't, although Cooker detects both modems correctly. I'm at a loss to explain why that is.

  15. In every version of Mandriva that I have used (2008, 2008.1, 2009)I have run into the same problem with my Huawei e220 modem. When you first plug it in it is recognised and the light moves from 2 flashes/sec to 1 flash/sec which is its way of telling you that it has a network to which you can connect. However on each version, within about 60 seconds, this has changed back to 2 flashes/sec which means it has reset itself. This behaviour makes the modem completely unusable in any default install since 2008.

     

    Way back in 2008 I discovered the solution, and this is it.

     

    You go here:

     

    http://nocturno.nsk.pt/scriptpages

     

    You follow the top link which mentions the Huawei e220, you copy the file and add it to /usr/share/hal/fdi/preprobe/20thirdparty. Your modem will then work normally.

     

    If I were a Mandriva dev, firstly I would investigate why this simple piece of code is not in Mandriva by default and secondly I would send a heartfelt email of congratulation (maybe even a donation?) to the writer of this piece of code in recognition of his bailing out every version of Mandriva since he wrote the script.

  16. Looks like I am not the only one then. Although 2.6.27.5-1 is running well for me now despite the fact it has KDE4 on it. (it did survive a reboot)!

     

    An interesting fact that I have just found out is that I problably did have internet access through the chroot platform, but in fact the repositories I was using in urpmi.cfg were off line - i found this out when I tried to update it using MCC from within the distro. I changed repositories and this resulted in it pulling in quite a few updates one of which was 2.6.27.5-2.1 but now this won't boot!! Anyway enough for one day, I'll stick with 5.1 for the time being.

     

    Thanks again ffi.

     

    EDIT. Correction 5-2.1 does work - I just had a typo in menu.lst.

  17. Well I just don't know what to say about this.

     

    I liked the idea of ffi's suggestion to use chroot, so I did. I then tried to run 'urpmi --auto-update' but all I got was 'no connection' even after I transferred the 'resolv.conf'. I gave up on this, tried booting again and it came up straight away :huh: ??.

     

    So I guess this is solved, although I really have no idea how, and I am almost prepared for it to fall over again next time I reboot!

     

    Anyway thanks for your help.

  18. Thaks for that Lex, but I already have kernel 2.6.26 installed and it doesn't boot either - exactly the same error.

     

    Out of interest I copied rc.sysinit from Mandriva 2008.1 and booted with that instead. (Don't try this at home kids!! :lol2: ) The boot failed at exactly the same section - ie the fsck section (although on the 2008.1 file this is line 676 not 669 although the coding looks very much the same to my untrained eyes).

     

    I don't know if that means anything to anyone.

     

    I will probably have to end up reinstalling my disk image again, but I will leave it open for a few days just to see if anybody comes up with anything.

  19. Last night I did some updates that enabled the 2.6.27 kernel to work with the Nvidia graphics card. They worked, and although I had a bit of trouble with 'nvidia-settings' freezing while I was trying to configure, I eventually got it to work as it did with the previous kernel (ie. pretty well). This morning I booted again and after waiting about 5 minutes for that damned splash screen to go away (something really has to be done about that) I got the following error:

     

    Checking Root Filesystem.

    etc/rc.d/rc.sysinit : line 669: 2712 segmentation fault

     

    You then get the usual screen about entering the root password for maintenance. I did this and ran e2fsck with the -fp options and it completed with no errors.

     

    Any other ideas as to what I might do? I doubt very much I would be able to tell what the error was in rc.sysinit (it has over 1000 lines). I could replace it, but where would I get a replacement from? I could reinstall an image from about a week ago but that is only a last resort really as I will lose so many updates.

     

    Any ideas?

     

    EDIT. I just had a look at rc.sysinit from another distro. this is what it looks like:

     

    Fsck()

    {

    fsck $*

    rc=$?

     

    if [ "$rc" -eq "0" ]; then

    echo_success

    echo

    elif [ "$rc" -eq "1" ]; then

    echo_passed

    echo

    elif [ "$rc" -eq "2" -o "$rc" -eq "3" ]; then

    gprintf "Unmounting file systems\n"

    umount -a

    mount -n -o remount,ro /

    gprintf "Automatic reboot in progress.\n"

    reboot -f

    fi

     

    Line 669 is the opening curly bracket after the line with 'fsck'. Is there something wrong with that? (it does have a corresponding closing bracket way down at line 711)

×
×
  • Create New...