Loading ...
Sorry, an error occurred while loading the content.

build problem with opkg-cl hanging

Expand Messages
  • Steve
    I have not been able to work around the issue of opkg-cl wanting to know what to do about overwriting dev_table when building the rootfs. I have run the script
    Message 1 of 6 , May 7, 2010
    • 0 Attachment
      I have not been able to work around the issue of opkg-cl wanting to know what to do about overwriting dev_table when building the rootfs. I have run the script in the directory and it builds the images but, when I download and boot an image, it always stops saying 'unable to open initial console' which I think is a nice way of saying there is no initrd. Anyone care to whack me with a cluebat on this?
      Thanks,
      Steve
    • Mike Westerhof (mwester)
      ... A recent fix committed into the OE repository had the side-effect of breaking SlugOS builds. I have not had time to investigate a fix, but since at
      Message 2 of 6 , May 7, 2010
      • 0 Attachment
        Steve wrote:
        > I have not been able to work around the issue of opkg-cl wanting to know what to do about overwriting dev_table when building the rootfs. I have run the script in the directory and it builds the images but, when I download and boot an image, it always stops saying 'unable to open initial console' which I think is a nice way of saying there is no initrd. Anyone care to whack me with a cluebat on this?
        >

        A recent "fix" committed into the OE repository had the side-effect of
        breaking SlugOS builds. I have not had time to investigate a fix, but
        since at least one other person besides myself is running into this
        problem, I will raise the priority of trying to resolve this. It is
        complicated by the fact that presumably the change fixed a problem for
        someone else, so I cannot just revert their change...
        -Mike
      • Hans Henry von Tresckow
        On Fri, May 7, 2010 at 12:02 PM, Mike Westerhof (mwester) ... Please try the attached patch. I looked at the logs, and opkg-cl was hanging while waiting for
        Message 3 of 6 , May 8, 2010
        On Fri, May 7, 2010 at 12:02 PM, Mike Westerhof (mwester) <mwester@...> wrote:
         

        Steve wrote:
        > I have not been able to work around the issue of opkg-cl wanting to know what to do about overwriting dev_table when building the rootfs. I have run the script in the directory and it builds the images but, when I download and boot an image, it always stops saying 'unable to open initial console' which I think is a nice way of saying there is no initrd. Anyone care to whack me with a cluebat on this?
        >

        A recent "fix" committed into the OE repository had the side-effect of
        breaking SlugOS builds. I have not had time to investigate a fix, but
        since at least one other person besides myself is running into this
        problem, I will raise the priority of trying to resolve this. It is
        complicated by the fact that presumably the change fixed a problem for
        someone else, so I cannot just revert their change...
        -Mike

        Please try the attached patch. I looked at the logs, and opkg-cl was hanging while waiting for the "user" to decide which device tables to use. I added -force-defaults to  IPKG_ARGS and the build worked. I haven't tested the image yet though. If this works, I'll send the patch to the OE list.



        --
        Henry von Tresckow (hvontres)
      • Hans Henry von Tresckow
        On Sat, May 8, 2010 at 2:50 PM, Hans Henry von Tresckow ... Please try this version of the patch instead, the previous picked up some extra stuff.... need to
        Message 4 of 6 , May 8, 2010
        On Sat, May 8, 2010 at 2:50 PM, Hans Henry von Tresckow <hvontres@...> wrote:
         
        [Attachment(s) from Hans Henry von Tresckow included below]



        On Fri, May 7, 2010 at 12:02 PM, Mike Westerhof (mwester) <mwester@...> wrote:
         

        Steve wrote:
        > I have not been able to work around the issue of opkg-cl wanting to know what to do about overwriting dev_table when building the rootfs. I have run the script in the directory and it builds the images but, when I download and boot an image, it always stops saying 'unable to open initial console' which I think is a nice way of saying there is no initrd. Anyone care to whack me with a cluebat on this?
        >

        A recent "fix" committed into the OE repository had the side-effect of
        breaking SlugOS builds. I have not had time to investigate a fix, but
        since at least one other person besides myself is running into this
        problem, I will raise the priority of trying to resolve this. It is
        complicated by the fact that presumably the change fixed a problem for
        someone else, so I cannot just revert their change...
        -Mike

        Please try the attached patch. I looked at the logs, and opkg-cl was hanging while waiting for the "user" to decide which device tables to use. I added -force-defaults to  IPKG_ARGS and the build worked. I haven't tested the image yet though. If this works, I'll send the patch to the OE list.



        --
        Henry von Tresckow (hvontres)
        .\\


        Please try this version of the patch instead, the previous picked up some extra stuff.... need to work on my git-wondo.... :)

        --
        Henry von Tresckow (hvontres)
      • Steve
        I added the arguments to slugos_image.inc and it finished building and works! Thanks for finding and fixing this issue. I didn t apply it via the patch system
        Message 5 of 6 , May 8, 2010
        • 0 Attachment
          I added the arguments to slugos_image.inc and it finished building and works! Thanks for finding and fixing this issue. I didn't apply it via the patch system because it looked so straightforward and I wasn't sure how to do it. FYI, the build was complete except for the last steps.

          There are a few issues with 5.4-alpha however. It is not possible to log in via telnet or ssh. In both cases, the connections are refused. It also got its IP address via DHCP which is fine but took a moment to find it. When the network comes up, /etc/udhcpc.d/50default complains that it can't create /etc/resolve.conf because there is no room on the device. 'df' indicates that /dev/root is 6400 blocks big, 6400 blocks used and 0 available. Sounds like that might be the issue. All information from the serial port of course.

          Thanks again for getting me this far,
          Steve

          --- In nslu2-linux@yahoogroups.com, Hans Henry von Tresckow <hvontres@...> wrote:
          >
          > On Sat, May 8, 2010 at 2:50 PM, Hans Henry von Tresckow
          > <hvontres@...>wrote:
          >
          > >
          > > [Attachment(s) <#12879eb90471d3e6_TopText> from Hans Henry von Tresckow
          > > included below]
          > >
          > >
          > >
          > > On Fri, May 7, 2010 at 12:02 PM, Mike Westerhof (mwester) <mwester@...
          > > > wrote:
          > >
          > >>
          > >>
          > >> Steve wrote:
          > >> > I have not been able to work around the issue of opkg-cl wanting to know
          > >> what to do about overwriting dev_table when building the rootfs. I have run
          > >> the script in the directory and it builds the images but, when I download
          > >> and boot an image, it always stops saying 'unable to open initial console'
          > >> which I think is a nice way of saying there is no initrd. Anyone care to
          > >> whack me with a cluebat on this?
          > >> >
          > >>
          > >> A recent "fix" committed into the OE repository had the side-effect of
          > >> breaking SlugOS builds. I have not had time to investigate a fix, but
          > >> since at least one other person besides myself is running into this
          > >> problem, I will raise the priority of trying to resolve this. It is
          > >> complicated by the fact that presumably the change fixed a problem for
          > >> someone else, so I cannot just revert their change...
          > >> -Mike
          > >>
          > >
          > > Please try the attached patch. I looked at the logs, and opkg-cl was
          > > hanging while waiting for the "user" to decide which device tables to use. I
          > > added -force-defaults to IPKG_ARGS and the build worked. I haven't tested
          > > the image yet though. If this works, I'll send the patch to the OE list.
          > >
          > >
          > >
          > > --
          > > Henry von Tresckow (hvontres)
          > > .\\
          > >
          > >
          > >
          >
          > Please try this version of the patch instead, the previous picked up some
          > extra stuff.... need to work on my git-wondo.... :)
          >
          > --
          > Henry von Tresckow (hvontres)
          >
        • Hans Henry von Tresckow
          ... have to take another look at the free space issue. -- Henry von Tresckow (hvontres)
          Message 6 of 6 , May 8, 2010
          • 0 Attachment
            On Sat, May 8, 2010 at 4:19 PM, Steve <steve@...> wrote:
             



            I added the arguments to slugos_image.inc and it finished building and works! Thanks for finding and fixing this issue. I didn't apply it via the patch system because it looked so straightforward and I wasn't sure how to do it. FYI, the build was complete except for the last steps.

            There are a few issues with 5.4-alpha however. It is not possible to log in via telnet or ssh. In both cases, the connections are refused. It also got its IP address via DHCP which is fine but took a moment to find it. When the network comes up, /etc/udhcpc.d/50default complains that it can't create /etc/resolve.conf because there is no room on the device. 'df' indicates that /dev/root is 6400 blocks big, 6400 blocks used and 0 available. Sounds like that might be the issue. All information from the serial port of course.

            Thanks again for getting me this far,
            Steve

            I am glad it worked that far. I'll send that patch upstream then . We'll have to take another look at the free space issue.

            --
            Henry von Tresckow (hvontres)
          Your message has been successfully submitted and would be delivered to recipients shortly.