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

Re: [nslu2-linux] Re: 4.20 and disk power down

Expand Messages
  • Robert Hammond
    In message , Matthias Appel writes ... I have generated an /unslung/rc.sysinit script file but it seems
    Message 1 of 21 , Jun 2, 2005
    View Source
    • 0 Attachment
      In message <168038247.20050602203259@...>, Matthias Appel
      <lists@...> writes
      >Robert Hammond wrote:
      >
      >> Thanks, I now have the two scripts, rc.sysinit seems to have the
      >> relevant ram disk scripts.
      >
      >I'm also interested in this. Can you please create a wiki pages which
      >documents your steps?
      >
      I have generated an /unslung/rc.sysinit script file but it seems not to
      work as expected. Also had little time to sort this due to sick child
      in the house (now on the mend).

      The script section that I am using is basically unchanged from that used
      in 3.18 so sets up a small ram disk for /var and /dev. Something is
      failing here, a df command indicates that both /var or /dev are still
      running from hard disk!!. I am sure that this is because the ram disks
      are being cancelled in some way because /var and /dev also both exist in
      the existing disk tree structure as standard folders.

      Some advice is needed if possible

      What should I do with the existing /var /var.state /dev /dev.state
      folders in the disk tree??

      My gut feeling here is just to delete them all (booting to jffs2 and
      deleting through /mnt/sda1/) followed by a re-boot but I am not too sure
      of the outcome of this and would rather not rash my slug.

      Also not too sure of the function of /var.state and /dev.state.

      --
      Robert Hammond
      PGP:0x154144DA
    • Robert Hammond
      In message , Matthias Appel writes ... I am posting a working script here first for others to try, my
      Message 2 of 21 , Jun 3, 2005
      View Source
      • 0 Attachment
        In message <168038247.20050602203259@...>, Matthias Appel
        <lists@...> writes
        >Robert Hammond wrote:
        >
        >> Thanks, I now have the two scripts, rc.sysinit seems to have the
        >> relevant ram disk scripts.
        >
        >I'm also interested in this. Can you please create a wiki pages which
        >documents your steps?
        >
        I am posting a working script here first for others to try, my
        intention is to post to the wiki in a few days if all is well.

        Instructions-

        Copy the following script section into a new file called rc.sysinit,
        you need to locate this file in the /unslung folder located on the disk
        drive and not the one located in the flash ram. To locate the correct
        folder does depend if you are running off of the hd rootfs or not. The
        best way to guarantee the correct unslung folder is to follow
        /mnt/sda1/unslung (note - may be sda2 depending upon USB socket used for
        disk drive).
        chmod the file with permissions 755.


        #!/bin/sh
        # Intended for unslung firmware 4.x only
        # Script section to create ramdisk for /var
        mkdir -p /var/lock/samba >/dev/null

        /bin/echo "Creating ramfs for /var:"
        /bin/mount -t ramfs ramfs /mnt/tmpmnt -o maxsize=256
        /bin/cp -rp /var/* /mnt/tmpmnt/
        /bin/mkdir -p /var.state
        /bin/mount -o bind /var /var.state
        /bin/mount -o bind /mnt/tmpmnt /var
        /bin/umount /mnt/tmpmnt
        /bin/killall syslogd
        /bin/killall klogd

        # Return and continue execution of system rc.sysinit
        return 1
        # [EOF]


        Re-boot your slug

        Interestingly this ram disk does not show up when using the df command,
        but does show when using the mount command. The slug disk drive does
        seem to power down for extended periods so does seem to be working.


        --
        Robert Hammond
        PGP:0x154144DA
      • Robert Hammond
        In message , Robert Hammond writes ... Note that this does not work correctly, I just noticed that
        Message 3 of 21 , Jun 3, 2005
        View Source
        • 0 Attachment
          In message <gEdR28BifJoCFwGY@...>, Robert Hammond
          <rob.hammond@...> writes
          >In message <168038247.20050602203259@...>, Matthias Appel
          ><lists@...> writes
          >>Robert Hammond wrote:
          >>
          >>> Thanks, I now have the two scripts, rc.sysinit seems to have the
          >>> relevant ram disk scripts.
          >>
          >>I'm also interested in this. Can you please create a wiki pages which
          >>documents your steps?
          >>
          >I am posting a working script here first for others to try, my
          >intention is to post to the wiki in a few days if all is well.
          >
          >Instructions-
          >
          >Copy the following script section into a new file called rc.sysinit,
          >you need to locate this file in the /unslung folder located on the disk
          >drive and not the one located in the flash ram. To locate the correct
          >folder does depend if you are running off of the hd rootfs or not. The
          >best way to guarantee the correct unslung folder is to follow
          >/mnt/sda1/unslung (note - may be sda2 depending upon USB socket used for
          >disk drive).
          >chmod the file with permissions 755.
          >
          >
          >#!/bin/sh
          ># Intended for unslung firmware 4.x only
          ># Script section to create ramdisk for /var
          >mkdir -p /var/lock/samba >/dev/null
          >
          >/bin/echo "Creating ramfs for /var:"
          >/bin/mount -t ramfs ramfs /mnt/tmpmnt -o maxsize=256
          >/bin/cp -rp /var/* /mnt/tmpmnt/
          >/bin/mkdir -p /var.state
          >/bin/mount -o bind /var /var.state
          >/bin/mount -o bind /mnt/tmpmnt /var
          >/bin/umount /mnt/tmpmnt
          >/bin/killall syslogd
          >/bin/killall klogd
          >
          ># Return and continue execution of system rc.sysinit
          >return 1
          ># [EOF]
          >
          >
          >Re-boot your slug
          >
          >Interestingly this ram disk does not show up when using the df command,
          >but does show when using the mount command. The slug disk drive does
          >seem to power down for extended periods so does seem to be working.
          >
          >
          Note that this does not work correctly, I just noticed that my disk
          shares in Windows are missing when this script section is loaded. Have
          now removed it.

          --
          Robert Hammond
          PGP:0x154144DA
        • Robert Hammond
          In message , Robert Hammond writes ... I do have one more important observation - when using the
          Message 4 of 21 , Jun 5, 2005
          View Source
          • 0 Attachment
            In message <MWY6iRB34LoCFwCa@...>, Robert Hammond
            <rob.hammond@...> writes
            >In message <gEdR28BifJoCFwGY@...>, Robert Hammond
            ><rob.hammond@...> writes
            >>In message <168038247.20050602203259@...>, Matthias Appel
            >><lists@...> writes
            >>>Robert Hammond wrote:
            >>>
            >>>> Thanks, I now have the two scripts, rc.sysinit seems to have the
            >>>> relevant ram disk scripts.
            >>>
            >>>I'm also interested in this. Can you please create a wiki pages which
            >>>documents your steps?
            >>>
            >>I am posting a working script here first for others to try, my
            >>intention is to post to the wiki in a few days if all is well.
            >>
            >>Instructions-
            >>
            >>Copy the following script section into a new file called rc.sysinit,
            >>you need to locate this file in the /unslung folder located on the disk
            >>drive and not the one located in the flash ram. To locate the correct
            >>folder does depend if you are running off of the hd rootfs or not. The
            >>best way to guarantee the correct unslung folder is to follow
            >>/mnt/sda1/unslung (note - may be sda2 depending upon USB socket used for
            >>disk drive).
            >>chmod the file with permissions 755.
            >>
            >>
            >>#!/bin/sh
            >># Intended for unslung firmware 4.x only
            >># Script section to create ramdisk for /var
            >>mkdir -p /var/lock/samba >/dev/null
            >>
            >>/bin/echo "Creating ramfs for /var:"
            >>/bin/mount -t ramfs ramfs /mnt/tmpmnt -o maxsize=256
            >>/bin/cp -rp /var/* /mnt/tmpmnt/
            >>/bin/mkdir -p /var.state
            >>/bin/mount -o bind /var /var.state
            >>/bin/mount -o bind /mnt/tmpmnt /var
            >>/bin/umount /mnt/tmpmnt
            >>/bin/killall syslogd
            >>/bin/killall klogd
            >>
            >># Return and continue execution of system rc.sysinit
            >>return 1
            >># [EOF]
            >>
            >>
            >>Re-boot your slug
            >>
            >>Interestingly this ram disk does not show up when using the df command,
            >>but does show when using the mount command. The slug disk drive does
            >>seem to power down for extended periods so does seem to be working.
            >>
            >>
            >Note that this does not work correctly, I just noticed that my disk
            >shares in Windows are missing when this script section is loaded. Have
            >now removed it.
            >
            I do have one more important observation - when using the above script,
            the conf partition on /dev/sda2 either never mounts or is somehow
            unmounted. The side effect of this is the dreaded disk un-formated
            message in the web interface and hence all of the disk shares failing.
            Luckily removing the script will recover from this.

            Creating a ram disk for /var using an /unslung script may be impossible
            with firmware 4.20 so could be better serviced at the slingbox stage in
            a later firmware release.

            Unfortunately the reason why this script section fails to work correctly
            is beyond my knowledge so perhaps another user who wants reliable disk
            spin down could take on this project.

            --
            Robert Hammond
            PGP:0x154144DA
          • Nathan Ellsworth
            Would it be possible to put /var on a Linux-formatted USB thumb drive on the Drive 1 port and then put your spin-downable drive with all your other data on the
            Message 5 of 21 , Jun 6, 2005
            View Source
            • 0 Attachment
              Would it be possible to put /var on a Linux-formatted
              USB thumb drive on the Drive 1 port and then put your
              spin-downable drive with all your other data on the
              Drive 2 port? Just a random thought ...

              Nathan

              --- Robert Hammond <rob.hammond@...> wrote:


              ---------------------------------
              In message <MWY6iRB34LoCFwCa@...>, Robert
              Hammond
              <rob.hammond@...> writes
              >In message <gEdR28BifJoCFwGY@...>, Robert
              Hammond
              ><rob.hammond@...> writes
              >>In message <168038247.20050602203259@...>,
              Matthias Appel
              >><lists@...> writes
              >>>Robert Hammond wrote:
              >>>
              >>>> Thanks, I now have the two scripts, rc.sysinit
              seems to have the
              >>>> relevant ram disk scripts.
              >>>
              >>>I'm also interested in this. Can you please create
              a wiki pages which
              >>>documents your steps?
              >>>
              >>I am posting a working script here first for others
              to try, my
              >>intention is to post to the wiki in a few days if
              all is well.
              >>
              >>Instructions-
              >>
              >>Copy the following script section into a new file
              called rc.sysinit,
              >>you need to locate this file in the /unslung folder
              located on the disk
              >>drive and not the one located in the flash ram. To
              locate the correct
              >>folder does depend if you are running off of the hd
              rootfs or not. The
              >>best way to guarantee the correct unslung folder is
              to follow
              >>/mnt/sda1/unslung (note - may be sda2 depending upon
              USB socket used for
              >>disk drive).
              >>chmod the file with permissions 755.
              >>
              >>
              >>#!/bin/sh
              >># Intended for unslung firmware 4.x only
              >># Script section to create ramdisk for /var
              >>mkdir -p /var/lock/samba >/dev/null
              >>
              >>/bin/echo "Creating ramfs for /var:"
              >>/bin/mount -t ramfs ramfs /mnt/tmpmnt -o maxsize=256
              >>/bin/cp -rp /var/* /mnt/tmpmnt/
              >>/bin/mkdir -p /var.state
              >>/bin/mount -o bind /var /var.state
              >>/bin/mount -o bind /mnt/tmpmnt /var
              >>/bin/umount /mnt/tmpmnt
              >>/bin/killall syslogd
              >>/bin/killall klogd
              >>
              >># Return and continue execution of system rc.sysinit
              >>return 1
              >># [EOF]
              >>
              >>
              >>Re-boot your slug
              >>
              >>Interestingly this ram disk does not show up when
              using the df command,
              >>but does show when using the mount command. The
              slug disk drive does
              >>seem to power down for extended periods so does seem
              to be working.
              >>
              >>
              >Note that this does not work correctly, I just
              noticed that my disk
              >shares in Windows are missing when this script
              section is loaded. Have
              >now removed it.
              >
              I do have one more important observation - when using
              the above script,
              the conf partition on /dev/sda2 either never mounts or
              is somehow
              unmounted. The side effect of this is the dreaded disk
              un-formated
              message in the web interface and hence all of the disk
              shares failing.
              Luckily removing the script will recover from this.

              Creating a ram disk for /var using an /unslung script
              may be impossible
              with firmware 4.20 so could be better serviced at the
              slingbox stage in
              a later firmware release.

              Unfortunately the reason why this script section fails
              to work correctly
              is beyond my knowledge so perhaps another user who
              wants reliable disk
              spin down could take on this project.

              --
              Robert Hammond
              PGP:0x154144DA


              [ Moderator Note: All new information should be
              recorded in the Wiki at http://www.nslu2-linux.org ]



              ---------------------------------
              Yahoo! Groups Links

              To visit your group on the web, go to:
              http://groups.yahoo.com/group/nslu2-linux/

              To unsubscribe from this group, send an email to:
              nslu2-linux-unsubscribe@yahoogroups.com

              Your use of Yahoo! Groups is subject to the Yahoo!
              Terms of Service.





              __________________________________
              Discover Yahoo!
              Get on-the-go sports scores, stock quotes, news and more. Check it out!
              http://discover.yahoo.com/mobile.html
            • Robert Hammond
              In message , Nathan Ellsworth writes ... I have been thinking along the same
              Message 6 of 21 , Jun 6, 2005
              View Source
              • 0 Attachment
                In message <20050606173259.65445.qmail@...>, Nathan
                Ellsworth <nathan@...> writes
                >Would it be possible to put /var on a Linux-formatted
                >USB thumb drive on the Drive 1 port and then put your
                >spin-downable drive with all your other data on the
                >Drive 2 port? Just a random thought ...
                >
                I have been thinking along the same lines but have little knowledge of
                file systems located on a ram stick, my understanding is that the stick
                would be formatted as vfat which could cause unpredictable problems for
                a system folder.


                Strange how a couple of days away can generate a solution to getting the
                ram disk working.

                I have been working backwards through the boot sequence contained in
                rc.sysinit and rc.1 and have discovered that the earliest that I seem to
                be able to create a working ram disk using a slightly modified script
                with out any noticeable bad effects is at rc.bootbin :)

                The next earlier script step is rc.1 its self, the script will fail if
                booting this early :(

                Very strange because it should boot before this, most probable reason
                is that there does seem to be some strange script lines early in the
                rc.1 file in 4.20 which can un-mount /dev/sda2 under certain
                circumstances so could be some sort of conflict here.

                I just need to test for another night or two so hope to post up a
                working beta script some time over the next couple of days, followed by
                a wiki page some days later.

                Do note that some packages use the disk folder /opt/var rather than /var
                for there log files (I think that samba3 is configured this way) so some
                package.conf files made need attention for reliable disk spin down.

                --
                Robert Hammond
                PGP:0x154144DA
              • Robert Hammond
                At last a working script for others to try. Short form Instruction (requires moderate slug knowledge) : - Copy the following script section into a new file
                Message 7 of 21 , Jun 7, 2005
                View Source
                • 0 Attachment
                  At last a working script for others to try.



                  Short form Instruction (requires moderate slug knowledge) : -

                  Copy the following script section into a new file called rc.bootbin
                  using a unix style text editor, you need to locate this file in the
                  /unslung folder located on the disk drive and not the folder located in
                  the flash ram. To locate the correct folder does depend if you are
                  running off of the hd rootfs or not. The best way to guarantee the
                  correct unslung folder is to follow /mnt/sda1/unslung (note - may be
                  sda2 depending upon USB socket used for disk drive).
                  chmod the file with permissions 755.

                  #!/bin/sh
                  # /unslung/rc.bootbin
                  # Intended for unslung firmware 4.x only
                  # Script section to create ramdisk for /var

                  /bin/echo "Creating ramfs for /var:"
                  /bin/mount -t ramfs ramfs /mnt/tmpmnt -o maxsize=256
                  /bin/cp -rp /var/* /mnt/tmpmnt/
                  /bin/mkdir -p /var.state
                  /bin/mount -o bind /var /var.state
                  /bin/mount -o bind /mnt/tmpmnt /var
                  /bin/umount /mnt/tmpmnt

                  # Return and continue execution of system rc.
                  return 1
                  # [EOF]


                  --
                  Robert Hammond
                  PGP:0x154144DA
                • Rod Whitby
                  ... Actually, it would be sda2 if you unslung to the conf partition instead of the data partition. It would be sdb1 or sdb2 if you used the disk2 port instead
                  Message 8 of 21 , Jun 7, 2005
                  View Source
                  • 0 Attachment
                    On 6/8/05, Robert Hammond <rob.hammond@...> wrote:
                    > The best way to guarantee the
                    > correct unslung folder is to follow /mnt/sda1/unslung (note - may be
                    > sda2 depending upon USB socket used for disk drive).

                    Actually, it would be sda2 if you unslung to the conf partition
                    instead of the data partition. It would be sdb1 or sdb2 if you used
                    the disk2 port instead of the disk1 port.

                    -- Rod
                  • mpmarks2000
                    Thanks so much for working on this Robert. I don t really want to use NSLU2 on a 24/7 basis unless I can get those drives to sleep somehow. large capacity 7200
                    Message 9 of 21 , Jun 15, 2005
                    View Source
                    • 0 Attachment
                      Thanks so much for working on this Robert. I don't really want to use
                      NSLU2 on a 24/7 basis unless I can get those drives to sleep somehow.
                      large capacity 7200 RPM drives get very hot and sweaty if they are
                      spinning in little boxes 24/7 - they wear out fast and make my
                      computer area hot too.

                      I have followed your instructions but I have yet to see my One Touch
                      go to sleep. Looking at /proc/mounts I can see that /var is on a
                      ramfs. But df shows /var still on the disk? So I'm a bit confused if
                      the script executed properly. Otherwise everything is working
                      normally. If I power the NSLU2 down the One Touch drives go to sleep
                      on time, but never if the Slug is awake.

                      I have a couple of questions for you.
                      1) How did you discover what processes were accessing the disk?
                      2) The script echos a note when it executes but I don't see the
                      message in /var/log/messages - is that simply because of timing or is
                      there a log entry somewhere that I missed?
                      3) Do you have other suggestions on how I can troubleshoot the
                      problem?

                      thanks!
                      /maurice



                      --- In nslu2-linux@yahoogroups.com, Robert Hammond <rob.hammond@n...>
                      wrote:
                      > At last a working script for others to try.
                      >
                      >
                      >
                      > Short form Instruction (requires moderate slug knowledge) : -
                      >
                      > Copy the following script section into a new file called rc.bootbin
                      > using a unix style text editor, you need to locate this file in the
                      > /unslung folder located on the disk drive and not the folder
                      located in
                      > the flash ram. To locate the correct folder does depend if you
                      are
                      > running off of the hd rootfs or not. The best way to guarantee
                      the
                      > correct unslung folder is to follow /mnt/sda1/unslung (note - may
                      be
                      > sda2 depending upon USB socket used for disk drive).
                      > chmod the file with permissions 755.
                      >
                      > #!/bin/sh
                      > # /unslung/rc.bootbin
                      > # Intended for unslung firmware 4.x only
                      > # Script section to create ramdisk for /var
                      >
                      > /bin/echo "Creating ramfs for /var:"



                      > /bin/mount -t ramfs ramfs /mnt/tmpmnt -o maxsize=256
                      > /bin/cp -rp /var/* /mnt/tmpmnt/
                      > /bin/mkdir -p /var.state
                      > /bin/mount -o bind /var /var.state
                      > /bin/mount -o bind /mnt/tmpmnt /var
                      > /bin/umount /mnt/tmpmnt
                      >
                      > # Return and continue execution of system rc.
                      > return 1
                      > # [EOF]
                      >
                      >
                      > --
                      > Robert Hammond
                      > PGP:0x154144DA
                    • mpmarks2000
                      An update. The script does work! My drive eventually went to sleep. Its not clear to me what Unslung is doing that keeps the drive active but eventually it
                      Message 10 of 21 , Jun 16, 2005
                      View Source
                      • 0 Attachment
                        An update. The script does work! My drive eventually went to sleep.
                        Its not clear to me what Unslung is doing that keeps the drive active
                        but eventually it went to sleep. Possibly the failure of a backup job.

                        So for my penance I added Robert's instructions to the Wiki as an
                        update to the "how to set the timeout on One Touch drives". I'm still
                        searching for a solution on how to put NON One Touch USB 2.0 drives to
                        sleep from Linux.

                        Thanks again Robert.
                        /maurice

                        --- In nslu2-linux@yahoogroups.com, "mpmarks2000" <marks@s...> wrote:
                        > Thanks so much for working on this Robert. I don't really want to use
                        > NSLU2 on a 24/7 basis unless I can get those drives to sleep somehow.
                        > large capacity 7200 RPM drives get very hot and sweaty if they are
                        > spinning in little boxes 24/7 - they wear out fast and make my
                        > computer area hot too.
                        >
                        > I have followed your instructions but I have yet to see my One Touch
                        > go to sleep. Looking at /proc/mounts I can see that /var is on a
                        > ramfs. But df shows /var still on the disk? So I'm a bit confused if
                        > the script executed properly. Otherwise everything is working
                        > normally. If I power the NSLU2 down the One Touch drives go to sleep
                        > on time, but never if the Slug is awake.
                        >
                        > I have a couple of questions for you.
                        > 1) How did you discover what processes were accessing the disk?
                        > 2) The script echos a note when it executes but I don't see the
                        > message in /var/log/messages - is that simply because of timing or is
                        > there a log entry somewhere that I missed?
                        > 3) Do you have other suggestions on how I can troubleshoot the
                        > problem?
                        >
                        > thanks!
                        > /maurice
                        >


                        <some omitted>
                        > > PGP:0x154144DA
                      • Robert Hammond
                        In message , mpmarks2000 writes ... Thanks for adding it to the Wiki page, I have been meaning to do this for some
                        Message 11 of 21 , Jun 16, 2005
                        View Source
                        • 0 Attachment
                          In message <d8rvj3+lusp@...>, mpmarks2000 <marks@...>
                          writes
                          >An update. The script does work! My drive eventually went to sleep.
                          >Its not clear to me what Unslung is doing that keeps the drive active
                          >but eventually it went to sleep. Possibly the failure of a backup job.
                          >
                          >So for my penance I added Robert's instructions to the Wiki as an
                          >update to the "how to set the timeout on One Touch drives". I'm still
                          >searching for a solution on how to put NON One Touch USB 2.0 drives to
                          >sleep from Linux.
                          >
                          >/maurice
                          >
                          Thanks for adding it to the Wiki page, I have been meaning to do this
                          for some time. Note that I have slightly changed and reformatted the
                          Wiki page mainly so that cut and paste will work better with the script
                          section.

                          I have my disk set to power down on the minimum time setting, 5 minutes
                          I think. Drive does still spin up always at midnight perhaps to check
                          out any cron jobs, also spins up at other times for unknown reasons but
                          does seem to sleep most of the time.

                          --
                          Robert Hammond
                          PGP:0x154144DA
                        • mpmarks2000
                          I m actually running 5.5 now, but I was trying to track down why sda1 was active sometimes and why free memory kept disappearing and I found a problem with the
                          Message 12 of 21 , Jul 12, 2005
                          View Source
                          • 0 Attachment
                            I'm actually running 5.5 now, but I was trying to track down why sda1
                            was active sometimes and why free memory kept disappearing and I found
                            a problem with the Samba config. Robert's rc.bootbin script puts /var
                            in a ramfs (in memory). This is supposed to be limited to 256K but
                            others have suggested that the ramfs memory limit is ignored, so more
                            memory could be actually used.

                            I noticed that the samba log file (in /var/log/samba/log.smbd) kept
                            growing as the disks were accessed because of an error in the
                            /etc/smb.conf. It contains the line 'username map = no' and tries to
                            open a file called 'no' every time the samba disks are accessed. So
                            the log keeps growing and growing because the file doesn't exist.
                            Commenting out that line in /etc/samba doesnt work because its
                            recreated on reboot. So I searched for all of the smb.conf files and
                            changed them all. Lo and behold on reboot the line was gone and no
                            more errors were logged. Free memory still seems to decrease (more
                            slowly), but at least one problem is solved. I hope that this fixed in
                            the next beta by removing the line or pointing usname map to the right
                            file.

                            /maurice

                            --- In nslu2-linux@yahoogroups.com, Robert Hammond <rob.hammond@n...>
                            wrote:
                            > In message <d8rvj3+lusp@e...>, mpmarks2000 <marks@s...>
                            > writes
                            > >An update. The script does work! My drive eventually went to sleep.
                            > >Its not clear to me what Unslung is doing that keeps the drive active
                            > >but eventually it went to sleep. Possibly the failure of a backup job.
                            > >
                            > >So for my penance I added Robert's instructions to the Wiki as an
                            > >update to the "how to set the timeout on One Touch drives". I'm still
                            > >searching for a solution on how to put NON One Touch USB 2.0 drives to
                            > >sleep from Linux.
                            > >
                            > >/maurice
                            > >
                            > Thanks for adding it to the Wiki page, I have been meaning to do this
                            > for some time. Note that I have slightly changed and reformatted the
                            > Wiki page mainly so that cut and paste will work better with the script
                            > section.
                            >
                            > I have my disk set to power down on the minimum time setting, 5
                            minutes
                            > I think. Drive does still spin up always at midnight perhaps to check
                            > out any cron jobs, also spins up at other times for unknown reasons
                            but
                            > does seem to sleep most of the time.
                            >
                            > --
                            > Robert Hammond
                            > PGP:0x154144DA
                          • Robert Hammond
                            In message , mpmarks2000 writes ... I also see this error, I already have a small rc.samba diversion script
                            Message 13 of 21 , Jul 13, 2005
                            View Source
                            • 0 Attachment
                              In message <db0m73+hmlp@...>, mpmarks2000 <marks@...>
                              writes
                              <snip>
                              >
                              >I noticed that the samba log file (in /var/log/samba/log.smbd) kept
                              >growing as the disks were accessed because of an error in the
                              >/etc/smb.conf. It contains the line 'username map = no' and tries to
                              >open a file called 'no' every time the samba disks are accessed. So
                              >the log keeps growing and growing because the file doesn't exist.
                              >Commenting out that line in /etc/samba doesnt work because its
                              >recreated on reboot. So I searched for all of the smb.conf files and
                              >changed them all. Lo and behold on reboot the line was gone and no
                              >more errors were logged. Free memory still seems to decrease (more
                              >slowly), but at least one problem is solved. I hope that this fixed in
                              >the next beta by removing the line or pointing usname map to the right
                              >file.
                              >
                              I also see this error, I already have a small rc.samba diversion script
                              so will attempt a small script section to remove this line (should be
                              quite easy) perhaps using the sed script command. If it works I will
                              add it to the FAQ Wiki section.


                              Note that I also have strange event entries in my
                              /var/log/samba/log.nmbd log file.

                              I am also running the /var folder in ram but must admit have not seen
                              any reduction in free memory but will start to monitor it from today.
                              I would have thought that the most likely culprit is a misbehaved
                              package.

                              Also not that I was under the impression (not too sure where from) that
                              the Linksys firmware is configured with a cron job that periodically
                              clears the log files from the /var folder.
                              --
                              Robert Hammond
                              PGP:0x154144DA
                            • Robert Hammond
                              In message , Robert Hammond writes ... I have a tested script section for this, it needs to be
                              Message 14 of 21 , Jul 13, 2005
                              View Source
                              • 0 Attachment
                                In message <AxmwliEz6X1CFwoF@...>, Robert Hammond
                                <rob.hammond@...> writes
                                >In message <db0m73+hmlp@...>, mpmarks2000 <marks@...>
                                >writes
                                ><snip>
                                >>
                                >>I noticed that the samba log file (in /var/log/samba/log.smbd) kept
                                >>growing as the disks were accessed because of an error in the
                                >>/etc/smb.conf. It contains the line 'username map = no' and tries to
                                >>open a file called 'no' every time the samba disks are accessed. So
                                >>the log keeps growing and growing because the file doesn't exist.
                                >>Commenting out that line in /etc/samba doesnt work because its
                                >>recreated on reboot. So I searched for all of the smb.conf files and
                                >>changed them all. Lo and behold on reboot the line was gone and no
                                >>more errors were logged. Free memory still seems to decrease (more
                                >>slowly), but at least one problem is solved. I hope that this fixed in
                                >>the next beta by removing the line or pointing usname map to the right
                                >>file.
                                >>
                                >I also see this error, I already have a small rc.samba diversion script
                                >so will attempt a small script section to remove this line (should be
                                >quite easy) perhaps using the sed script command. If it works I will
                                >add it to the FAQ Wiki section.
                                >
                                >
                                >Note that I also have strange event entries in my
                                >/var/log/samba/log.nmbd log file.
                                >
                                >I am also running the /var folder in ram but must admit have not seen
                                >any reduction in free memory but will start to monitor it from today.
                                >I would have thought that the most likely culprit is a misbehaved
                                >package.
                                >
                                >Also not that I was under the impression (not too sure where from) that
                                >the Linksys firmware is configured with a cron job that periodically
                                >clears the log files from the /var folder.

                                I have a tested script section for this, it needs to be combined into a
                                working rc.samba text file such as the one on the Wiki at :-

                                <http://www.nslu2-linux.org/wiki/HowTo/EnableWinsServer>

                                The simple script section pasted below will search for the text
                                username map = no
                                if present will delete that line.


                                sed -i -e '/username map \= no/d' /etc/samba/smb.conf


                                --
                                Robert Hammond
                                PGP:0x154144DA
                              • mpmarks2000
                                Robert, what I was noticing was that the buffer memory+free memory dropped from 10M after a clean reboot to 1.5MB 24 hrs later. The probable cause is that
                                Message 15 of 21 , Jul 13, 2005
                                View Source
                                • 0 Attachment
                                  Robert, what I was noticing was that the buffer memory+free memory
                                  dropped from 10M after a clean reboot to 1.5MB 24 hrs later. The
                                  probable cause is that there are 2 more smbds running, making 4 total!
                                  At about 4MB each that accounts for the memory loss. I wish I knew why
                                  they were started.

                                  --- In nslu2-linux@yahoogroups.com, Robert Hammond <rob.hammond@n...>
                                  wrote:
                                  > In message <AxmwliEz6X1CFwoF@n...>, Robert Hammond
                                  > <rob.hammond@n...> writes
                                  > >In message <db0m73+hmlp@e...>, mpmarks2000 <marks@s...>
                                  > >writes
                                  > ><snip>
                                  > >>
                                  > >>I noticed that the samba log file (in /var/log/samba/log.smbd) kept
                                  > >>growing as the disks were accessed because of an error in the
                                  > >>/etc/smb.conf. It contains the line 'username map = no' and tries to
                                  > >>open a file called 'no' every time the samba disks are accessed. So
                                  > >>the log keeps growing and growing because the file doesn't exist.
                                  > >>Commenting out that line in /etc/samba doesnt work because its
                                  > >>recreated on reboot. So I searched for all of the smb.conf files and
                                  > >>changed them all. Lo and behold on reboot the line was gone and no
                                  > >>more errors were logged. Free memory still seems to decrease (more
                                  > >>slowly), but at least one problem is solved. I hope that this fixed in
                                  > >>the next beta by removing the line or pointing usname map to the right
                                  > >>file.
                                  > >>
                                  > >I also see this error, I already have a small rc.samba diversion
                                  script
                                  > >so will attempt a small script section to remove this line (should be
                                  > >quite easy) perhaps using the sed script command. If it works I will
                                  > >add it to the FAQ Wiki section.
                                  > >
                                  > >
                                  > >Note that I also have strange event entries in my
                                  > >/var/log/samba/log.nmbd log file.
                                  > >
                                  > >I am also running the /var folder in ram but must admit have not seen
                                  > >any reduction in free memory but will start to monitor it from today.
                                  > >I would have thought that the most likely culprit is a misbehaved
                                  > >package.
                                  > >
                                  > >Also not that I was under the impression (not too sure where from) that
                                  > >the Linksys firmware is configured with a cron job that periodically
                                  > >clears the log files from the /var folder.
                                  >
                                  > I have a tested script section for this, it needs to be combined
                                  into a
                                  > working rc.samba text file such as the one on the Wiki at :-
                                  >
                                  > <http://www.nslu2-linux.org/wiki/HowTo/EnableWinsServer>
                                  >
                                  > The simple script section pasted below will search for the text
                                  > username map = no
                                  > if present will delete that line.
                                  >
                                  >
                                  > sed -i -e '/username map \= no/d' /etc/samba/smb.conf
                                  >
                                  >
                                  > --
                                  > Robert Hammond
                                  > PGP:0x154144DA
                                • Robert Hammond
                                  In message , mpmarks2000 writes ... I seem to remember a post some where inferring that if you disable the UPNP
                                  Message 16 of 21 , Jul 14, 2005
                                  View Source
                                  • 0 Attachment
                                    In message <db4c8s+g40o@...>, mpmarks2000 <marks@...>
                                    writes
                                    >Robert, what I was noticing was that the buffer memory+free memory
                                    >dropped from 10M after a clean reboot to 1.5MB 24 hrs later. The
                                    >probable cause is that there are 2 more smbds running, making 4 total!
                                    >At about 4MB each that accounts for the memory loss. I wish I knew why
                                    >they were started.
                                    >
                                    I seem to remember a post some where inferring that if you disable the
                                    UPNP support in the Web interface under Administration/System then it
                                    will remove two of the running smbds.
                                    --
                                    Robert Hammond
                                    PGP:0x154144DA
                                  Your message has been successfully submitted and would be delivered to recipients shortly.