Author |
Message |
iscraigh
|
Posted: Sat Jul 18, 2009 4:23 pm |
|
Joined: Mon Mar 21, 2005 1:43 pm
Posts: 388
Location:
Nanaimo BC
|
Sorry it took me so long to try this.
I get to the same place....
updated backend and run config_diskless_frontend.bash script I notice an error about the nfs module.
I forge ahead.
I end up stuck at can't find /dev/nfs
Try to create it cant find /bin/mknod
When I look in the image on the backend mknod is present.
Delete and try again
Still no go
any suggestions?
Craig
|
|
Top |
|
|
Greg Frost
|
Posted: Sat Jul 18, 2009 5:26 pm |
|
Joined: Mon May 10, 2004 8:08 pm
Posts: 1891
Location:
Adelaide, Australia
|
That was the error I was getting before I submitted a new kernel config that compiled my network card into the kernal (rather than a module). What network device does your frnotend use?
|
|
Top |
|
|
iscraigh
|
Posted: Sat Jul 18, 2009 5:41 pm |
|
Joined: Mon Mar 21, 2005 1:43 pm
Posts: 388
Location:
Nanaimo BC
|
Tried on 2 machines one has a tulip
Code: lsmod Module Size Used by ipv6 260212 12 usbhid 36896 0 hid 40704 1 usbhid snd_seq_oss 31872 0 snd_seq_midi_event 8192 1 snd_seq_oss snd_seq 49968 4 snd_seq_oss,snd_seq_midi_event snd_seq_device 8332 2 snd_seq_oss,snd_seq snd_pcm_oss 40320 0 snd_mixer_oss 16512 1 snd_pcm_oss nvidia 7231068 36 analog 12192 0 ppp_generic 25244 0 slhc 7680 1 ppp_generic gameport 12300 1 analog serio_raw 7172 0 pcspkr 4352 0 tulip 51616 0 sg 26804 0 snd_intel8x0 30876 0 snd_ac97_codec 102436 1 snd_intel8x0 ac97_bus 3584 1 snd_ac97_codec ohci_hcd 25360 0 forcedeth 56720 0 shpchp 34068 0 pci_hotplug 28064 1 shpchp snd_pcm 70020 3 snd_pcm_oss,snd_intel8x0,snd_ac97_codec snd_timer 21384 2 snd_seq,snd_pcm ehci_hcd 36876 0 snd 50852 9 snd_seq_oss,snd_seq,snd_seq_device,snd_pcm_oss,snd_mixer_oss,snd_intel8x0,snd_ac97_codec,snd_pcm,snd_timer soundcore 8160 1 snd snd_page_alloc 10120 2 snd_intel8x0,snd_pcm usbcore 136976 4 usbhid,ohci_hcd,ehci_hcd i2c_nforce2 8580 0 i2c_core 22804 2 nvidia,i2c_nforce2 nvidia_agp 8348 1 agpgart 31572 2 nvidia,nvidia_agp evdev 11296 5 thermal 17180 0 processor 41388 1 thermal fan 6276 0 button 7824 0 battery 12036 0 ac 6020 0 rtc_cmos 12332 0 rtc_core 17564 1 rtc_cmos rtc_lib 4480 1 rtc_core ext3 126984 2 jbd 46996 1 ext3 mbcache 8708 1 ext3 sd_mod 26904 4 sr_mod 16708 0 cdrom 35360 1 sr_mod pata_acpi 6016 0 pata_amd 12292 3 libata 158240 2 pata_acpi,pata_amd scsi_mod 102036 4 sg,sd_mod,sr_mod,libata
The other
uses an sis900 module
Is this any help
Craig
|
|
Top |
|
|
Greg Frost
|
Posted: Sat Jul 18, 2009 6:07 pm |
|
Joined: Mon May 10, 2004 8:08 pm
Posts: 1891
Location:
Adelaide, Australia
|
If either of them use a module, then that is your problem. For nfsroot, the network module must be built into the kernel.
Are you running the latest R6 (pacman -Syu)? I dont understand why it is showing tulip as a module because for the latest kernel, that is built in. My network device does not show up in lsmod: Code: [gfrost@devnew net]$ dmesg | grep -i ether r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded [gfrost@devnew net]$ lsmod | grep r81 [gfrost@devnew net]$ What does this give you: Code: sudo updatedb locate tulip
|
|
Top |
|
|
iscraigh
|
Posted: Sat Jul 18, 2009 9:31 pm |
|
Joined: Mon Mar 21, 2005 1:43 pm
Posts: 388
Location:
Nanaimo BC
|
Okay it appears it is the nforce driver
forcedeth: Reverse Engineered nForce ethernet driver. Version 0.61 Code: [root@trev ~]# dmesg | grep -i ether forcedeth: Reverse Engineered nForce ethernet driver. Version 0.61. [root@trev ~]# locate tulip /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip/de2104x.ko /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip/de4x5.ko /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip/dmfe.ko /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip/tulip.ko /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip/uli526x.ko /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip/winbond-840.ko /lib/modules/2.6.28-LinHES/kernel/drivers/net/tulip/xircom_cb.ko /sys/bus/pci/drivers/tulip /sys/bus/pci/drivers/tulip/0000:01:07.0 /sys/bus/pci/drivers/tulip/bind /sys/bus/pci/drivers/tulip/module /sys/bus/pci/drivers/tulip/new_id /sys/bus/pci/drivers/tulip/uevent /sys/bus/pci/drivers/tulip/unbind /sys/module/tulip /sys/module/tulip/drivers /sys/module/tulip/holders /sys/module/tulip/initstate /sys/module/tulip/notes /sys/module/tulip/refcnt /sys/module/tulip/sections /sys/module/tulip/srcversion /sys/module/tulip/version /sys/module/tulip/drivers/pci:tulip /sys/module/tulip/notes/.note.gnu.build-id /sys/module/tulip/sections/.altinstr_replacement /sys/module/tulip/sections/.altinstructions /sys/module/tulip/sections/.bss /sys/module/tulip/sections/.data /sys/module/tulip/sections/.devexit.text /sys/module/tulip/sections/.devinit.data /sys/module/tulip/sections/.devinit.text /sys/module/tulip/sections/.exit.text /sys/module/tulip/sections/.gnu.linkonce.this_module /sys/module/tulip/sections/.init.text /sys/module/tulip/sections/.note.gnu.build-id /sys/module/tulip/sections/.parainstructions /sys/module/tulip/sections/.rodata /sys/module/tulip/sections/.rodata.str1.1 /sys/module/tulip/sections/.rodata.str1.4 /sys/module/tulip/sections/.smp_locks /sys/module/tulip/sections/.strtab /sys/module/tulip/sections/.symtab /sys/module/tulip/sections/.text /sys/module/tulip/sections/__bug_table /sys/module/tulip/sections/__param /usr/src/linux-2.6.28-LinHES/drivers/net/tulip /usr/src/linux-2.6.28-LinHES/drivers/net/tulip/Kconfig /usr/src/linux-2.6.28-LinHES/include/config/tulip.h /usr/src/linux-2.6.28-LinHES/include/config/net/tulip.h
|
|
Top |
|
|
Greg Frost
|
Posted: Sat Jul 18, 2009 10:08 pm |
|
Joined: Mon May 10, 2004 8:08 pm
Posts: 1891
Location:
Adelaide, Australia
|
Have you tried removing /ftpboot to allow the script to re-create it? If you created it when the old pre-upgraded kernel was installed, it will have a kernel image from the old kernel (without the ethernet modules built in).
|
|
Top |
|
|
iscraigh
|
Posted: Sun Jul 19, 2009 10:43 am |
|
Joined: Mon Mar 21, 2005 1:43 pm
Posts: 388
Location:
Nanaimo BC
|
Started from scratch after deleting nfsroot and tftp, still no go the only thing I see while it is building is
error nfs module not found
Code: IPRANGE=192.168.0.40 192.168.0.50 tftpd is already present ok: run: tftpd: (pid 2125) 0s ALLNAMESERVERS= 127.0.0.1, 192.168.0.254 dhcpd is already present ok: run: dhcpd: (pid 2167) 0s ok: down: mythbackend: 1s, normally up Already commented out skip-networking. Setting backend IP in mythtv's mysql settings ok: run: mythbackend: (pid 2222) 0s Already added line for /myth in /etc/exports Already added line for /data/var/cache/pacman in /etc/exports Creating /tftpboot directories `/usr/lib/syslinux/pxelinux.0' -> `/tftpboot/pxelinux.0' `/boot/vmlinuz26' -> `/tftpboot/vmlinuz26' Building kernel miniroot :: Begin build ERROR: module 'nfs' not found :: Parsing hook [base] :: Parsing hook [udev] :: Parsing hook [autodetect] :: Parsing hook [pata] :: Parsing hook [scsi] :: Parsing hook [sata] :: Parsing hook [filesystems] :: Parsing hook [fbsplash] :: Parsing hook [net] :: Generating module dependencies
Could that be the issue?
Thanks
Craig
|
|
Top |
|
|
Greg Frost
|
Posted: Mon Jul 20, 2009 7:36 pm |
|
Joined: Mon May 10, 2004 8:08 pm
Posts: 1891
Location:
Adelaide, Australia
|
That nfs error is a result of the change to build nfs into the kernel. It is not of concern. When I get the chance I'll test again on my system from a fresh install and see if I get any clues.
|
|
Top |
|
|
Greg Frost
|
Posted: Wed Jul 22, 2009 5:06 am |
|
Joined: Mon May 10, 2004 8:08 pm
Posts: 1891
Location:
Adelaide, Australia
|
Just how far is it getting?
Mine gets to this point and seems to stop: Code: Root device '/dev/nfs' doesn't exist, attempting to create it /bin/mknod "/dev/nfs" b 0 255 ::Initramfs Completed - control passing to kinit IP-Config: eth0 hardware address 00:06:nn:nn:nn:nn mtu 1500 DHCP RARP
If I ping the frontend address from the backend when it is stalled like this, it springs to life, completes the boot sequence booting into the frontend.
Are you seeing the same behaviour?
If I just leave it it also appears to boot into the frontend if I wait long enough.
|
|
Top |
|
|
iscraigh
|
Posted: Sat Jul 25, 2009 1:39 am |
|
Joined: Mon Mar 21, 2005 1:43 pm
Posts: 388
Location:
Nanaimo BC
|
Mine ends just after that with a line that says
unable to find create root device /dev/nfs
cannot fine /bin/mknod
So yours makes the node but mine doesn't. If I echo the bin from the root shell mknod is not there, If I look on the nfsroot folder in the backend it is present.
Then the usual messages about dropping to a root shell
I left it but it never started the frontend.
Craig
|
|
Top |
|
|
iscraigh
|
Posted: Sat Aug 01, 2009 9:59 am |
|
Joined: Mon Mar 21, 2005 1:43 pm
Posts: 388
Location:
Nanaimo BC
|
|
Top |
|
|
djb61230
|
Posted: Sat Aug 01, 2009 10:10 am |
|
Joined: Fri Jun 23, 2006 10:57 am
Posts: 106
|
I'm playing with R6 and I find that one of my clients will get stuck at that position if I do a "reboot". But it will NOT be stuck if I shutdown and power off the client, then power up.
Probably not helpful but something that I noticed.
|
|
Top |
|
|
lordsnooty
|
Posted: Wed Aug 12, 2009 5:57 pm |
|
Joined: Tue Feb 21, 2006 2:25 pm
Posts: 39
|
Can anyone help me with getting this to work with R5.5 and the Via Rhine chip found on their mini-boards? I am getting kernel panic "no init found" when attempting a boot, but looking at the config I see that this driver is already compiled into the kernel. I try init=/bin/bash, that cannot be executed. Any ideas where I should take it next? (I can't immediately install to hard disk, not enough power cables!)
Cheers
|
|
Top |
|
|
iscraigh
|
Posted: Sat Nov 07, 2009 2:36 pm |
|
Joined: Mon Mar 21, 2005 1:43 pm
Posts: 388
Location:
Nanaimo BC
|
Thought I would play with this again.
I run the config_diskless_frontend.bash script
It makes all the files etc, tells me to boot the frontend.
I boot the front end it pulls an ip from the range specified when I ran the script.
I rerun the script, get the error unable to automatically determine the mac address etc.
I look in the daemon.log
Code: Nov 7 12:30:43 hdbe dhcpd: DHCPDISCOVER from 00:22:68:69:b9:88 via eth0 Nov 7 12:30:43 hdbe dhcpd: DHCPOFFER on 192.168.0.45 to 00:22:68:69:b9:88 via eth0 Nov 7 12:30:45 hdbe dhcpd: DHCPREQUEST for 192.168.0.45 (192.168.0.22) from 00:22:68:69:b9:88 via eth0 Nov 7 12:30:45 hdbe dhcpd: DHCPACK on 192.168.0.45 to 00:22:68:69:b9:88 via eth0 Nov 7 12:30:50 hdbe dhcpd: DHCPDISCOVER from 00:22:68:69:b9:88 via eth0 Nov 7 12:30:50 hdbe dhcpd: DHCPOFFER on 192.168.0.45 to 00:22:68:69:b9:88 via eth0 Nov 7 12:30:50 hdbe dhcpd: DHCPREQUEST for 192.168.0.45 (192.168.0.22) from 00:22:68:69:b9:88 via eth0 Nov 7 12:30:50 hdbe dhcpd: DHCPACK on 192.168.0.45 to 00:22:68:69:b9:88 via eth0
It appears to have record of the mac address ip combo.
Any idea what I can do to get this running?
Thanks
Craig
|
|
Top |
|
|
Greg Frost
|
Posted: Sat Nov 07, 2009 2:45 pm |
|
Joined: Mon May 10, 2004 8:08 pm
Posts: 1891
Location:
Adelaide, Australia
|
The bit about running the script again is only necessary if you want to setup 2 frontends. It allows the system to determine which nfsroot to serve to which mac address. If you only have one frontend, everything should work without the need to run it again.
I see you just posted this. Join me in #linhes for a more detailed discussion if you are available.
Out of interest, how long has your backend been up? The script works by checking /var/log/daemon.log and I just booted one of my frontends and found that the mount logs went to /var/log/daemon.log.1
So it looks like the logs have been rotated, but its still writing to the old log. If this is the case with your system, rebooting the backend may solve your problem (because the logs will go to /var/log/daemon.log again).
|
|
Top |
|
|