Re: [suse-autoinstall] Autoinstall with PXE
I always get the "Insert modules Disk 3" when running with autoyast=nfs://172.16.100.1/opt/suse set on kernel cmdline. Is this because yast2 cannot find the modules
it
need? Or does it always ask this question? How do I get past it? (our systems only has a serial console, no floppy, no cdrom).
What do you have in autoyast=nfs://172.16.100.1/opt/suse ? If you have
copied the CDs to this directory and you wish to install over NFS, you
should use install=nfs://172.16.100.1/opt/suse
This is my commandline. init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE It looks like linuxrc cannot find any modules on the ramdisk. Adding root=/dev/ram0 does not help either. Last out messages from kernel : : RAMDISK: Compressed image found at block 0 Uncompressing.........................................done. Freeing initrd memory: 65536k freed VFS: Mounted root (minix filesystem). Freeing unused kernel memory: 120k freed Moving into tmpfs... done. ramdisk /dev/ram0 freed
SuSE installation program v1.3 (c) 1996-2002 SuSE Linux AG <<< Starting hardware detection... Activating usb devices... done Searching for infofile......
/Jörgen Karlsson
* Jörgen Karlsson <jorgen.karlsson@uab.ericsson.se> [May 23. 2002 14:13]:
I always get the "Insert modules Disk 3" when running with autoyast=nfs://172.16.100.1/opt/suse set on kernel cmdline. Is this because yast2 cannot find the modules
it
need? Or does it always ask this question? How do I get past it? (our systems only has a serial console, no floppy, no cdrom).
What do you have in autoyast=nfs://172.16.100.1/opt/suse ? If you have
copied the CDs to this directory and you wish to install over NFS, you
should use install=nfs://172.16.100.1/opt/suse
This is my commandline.
init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE
It looks like linuxrc cannot find any modules on the ramdisk.
Adding root=/dev/ram0 does not help either.
Last out messages from kernel
: : RAMDISK: Compressed image found at block 0 Uncompressing.........................................done. Freeing initrd memory: 65536k freed VFS: Mounted root (minix filesystem). Freeing unused kernel memory: 120k freed Moving into tmpfs... done. ramdisk /dev/ram0 freed
SuSE installation program v1.3 (c) 1996-2002 SuSE Linux AG <<< Starting hardware detection... Activating usb devices... done Searching for infofile......
Could you please try forcing loading the modules from the command line using the 'insmod' keyword, i.e.: autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=<module name> If you are using the initrd from suse/images/boot/initrd then all modules should be there. Have you modifed the initrd? Regards, Anas
/Jörgen Karlsson
-- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
-- Anas Nashif <nashif@suse.com>, SuSE Linux AG Montreal (Laval), Canada
Could you please try forcing loading the modules from the command line using the 'insmod' keyword, i.e.: autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=<module name>
I will try if insmod=eepro100 (or insmod=eepro100.o ??) solves the problem. All modules I need seems to be in /modules. As I said before I get a message box telling me to insert modules disk3. If you press the no button you revert back to manual installation. In manual mode there is a menu selection where you can load modules manually. I tried to load the eepro100 driver and the sym53c8xx drivers that way but it failed. My theory is that after makeing the new root and chrooting linuxrc fails to find the modules on the new root. But I don't know..
If you are using the initrd from suse/images/boot/initrd then all modules should be there. Have you modifed the initrd?
No. But I modified it once and added sash (statically linked) to the initrd image and replaced linuxrc with sash: mv linuxrc linuxrc.orig ln -s bin/sash linuxrc Then I booted, got a sash prompt and verified that the initrd was correctly mounted and that all the files where there. Then I tried to run the original linuxrc file. cd / ./linuxrc.orig Linuxrc created a new root, added some new files and chrooted to that it. Then it returned to the sash prompt. OK so I run linuxrc.orig again. Then the autoinstallation started! I could see from the ethereal snoop that the install image on the installation server was mounted. Then it crashed when partitioning the disks. I didn't look closely on what the problem was but the autoinstallation started. Could be errors in my xml file. regards Jörgen
Hi, I have found out what caused my problems. Our PXE bootstrap loader changed the root_dev field in the bzimage to /dev/ram0 (that was what we needed before). I modified the PXE bootstrap to leave that field unchanged (rdev on images/boot/linux gives /dev/sda1). Then everything started to work as expected:) But autoinstallation only works on eth1 !! 1. Manual installation with PXE (eth0) -------------------------------------- kernel cmdline: init=/linuxrc rw insmod=eepro100 Works !! 2. Autoinstallation with PXE (eth0) - 1 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100 Installation crashes everytime. When suse/images/root is downloaded only 10Mb of 33Mb is downloaded everytime. Tracing with ethereal showed that after 10Mb download the network is dead. ALT-F5: (Noticed that several shell commands did not work, e.g. ls, less... dmesg showed following output: : nfs: RPC call returned 101 RPC: sendmsg returned 101 nfs: RPC call returned 101 RPC: sendmsg returned 101 : : Error -3 while decompressing! c2d05a68(-33528512)-> ffe90000(4096 Error -3 while decompressing! c2xxxxxx ........etc : ifconfig -a shows that the ip-address on eth0 had disappeared. Rx bytes was 11095548 and TX bytes 401112. (in line with what the ethereal trace showed). NOTE: linuxrc does not check the result codes from the download before loop back mounting the image file. 3. Autoinstallation with PXE (eth1) - 1 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100 netdevice=eth1 Moved the cable from eth0 to eth1 on client. Did not work as linuxrc used eth0 everytime. 4. Autoinstallation with PXE (eth1) - 2 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100 netdevice=eth1 Downloaded new initrd/linux from: ftp://ftp.suse.com/pub/people/snwint/8.0/boot/ (fix for eth1) Works!!!!!!!!!! Successfull autoinstallation!!!!!!!! 5. Autoinstallation with PXE (eth0) - 2 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100 I switched back using eth0. Maybe the new initrd did it? No, does not work, same result as in 2. I tried several times. The only way to get a successful install is 4 above!!!! /Jörgen Karlsson
/Jörgen Karlsson
-- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
-- Anas Nashif <nashif@suse.com>, SuSE Linux AG Montreal (Laval), Canada
* Jörgen Karlsson <publius@chello.se> [May 27. 2002 23:41]:
Hi,
1. Manual installation with PXE (eth0) -------------------------------------- kernel cmdline: init=/linuxrc rw insmod=eepro100
Works !!
Where does linuxrc retrieves the root image from? Did you specify that later in the process?
2. Autoinstallation with PXE (eth0) - 1 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100
Installation crashes everytime.
When suse/images/root is downloaded only 10Mb of 33Mb is downloaded everytime. Tracing with ethereal showed that after 10Mb download the network is dead.
ALT-F5:
(Noticed that several shell commands did not work, e.g. ls, less...
dmesg showed following output: : nfs: RPC call returned 101 RPC: sendmsg returned 101 nfs: RPC call returned 101 RPC: sendmsg returned 101 : : Error -3 while decompressing! c2d05a68(-33528512)-> ffe90000(4096 Error -3 while decompressing! c2xxxxxx ........etc :
ifconfig -a shows that the ip-address on eth0 had disappeared. Rx bytes was 11095548 and TX bytes 401112. (in line with what the ethereal trace showed).
NOTE: linuxrc does not check the result codes from the download before loop back mounting the image file.
Investigating...
3. Autoinstallation with PXE (eth1) - 1 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100 netdevice=eth1
Moved the cable from eth0 to eth1 on client.
Did not work as linuxrc used eth0 everytime.
Ok, this is fixed in the initrd in 4.
4. Autoinstallation with PXE (eth1) - 2 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100 netdevice=eth1
Downloaded new initrd/linux from: ftp://ftp.suse.com/pub/people/snwint/8.0/boot/ (fix for eth1)
Works!!!!!!!!!! Successfull autoinstallation!!!!!!!!
5. Autoinstallation with PXE (eth0) - 2 --------------------------------------- kernel cmdline: init=/linuxrc rw autoyast=nfs://172.16.100.1/opt/SuSE/main.xml install=nfs://172.16.100.1/opt/SuSE insmod=eepro100
I switched back using eth0. Maybe the new initrd did it?
No, does not work, same result as in 2.
Well, I would switch NICs (eth0<->eth1) and try again to make sure NIC on eth0 is OK, although you had a successful installation in 1. Regards, Anas
I tried several times. The only way to get a successful install is 4 above!!!!
/Jörgen Karlsson
-- Anas Nashif <nashif@suse.com>, SuSE Linux AG Montreal (Laval), Canada
Anas Nashif wrote:
Where does linuxrc retrieves the root image from? Did you specify that later in the process?
I never specified the root image.
Well, I would switch NICs (eth0<->eth1) and try again to make sure NIC on eth0 is OK, although you had a successful installation in 1.
The NICs can't be swapped as they are mounted on the mother board. But I get the same result on two different machines. /Jörgen Karlsson
Regards, Anas
I tried several times. The only way to get a successful install is 4 above!!!!
/Jörgen Karlsson
-- Anas Nashif <nashif@suse.com>, SuSE Linux AG Montreal (Laval), Canada
-- To unsubscribe, e-mail: suse-autoinstall-unsubscribe@suse.com For additional commands, e-mail: suse-autoinstall-help@suse.com
participants (3)
-
Anas Nashif -
Jörgen Karlsson -
Jörgen Karlsson