RE: [External] Re: GRUB EFI blues - Debian 9/FAI 5.3.6
Honeywell Internal > The FAI classes are defined in class/. By default FAI defines the class > GRUB_PC if the architecture is i386 or amd64. You should try to change this to > GRUB_EFI in class/60-misc. > > -- > regards Thomas Thank you, Thomas. I downloaded the source code for FAI 5.3.6 and came to a similar conclusion. Here's what I did in one of my class/ files: if test -d /sys/firmware/efi ; then echo GRUB_EFI fi I also had to reorganize my files under disk_config/* so I could support both BIOS and EFI based systems simultaneously. Bob
RE: [External] Re: GRUB EFI blues - Debian 9/FAI 5.3.6
Please accept my apology in advance, I cannot get this e-mail client to bottom post and be readable. Here's what I've found: 1. My thumb drive appears to be booting in EFI mode as I can see /sys/firmware/efi once the system is installing. In addition, dmesg.log shows: [0.00] efi: EFI v2.50 by American Megatrends [0.00] efi: ACPI=0x9e1b2000 ACPI 2.0=0x9e1b2000 SMBIOS=0xf05e0 SMBIOS 3.0=0xf0600 MPS=0xfca20 ESRT=0x9a7e0bd8 2. I suspect another issue was the fact I was missing a partition for /boot/efi with a "vfat" filesystem. I've fixed this as well. 3. I thought I might need the grub-efi package installed in the NFSROOT and I tried to do so but this caused conflicts with fai-make-nfsroot At this point, I think getting GRUB to install is the main issue. If I could get FAI to create GRUB_EFI that would be wonderful as well. From: linux-fai [mailto:linux-fai-boun...@uni-koeln.de] On Behalf Of Andreas Heinlein Sent: Wednesday, April 18, 2018 3:22 AM To: linux-fai@uni-koeln.de Subject: [External] Re: GRUB EFI blues - Debian 9/FAI 5.3.6 Am 18.04.2018 um 10:14 schrieb tt-...@kky.ttu.ee<mailto:tt-...@kky.ttu.ee>: I can second to that. I installed a SuperMicro X10SLM-F based server last month and did not find any option in the BIOS to PXE-boot FAI into UEFI mode. Ended up using disklabel:gpt-bios and GRUB_PC. I did not try to boot off an USB stick, so it is worth investigating if an option exists for booting that in UEFI mode. From my experiments I was left with the impression that it is not easy (or even possible) to “cross-install” UEFI-boot-capable disk if the system was booted into legacy (BIOS) mode. If someone has found a way to do it, I would also appreciate suggestions. Regards, Toomas From: linux-fai <linux-fai-boun...@uni-koeln.de><mailto:linux-fai-boun...@uni-koeln.de> On Behalf Of Andreas Heinlein Sent: Wednesday, April 18, 2018 9:56 AM To: linux-fai@uni-koeln.de<mailto:linux-fai@uni-koeln.de> Subject: Re: GRUB EFI blues - Debian 9/FAI 5.3.6 Am 18.04.2018 um 00:28 schrieb Bob Apodaca: I think the first issue is FAI is setting the GRUB_PC class instead of the GRUB_EFI class and I'm not sure why. I am pretty sure this depends on how the installation was started. That means you will have to boot your FAI installation using UEFI as well. This can be a bit tricky if you want to install from network - I also tried setting up PXE with UEFI some time ago and failed. Bye, Andreas I am pretty sure it is not possible to set up grub-efi correctly when booted in legacy mode. While it is possible to detect that we are actually running an EFI-capable machine (dmidecode or lshw can detect that), we cannot access the efi variables under /sys/efi since the firmware doesn't expose them to the host when running under CSM aka "Legacy mode". Booting from USB with UEFI is possible, in fact I have such a USB device here somewhere. But I need to remember what I did, it was not (yet) completed in FAI at that time. I remember I wanted to make some patches available, but never found the time. This is almost a year ago now. What you basically need is a small FAT partition preferrably of type 'ef' (EFI Boot Partition) on the USB drive, which contains a grub efi image as EFI/BOOT/BOOTX64.EFI. That image can be created with grub-mkimage and needs to include at least all modules for reading the "main" partition and the grub.cfg on it. That will be mostly ext filesystem and msdos partition table, I think. That image should also include an embedded config file with a one-liner like configfile (hd0,msdos1) if the main partition is the second on the USB drive. I will try to find this again and post it here. Bye, Andreas
RE: [External] Re: GRUB EFI blues - Debian 9/FAI 5.3.6
Honeywell Internal From: linux-fai [mailto:linux-fai-boun...@uni-koeln.de] On Behalf Of Andreas Heinlein Sent: Wednesday, April 18, 2018 3:22 AM To: linux-fai@uni-koeln.de Subject: [External] Re: GRUB EFI blues - Debian 9/FAI 5.3.6 Am 18.04.2018 um 10:14 schrieb tt-...@kky.ttu.ee<mailto:tt-...@kky.ttu.ee>: I can second to that. I installed a SuperMicro X10SLM-F based server last month and did not find any option in the BIOS to PXE-boot FAI into UEFI mode. Ended up using disklabel:gpt-bios and GRUB_PC. I did not try to boot off an USB stick, so it is worth investigating if an option exists for booting that in UEFI mode. From my experiments I was left with the impression that it is not easy (or even possible) to “cross-install” UEFI-boot-capable disk if the system was booted into legacy (BIOS) mode. If someone has found a way to do it, I would also appreciate suggestions. Regards, Toomas From: linux-fai <linux-fai-boun...@uni-koeln.de><mailto:linux-fai-boun...@uni-koeln.de> On Behalf Of Andreas Heinlein Sent: Wednesday, April 18, 2018 9:56 AM To: linux-fai@uni-koeln.de<mailto:linux-fai@uni-koeln.de> Subject: Re: GRUB EFI blues - Debian 9/FAI 5.3.6 Am 18.04.2018 um 00:28 schrieb Bob Apodaca: I think the first issue is FAI is setting the GRUB_PC class instead of the GRUB_EFI class and I'm not sure why. I am pretty sure this depends on how the installation was started. That means you will have to boot your FAI installation using UEFI as well. This can be a bit tricky if you want to install from network - I also tried setting up PXE with UEFI some time ago and failed. Bye, Andreas I am pretty sure it is not possible to set up grub-efi correctly when booted in legacy mode. While it is possible to detect that we are actually running an EFI-capable machine (dmidecode or lshw can detect that), we cannot access the efi variables under /sys/efi since the firmware doesn't expose them to the host when running under CSM aka "Legacy mode". Booting from USB with UEFI is possible, in fact I have such a USB device here somewhere. But I need to remember what I did, it was not (yet) completed in FAI at that time. I remember I wanted to make some patches available, but never found the time. This is almost a year ago now. What you basically need is a small FAT partition preferrably of type 'ef' (EFI Boot Partition) on the USB drive, which contains a grub efi image as EFI/BOOT/BOOTX64.EFI. That image can be created with grub-mkimage and needs to include at least all modules for reading the "main" partition and the grub.cfg on it. That will be mostly ext filesystem and msdos partition table, I think. That image should also include an embedded config file with a one-liner like configfile (hd0,msdos1) if the main partition is the second on the USB drive. I will try to find this again and post it here. Bye, Andreas
GRUB EFI blues - Debian 9/FAI 5.3.6
I have a Supermicro X11SAE motherboard with an M.2 SATA drive that requires UEFI to boot. I've created a bootable USB drive, I've changed the BIOS settings to use UEFI and I can boot and appear to install the system. However, the system will not boot after the install is completed and the system does not recognize the drive. I think the first issue is FAI is setting the GRUB_PC class instead of the GRUB_EFI class and I'm not sure why. I've also attempted to run GRUB_EFI/10-setup manually after the installation is complete, this does not allow the system to boot either. Help.
debconf and exim4
Honeywell Internal I'm trying to pre-seed answers for exim4-config by creating a file in debconf/FAIBASE. For the most part, it appears to work. Here's the line that's giving me issues: $ grep other debconf/FAIBASE exim4-configexim4/dc_other_hostnamesstring However, when I'm in the chrooted install environment, I'm getting this: root@phxtst09:/# debconf-get-selections | grep exim4-config | grep other exim4-configexim4/dc_other_hostnamesstring subdomain.example.com So, it's not blank. Any ideas?
RE: cryptsetup with FAI 4.3.1 (Debian Jessie)
From: linux-fai [mailto:linux-fai-boun...@uni-koeln.de] On Behalf Of Bob Apodaca Sent: Tuesday, May 16, 2017 8:48 AM To: linux-fai@uni-koeln.de Subject: cryptsetup with FAI 4.3.1 (Debian Jessie) I'm trying to make an encrypted partition and getting errors. Here are my files: # disk_config disk1 disklabel:gpt bootable:1 fstabkey:uuid primary - 1M - - primary / 8192 ext4 rw,errors=remount-ro createopts="-L root" primary swap 1024 swap sw createopts="-L swap" primary /home 5120 ext4 rw,errors=remount-ro createopts="-m 1 -L home" primary /var/log 2048 ext4 rw,errors=remount-ro createopts="-L var_log" primary /tmp 4096 ext4 rw,errors=remount-ro createopts="-L tmp" primary - 2048-- - disk_config cryptsetup luks:"test" /var/data disk1.7 ext3 rw,errors=remount-ro createopts="-L data" and here is some output from fai.log: Executing: wipefs -a /dev/sda7 ... Executing: parted -s /dev/sda mkpart primary "" 21476933632B 80026344959B ... Executing: head -c 2048 /dev/urandom | od | tee /tmp/fai/crypt_dev_sda7 Executing: yes YES | cryptsetup luksFormat /dev/sda7 /tmp/fai/crypt_dev_sda7 -c aes-cbc-essiv:sha256 -s 256 Executing: cryptsetup luksOpen /dev/sda7 crypt_dev_sda7 --key-file /tmp/fai/crypt_dev_sda7 Executing: yes 'test' | cryptsetup luksAddKey --key-file /tmp/fai/crypt_dev_sda7 /dev/sda7 Executing: yes 'test' | cryptsetup luksRemoveKey /dev/sda7 /tmp/fai/crypt_dev_sda7 Executing: mkfs.ext3 -L data /dev/mapper/crypt_dev_sda7 /dev/sda5 UUID=073bb05a-c5c0-454c-8ba7-034b4d927da7 /dev/sda5 LABEL=var_log /dev/sda4 UUID=342b98dc-bc5f-40fd-bb7d-001b38c057ee /dev/sda4 LABEL=home /dev/sda3 UUID=b055a193-3dfe-4c4a-b569-84534ddd39b0 /dev/sda3 LABEL=swap /dev/sda6 UUID=76199b90-db2d-4b2a-b43c-f721639379d0 /dev/sda6 LABEL=tmp /dev/sda2 UUID=8270d805-b35d-46fb-bb8f-d602aec8bd26 /dev/sda2 LABEL=root Calling task_mountdisks Enable swap device /dev/sda3 Mounting UUID=8270d805-b35d-46fb-bb8f-d602aec8bd26 to /target/ mount: unknown filesystem type 'ext4' Error in task mountdisks. Code: 885 Traceback: task_error task_mountdisks task task_install task task_action task main Source hook: savelog.LAST.source ERRORS found in log files. See /tmp/fai/error.log -- Two apologies: 1. My mail client stinks 2. I had rebuilt the NFSROOT and forgot to update the kernel/initrd on my USB stick causing this error.
RE: /tmp read-only with Debian Stretch and FAI 5.3.4
> For a 4.x kernel you need rootovl instead of aufs as kernel parameter. boot=live is now obsolete. That was the issue. Thanks.
/tmp read-only with Debian Stretch and FAI 5.3.4
I know there are several threads already going related to this topic, sorry to start another. I'm using a GRUB bootable USB stick to start FAI. I've already attempted the suggestions here: https://lists.uni-koeln.de/pipermail/linux-fai/2016-August/011454.html Here's my /etc/exports on the FAI server (192.168.80.138): # Force first export NFSv4, everything else NFSv3 (for FAI) /srv/nfsv4 192.168.80.0/24(fsid=0,ro,no_subtree_check) /srv/fai *(async,ro) Here's the stanza from my GRUB entry: menuentry 'Debian 9 (Stretch) 64 bit FAI using DHCP' { insmod part_msdos echo'Loading Linux 4.9.0-1-amd64 ...' linux /stretch64/vmlinuz-4.9.0-1-amd64 ip=eth0:dhcp root=nfs:192.168.80.138:/srv/fai/nfsroot,nfsvers=3 aufs boot=live FAI_CONFIG_SRC=nfs://192.168.80.138/srv/fai/config,nfsvers=3 FAI_FLAGS=verbose,sshd,createvt FAI_ACTION=install BOOT_IMAGE=vmlinuz-4.9.0-1-amd64 HOSTNAME=faitest echo'Loading initial RAM disk ...' initrd /stretch64/initrd.img-4.9.0-1-amd64 } I am still having the read-only /tmp issue. Any other suggestions?
RE: Recommended specs for a FAI server
Sent: Wednesday, August 17, 2016 1:53 PM To: linux-fai@uni-koeln.de Subject: Recommended specs for a FAI server I browsed through the mailing list archive but could not find any infos on this: What would be the "recommended" specifications for a FAI server? I know these things can vary from one use case to another, but what about a sensible setup? Maybe more broadly, what are your FAI server specs (if you use a standalone machine/VM)? - I use a standalone machine, it's a Pentium 4 processor @ 2.8GHz with 512MB. We usually only install one client at a time, but I'm sure the system could handle more. I have another system that is an apt proxy to cache packages so that may take some of the load off the main FAI server. HTH, Bob PS: Sorry for the bad formatting, my e-mail client is not so good for mailing lists.
RE: $ROOTCMD commands fail
>From: linux-fai [mailto:linux-fai-boun...@uni-koeln.de] On Behalf Of Thomas Lange >Sent: Wednesday, May 18, 2016 1:45 PM >To: fully automatic installation for Linux>Subject: Re: $ROOTCMD commands fail > >> On Wed, 18 May 2016 15:38:02 -0500, Michael Kriss said: > >> $ROOTCMD cd /home/kriss/src_dir ; make install Please try > >$ROOTCMD bash -c "cd /home/kriss/src_dir ; make install" Be sure you are using "/bin/bash" as your interpreter and not "/bin/sh".
RE: setting up deb mirror
My apologies in advance, this e-mail client does not appear to permit bottom posting nor do I have the freedom to choose my e-mail client. I use a APT mirror and I put the settings in /etc/fai/apt/sources.list on the FAI server (this is for FAI 4.3.1). Once I do “fai-make-nfsroot” the file gets copied to the NFSROOT on the FAI server. Did you run “fai-make-nfsroot” after you updates /etc/fai/apt/sources.list? From: linux-fai [mailto:linux-fai-boun...@uni-koeln.de] On Behalf Of Ronald Steele Sent: Tuesday, March 15, 2016 1:02 PM To: fully automatic installation for LinuxSubject: setting up deb mirror What is the correct way to configure faiserver when the clients are on a private, un-routed network? I’ve run "fai-mirror /srv/mirror" on faiserver. However it’s not clear to me how this gets served to the client. I’ve tried setting /etc/fai/apt/sources.list to point to faiserver, and that failed. I’ve tried defining FAI_DEBMIRROR=nfs://faiserver/srv/mirror. This caused the install to stop saying it could not perform the nfs mount. However, I could perform the mount from the command line so I know the file system is exported properly. This if FAI 5.0.1 Thanks, Ron
Re: Hooks not working with 4.3.1?
Brilliant! It worked just as you suggested. Thanks! On 01/29/2016 12:21 AM, Carsten Aulbert wrote: Hi Bob On 01/28/2016 08:31 PM, Bob Apodaca wrote: I can run the script with an interactive shell and it works. Any ideas? If I recall correctly, FAI is using run-parts with the "LSB" option, i.e. your script name does not match the expected regexps: " If the --lsbsysinit option is given, then the names must not end in .dpkg-old or .dpkg-dist or .dpkg-new or .dpkg-tmp, and must belong to one or more of the following namespaces: the LANANA-assigned namespace (^[a-z0-9]+$); the LSB hierarchical and reserved namespaces (^_?([a-z0-9_.]+-)+[a-z0-9]+$); and the Debian cron script namespace (^[a-zA-Z0-9_-]+$). " Can you try renaming it to 10-script and check again? HTH Carsten
Hooks not working with 4.3.1?
I am using FAI with Debian 8 ("Jessie") and each time I re-run fai-make-nfsroot my hooks are not run. I have them set-up as follows: # ls -l /etc/fai/nfsroot-hooks/ total 4 -rwxr-xr-x 1 root root 197 Sep 15 09:31 scripts.sh The scripts.sh has the following: - #!/bin/sh install -g root -o root -m 0644 /etc/dhcp/dhclient.conf /srv/fai/nfsroot/etc/dhcp/dhclient.conf - I can run the script with an interactive shell and it works. Any ideas?
/etc/resolv.conf with FAI 4.3.1
During the installation, the /etc/resolv.conf file is not the file I expected. The FAI server does have the correct file, but the target has 2 files: /etc/resolv.conf /etc/resolv.conf-installserver The target is using DHCP, so I suspect the the resolv.conf is written when the IP address is acquired. Is there a way to over-ride this? The DHCP server does not have the correct search path (and I cannot update the DHCP server). Can I install and use the resolvconf package in the FAI root to fix this?
Re: putting a label on a disk
On 07/16/2015 09:22 AM, John G Heim wrote: All, Question related to those I asked earlier in the week... Can setup-storage put a label on a disk partition so that it can be mounted via the label rather than the uuid? I am guessing not beccause parted doesn't do that (as far as I can tell). I guess I can use e2label and xfs_admin in a script. But But I think I am going to have to reproduce some of the logic in setup-storage in order to make sure the label matches the partition. You can use createopts to pass an option to mkfs, like this: primary / 8192-ext4 rw,errors=remount-ro createopts=-L root Bob
Re: installing to a wiped disk
On 07/10/2015 07:46 AM, John G Heim wrote: My setup-storage config is below. I copied it from a working FAI setup. Our current FAI server is debian wheezy with FAI 4.3.1. I'm setting up an ubuntu vivid server with FAI 4.3.3. I mentioned in another message that I happen to be blind. But I did listen to the setup-storage documentation and I didn't hear anything about creating a partition table. It is somewhat tedious listening to documentation, as you might imagine, and I may have missed it. disk_config disk1 bootable:1 align-at:4k primary / 64G ext3 rw,relatime logical swap 4G-8G swap rw logical /home 1%- ext3 rw,relatime Hi John, I am using disklabel:gpt to create a GPT style partition table. You can also use disklabel:msdos to make the old style partition table. This belongs on your disk_config line before you specify partitions, file systems, etc. Some more information here: http://wiki.fai-project.org/wiki/Setup-storage Bob
Re: initd error: touch: not found
On 05/28/2015 02:26 PM, Bob Apodaca wrote: I am using Debian Jessie (AMD64) and am getting the following message when the system reboots: /init: 401: /init: touch: not found I did a few searches and some seemed to suggest rebuilding the initrd. I've tried this, but it doesn't seem to help. Anyone else seen this? If anyone else has this issue, it is not related to FAI but to the initramfs-tools package: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=783291
How do you disable a service with systemd?
I am using Debian Jessie and things are going well, with a couple exceptions. I am installing lightdm, but on some systems I want to disable it. I was trying to do: systemctl disable lightdm However, systemd knows this is a chroot environment and the command fails. I've also tried the following: $ROOT_CMD rm -f /etc/systemd/system/display-manager.service $ROOT_CMD /usr/sbin/update-rc.d lightdm defaults $ROOT_CMD /usr/sbin/update-rc.d lightdm disable Any suggestions?
initd error: touch: not found
I am using Debian Jessie (AMD64) and am getting the following message when the system reboots: /init: 401: /init: touch: not found I did a few searches and some seemed to suggest rebuilding the initrd. I've tried this, but it doesn't seem to help. Anyone else seen this?
Re: fai-installation with static ip-address
On 07/31/2014 03:50 AM, Werner Pommerer wrote: Hello, can anyone give an example, how to do a fai-installation with static ip-addresses for the clients? There are no dhcp-servers in our subnets, where the servers are. I have done this, though it was not the easiest task. My first step was to create a bootable CD-ROM that uses GRUB to load the FAI kernel and initrd (if you search the FAI archives I believe I posted the steps previously). Next, I created some custom entries where I could specify the IP address, similar to: menuentry 'Wheezy 32 bit FAI using static IP' { insmod part_msdos echo'Loading Linux 3.2.0-4-686-pae ...' linux /wheezy/vmlinuz-3.2.0-4-686-pae ip=192.168.1.222::192.168.1.1:255.255.255.0::eth0:off root=/dev/nfs nfsroot=192.169.1.2:/srv/fai/nfsroot:vers=3 aufs boot=live FAI_CONFIG_SRC=nfs://192.168.1.2/srv/fai/config FAI_FLAGS=verbose,sshd,createvt FAI_ACTION=install BOOT_IMAGE=vmlinuz-3.2.0-4-686-pae HOSTNAME= echo'Loading initial ramdisk ...' initrd /wheezy/initrd.img-3.2.0-4-686-pae } I've obscured the IP addresses, for your reference: 192.168.1.222 is the IP address of the target system 192.168.1.1 is the network gateway 192.168.1.2 is the IP of the FAI server and I have to fill-in the HOSTNAME= parameter so FAI will know what type of installation to perform.
Re: fai-cd -B gives an error
On 05/28/2014 12:09 AM, werner.pomme...@uni-hohenheim.de wrote: Thank you. I am very interested. Could you explain in more detail for me? Werner I have created a bootable CD-ROM using GRUB, but I do not use fai-cd. I only use the CD-ROM to boot the kernel and load the initrd, after those steps the system boots using NFS. I have also successfully used a similar process to boot from a USB thumb drive. If this set-up would help you I can attempt to explain how to create your own. Bob Configure your FAI installation as you would for a normal PXE or DHCP boot process. A some point you will run the command: make-nfs-root Now create a folder where you will begin to create your CD-ROM, I call mine cdroot Create a boot directory and grub subdirectory: mkdir -p cdroot/boot/grub Now create a grub.cfg in boot/grub, here's a starting point: set menu_color_normal=cyan/blue set menu_color_highlight=white/blue set timeout=30 menuentry 'Wheezy 32 bit FAI using DHCP' { insmod part_msdos echo'Loading Linux 3.2.0-4-686-pae ...' linux /wheezy/vmlinuz-3.2.0-4-686-pae ip=eth0:dhcp root=/dev/nfs nfsroot=192.168.1.166:/srv/fai/nfsroot:vers=3 aufs boot=live FAI_CONFIG_SRC=nfs://192.168.1.166/srv/fai/config FAI_FLAGS=verbose,sshd,createvt FAI_ACTION=install BOOT_IMAGE=vmlinuz-3.2.0-4-686-pae HOSTNAME=fai echo'Loading initial ramdisk ...' initrd /wheezy/initrd.img-3.2.0-4-686-pae } You will need to replace 192.168.1.166 with the IP address of your FAI server. Now create a wheezy subdirectory in your cdroot directory and copy the kernel and initrd files from $NFSROOT/boot directory to your cdroot boot directory. Now create your ISO image with this command: grub-mkrescue --modules=linux ext2 fshelp ls boot pc -o grub2-boot.iso cdroot/ For Debian, I had to install the xorriso package to properly build the ISO image. Burn your ISO image. When the system boots, use Control-x to edit the settings and change the hostname from fai to the proper hostname for the target system. To transfer to a USB thumb drive: 1. Do a search for GRUB bootable USB for instructions on creating a bootable GRUB USB stick. 2. Simply copy all your files from your cdroot directory structure to the USB stick. Bob
Re: fai-cd -B gives an error
On 05/26/2014 05:25 AM, Werner Pommerer wrote: Hello, I would like to create a boot only ISO image which does a network installation. fai-cd -B /tmp/test.iso NFSROOT /srv/fai/nfsroot mounted Found dracut inside nfsroot, but booting a FAI CD with dracut inside the nfsrroot does not yet work. Using grub2 preparing grub2 eltorito.img using existing core.img from NFSROOT no memtest86+.bin found, omit memtest boot option Writing FAI CD-ROM image to /tmp/test.iso. This may need some time. re-directing all messages to /dev/null ERROR: genisoimage failed. Does anyone have any ideas what might be going wrong? Regards Werner Pommerer I have created a bootable CD-ROM using GRUB, but I do not use fai-cd. I only use the CD-ROM to boot the kernel and load the initrd, after those steps the system boots using NFS. I have also successfully used a similar process to boot from a USB thumb drive. If this set-up would help you I can attempt to explain how to create your own. Bob
Nagios dependencies not installed
We are deploying Nagios and I attempted to add it to our base installation. The package is installed, but none of the dependencies are being installed. I did not see anything helpful in the installation log nor the error log. My package_config/FAIBASE file looks like this: PACKAGES aptitude DEBIAN bind9-host ... nagios-nrpe-server ... time update-inetd [EOF] Using Debian wheezy: # dpkg -l | grep fai ii fai-client 4.0.8~deb7u1 all Fully Automatic Installation client package ii fai-doc 4.0.8~deb7u1 all Documentation for FAI ii fai-server 4.0.8~deb7u1 all Fully Automatic Installation server package ii fai-setup-storage4.0.8~deb7u1 all automatically prepare storage devices Thank you for your consideration.
ifclass syntax within if statement
When installing AMD64 systems, I have created a small script: #!/bin/sh if ifclass AMD64 ; then # https://wiki.debian.org/Multiarch/Implementation $ROOTCMD dpkg --add-architecture i386 fi However, when the target system is installed it appears the command has not run: # dpkg --print-foreign-architectures I would appreciate any suggestions.
Re: Problems with wheezy installation on multi nic servers
On 08/26/2013 06:05 AM, Steffen Eichler wrote: Hi, I've installed a fai server with version 4.0.7. If I start a installation on a server with more than one network interface the following options in pxelinux.cfg APPEND initrd=initrd.img-3.2.0-4-amd64 ip=dhcp root=/dev/nfs nfsroot=XXX.XXX.XXX.XXX:/srv/fai/wheezy/nfsroot aufs rd_NO_LVM FAI_CONFIG_SRC=nfs://XXX.XXX.XXX.XXX/srv/fai/wheezy/config/ FAI_FLAGS=verbose,sshd,createvt root=/dev/nfs boot=live FAI_ACTION=install vga=785 DHCP configuration of interfaces fails and fai-client is not able to mount nfs. After that installation runs in kernel panic. I have tried this from grub, not pxelinux.cfg, but I suspect it will work: ip=eth0:dhcp Or whatever ethernet device you want it to use. Try that, see if it helps. Bob
Re: update to wheezy / r8169 card
Two things for you to check: 1. Perhaps a firmware issue, on Debian I found a package named firmware-realtek. 2. I have several systems with 2 network cards and there is an approximately 2 minute time-out when the second network card is not connected to a network where it can grab an address. If this is your situation, you can ignore the other network card with an option similar to: ip=eth0:dhcp in your boot configuration. Bob On 07/29/2013 05:44 AM, Michał Dwużnik wrote: Hi, I'm in the middle of a big leap to wheezy on my stations. The faisrv itself is upgraded to wheezy, I checked booting and installing squeeze from wheezy server, everything went fine for the test stations. Installs on dell optiplex 3010 machines were smooth (r8169 network card I was a bit worried about...). After that I decided to try on a new nfsroot, built for wheezy install on the stations Having built the nfsroot I find stations correctly booting from tftp with kernel 3.2 (pxelinux file is default fai-generated label fai-generated kernel vmlinuz-3.2.0-4-amd64 append initrd=initrd.img-3.2.0-4-amd64 ip=dhcp root=/dev/nfs nfsroot=/srv/fai/nfsroot aufs nfsopts=-onfsvers=3 ) Kernel boots ok, initrd seems OK The boot process apparently stops when trying to reconfigure the network adapter r8169... 'eth0 link is up' and this does not proceed to the later stages. Any hints on that? Regards Michal -- Michal Dwuznik
Use script in /usr/local/bin during install
Forgive me if this is a repeat, I sent this over 90 minutes ago and have not seen it yet. I need to run a custom script during an FAI install. My first thought was to install the script in NFSROOT/usr/local/bin. However, when I booted my install target /usr/local/bin was empty. I'd appreciate any suggestions.
Re: FAI with NFS v4 on Debian Wheezy
Thank you for the responses and I did not mean to imply this is an issue with FAI, just thought some others on the FAI list may have had some experience with it. I do have idmapd running on both the server and install client. It does not appear to be helping. I am going to try NFS v3. Bob On 05/07/2013 07:33 AM, Toomas Tamm wrote: Hello! I think I got it the other (incorrect) way in my original post. Sorry! In wheezy, FQDN minus hostname is the default (Domain is not set in /etc/idmapd.conf by default), while in squeeze, it is fixed to localdomain in the provided /etc/idmapd.conf. The bottom line is, one must ensure that the domain is identical on the server and client. Otherwise strange things happen with UIDs and GIDs. Toomas On Tue, 2013-05-07 at 13:06 +0200, Andreas B. Mundt wrote: Hi, On Tue, May 07, 2013 at 10:23:58AM +0300, Toomas Tamm wrote: [...] 2) ensure that all idmapd-s have the same value for Domain. Squeeze defaults to a portion of your fqdn, while wheezy defaults to localdomain. This is set in /etc/idmapd.conf and better make it identical across your entire network. If you do not specify the domain, it will be picked up from the fqdn. man idmapd.conf: Domain The local NFSv4 domain name. An NFSv4 domain is a namespace with a unique username-UID and groupname-GID mapping. (Default: Host's fully-qualified DNS domain name) I use the following /etc/idmapd.conf: [General] Verbosity = 0 Pipefs-Directory = /var/lib/nfs/rpc_pipefs # set your own domain here, if id differs from FQDN minus hostname # Domain = localdomain [Mapping] Nobody-User = nobody Nobody-Group = nogroup and it works fine here for NFSv4 mounted home directories (cf. debian-lan project). Best regards, Andi
FAI with NFS v4 on Debian Wheezy
I have set-up an FAI server and have a client I am using to refine my files and scripts. The issue I am having is file and group ownership: ls -l /etc/cups/cupsd.conf -rw-r--r-- 1 4294967294 4294967294 4670 Mar 14 13:51 cupsd.conf The nfsstat -m shows I am using NFS v4 on the client for the /var/lib/fai/config directory. I have nfs-common installed and /etc/idmapd.conf is configured on both the client and server. More confusing, when I do this command on the client: # ls -l /var/lib/fai/config/files/ total 24 drwxr-xr-x 22 root root 4096 May 6 09:59 etc drwxr-xr-x 3 4294967294 4294967294 4096 Nov 7 2006 home drwx-- 3 4294967294 4294967294 4096 Apr 23 12:15 lib drwxr-xr-x 5 root root 4096 Mar 30 2011 root drwxr-xr-x 3 4294967294 4294967294 4096 Feb 4 2011 usr drwxr-xr-x 3 4294967294 4294967294 4096 May 22 2007 var on the server all of these are owned by root:root. I need some suggestions. Thank you.
Re: Creating GPT BIOS partition with FAI 4.0.6
On 03/19/2013 11:16 AM, andrew bezella wrote: unless this has changed in fai4, you would want to use disklabel:gpt-bios instead of gpt. the necessary bios_grub partition should then be created automatically. That's what I need to know. Thank you.