S390x

From Alpine Linux
Revision as of 03:40, 13 June 2018 by Tmh1999 (talk | contribs)

The installer

The Alpine s390x installer is a netboot installer: it requires a kernel and an initramfs from the official repository.

Installations on KVM and z/VM are supported. Installation on LPAR is not yet supported at the moment.

Kernel parameters (and parmfile)

The Alpine s390x installer requires following kernel parameters to work:

ip=dhcp : use DHCP for network configuration.

ip=client-ip:server-ip:gw-ip:netmask:hostname:device:autoconf:dns1:dns2 : use static IP configuration, each field is separated by a colon :

  • client-ip ip address of the guest VM where we are going to run the installer
  • server-ip not used, leave blank or fill with none
  • gw-ip the gateway ip address
  • netmask the netmask
  • hostname not used, leave blank or fill with none
  • device the network interface of the guest VM, default is eth0 if left blank
  • autoconf not used, leave blank or fill with none or off
  • dns1 address of the DNS server
  • dns2 address of the 2nd DNS server

alpine_repo= : the location of the Alpine repository from which packages are downloaded.

For stable release, use http://dl-cdn.alpinelinux.org/alpine/v3.8/main
For rolling release, use http://dl-cdn.alpinelinux.org/alpine/edge/main

modloop= : the remote location of the image containing kernel's modules, required for LVM and raid setup.

The remote location of the modloop image is the same as the kernel and initramfs (see below)

ssh_key= : the remote location of your SSH public key which is used to allow SSH connection into the installer. It will be downloaded and copied into /root/.ssh/authorized_keys in the installer.

Note: Fetching public keys from HTTPS and FTPS are supported. HTTP and FTP (plaintext) are NOT supported.
Note: Alpine Linux security policies are against using passwords for root access via SSH as we consider using public-private key pairs is a much better approach.


z/VM only

dasd= : the addresses of the DASD devices, either ECKD or FBA DASDs. Each device is separated with a comma.

s390x_net= : the network interface type and its subchannels. At the moment, only QETH layer 2 is supported, thus the name qeth_l2 is used (see below).


Pre-installation

KVM

On your running s390x host, download the kernel and initramfs.

Create a virtual disk: 

$ qemu-img create alpine_disk.qcow2 5G

Start qemu: (modify ip= alpine_repo= ssh_key= for your needs) 

$ qemu-system-s390x -M s390-ccw-virtio \
       -m 2048 -smp 2 -nographic -enable-kvm \
       -hda alpine_disk.qcow2 \
       -net nic -net tap,ifname=tap0,script=no'
       -kernel vmlinuz-vanilla \
       -initrd initramfs-vanilla \
       -append "ip=192.168.1.2::192.168.1.1:255.255.255.0:none:eth0:none:8.8.8.8 alpine_repo=http://dl-cdn.alpinelinux.org/alpine/v3.8/main modloop=http://dl-cdn.alpinelinux.org/alpine/v3.8/releases/s390x/netboot-3.8.0_rc2/modloop-vanilla ssh_key=https://your-website.com/your-ssh-key.pub"
Note: If you have direct access to the qemu console, ssh_key might not be required

z/VM

To ease out the process of downloading the images, punch the readers, ipl, etc., ZNETBOOT is used.

Create the parm file

On your workstation/laptop, create a file named alpine.znetboot in your home directory with contents below (modify dasd=s390x_net= ip= alpine_repo= ssh_key= for your needs) 

ZNETBOOT_KERNEL=http://dl-cdn.alpinelinux.org/alpine/v3.8/releases/s390x/netboot-3.8.0_rc2/vmlinuz-vanilla
ZNETBOOT_INITRD=http://dl-cdn.alpinelinux.org/alpine/v3.8/releases/s390x/netboot-3.8.0_rc2/initramfs-vanilla
ZNETBOOT_PROGRESS=1M

alpine_repo=http://dl-cdn.alpinelinux.org/alpine/v3.8/main
modloop=http://dl-cdn.alpinelinux.org/alpine/v3.8/releases/s390x/netboot-3.8.0_rc2/modloop-vanilla

dasd=0.0.04c0,0.0.05d1
s390x_net=qeth_l2,0.0.0560,0.0.0561,0.0.0562
ip=192.168.1.2::192.168.1.1:255.255.255.0:none:eth0:none:8.8.8.8

ssh_key=https://your-website.com/your-ssh-key.pub


Upload to z/VM system via 3270 client

Figure 1.

On your workstation/laptop, download 2 files znetboot.exec, and curl.rexx to your home directory.

Open 3270 client and log in the z/VM system with your z/VM username and password.

Upload 3 files alpine.znetboot, znetboot.exec, curl.rexx to the z/VM environment using the 3270 client (this tutorial uses x3270). On the top left corner, click "File", then "File Transfer". (Figure 1.)


Do following steps : (Figure 2.)

Figure 2.
  • On "Local File Name" box, enter alpine.znetboot (the file in your laptop/workstation, at ~/alpine.znetboot)
  • On "Host File Name" box, enter alpine znetboot (the file will be in z/VM console)
Note: Beware the difference between the dot . and the space characters in the file names.
  • Choose Send to host
  • Choose Host is VM/CMS
  • Choose either Fixed or Variable for Record Format
Enter a number for LRECL and BLKSIZE, respectively
Note: Any line in alpine.znetboot that has more 80 chars (columns) will be splitted into more than 1 line when uploaded to z/VM (CMS) console via 3270. Count the number of characters/column of the longest line in your alpine.znetboot and fill it in LRECL. There is no restriction for BLKSIZE but 80 is preferred.
  • Click Transfer File box

Repeat the same steps with znetboot.exec and curl.rexx files.


(Optional) Check the configuration files

On 3270 client, enter following commands to check if the configuration files are correctly transferred:

xedit alpine znetboot

xedit znetboot exec

xedit curl rex

or filel and put xedit on CMD column to edit respective file.


Start ZNETBOOT

On 3270 client, type below command and wait till Figure 3.:

Figure 3.

znetboot alpine


Installation

If you install on KVM, steps in this part does not involve the interaction with the console starting qemu anymore. Everything is done in the terminal with SSH client.

If you install on z/VM, steps in this part does not involve the interaction with the 3270 client anymore. Everything is done in the terminal with SSH client.

Either installing in KVM or z/VM environments, from your workstation/laptop, you will be able to run: 

$ ssh root@192.168.1.2 (change ip address to what you specified earlier)

Remaining steps are similar to installing Alpine on other architectures (x86, arm, ppc, etc.), either on KVM (using virtio/SCSI disks) or on z/VM with FBA DASDs. Installing on ECKD DASDs requires an additional step, as described below.


Example

Below is the detailed walkthrough of installing Alpine on a single ECKD DASD using LVM and extend that LVM to the second ECKD DASD.


After SSH-ing into the Alpine installer, run: 

# setup-alpine

Select keyboard layout [none]:

press Enter for none

Enter system hostname (short form, e.g. 'foo') [localhost]:

enter your hostname
Available interfaces are: eth0.
Enter '?' for help on bridges, bonding and vlans.
Which one do you want to initialize? (or '?' or 'done') [eth0]:
type 'eth0' or press Enter

Ip address for eth0? (or 'dhcp', 'none', '?') [192.168.1.2]

enter ip address or 'dhcp'

Netmask? [255.255.255.0]

enter netmask

Gateway? (or 'none') [192.168.1.1]

enter gateway's ip address

Do you want to do any manual network configuration? [no]

enter 'no' or press Enter

DNS domain name? (e.g 'bar.com') []

enter domain name or press Enter for none

DNS nameserver(s)? [8.8.8.8 ]

enter DNS nameserver

Changing password for root

enter root password

Which timezone are you in? ('?' for list) [UTC]

enter timezone or '?' for list of timezones

HTTP/FTP proxy URL? (e.g. 'http://proxy:8080', or 'none') [none]

enter proxy or press Enter for none

Enter mirror number (1-27) or URL to add (or r/f/e/done) [f]:

enter a number or 'r' or 'f' or 'e' or 'done' as described

Which SSH server? ('openssh', 'dropbear' or 'none') [openssh]

enter SSH server or press Enter for openssh

Which NTP client to run? ('busybox', 'openntpd', 'chrony' or 'none') [chrony]

enter 'busybox' or press Enter for chrony

(next step is the additional step for ECKD DASDs on z/VM) 

Available ECKD DASD(s) are:
  0.0.04c0	(3390/0c 3990/e9 IBM)
  0.0.05d1	(3390/0c 3990/e9 IBM)
Which ECKD DASD(s) would you like to be formatted using dasdfmt? (enter '?' for help) [all]
enter 'all' or '0.0.04c0 0.0.05d1' (separated by a space) to format all/both DASDs
enter '0.0.04c0' or '0.0.05d1' to format respective DASD
enter '?' for help

WARNING: Erase ECKD DASD 0.0.04c0? [y/N]:

enter 'y' to format
Available disks are:
  dasda	(2.5 GB IBM 0.0.04c0)
Which disk(s) would you like to use? (or '?' for help or 'none') [dasda]
enter 'dasda' or press Enter
The following disk is selected:
  dasda	(2.5 GB IBM 0.0.04c0)
How would you like to use it? ('sys', 'data', 'lvm' or '?' for help) [?]
enter 'lvm'
The following disk is selected (with LVM):
  dasda	(2.5 GB IBM 0.0.04c0)
How would you like to use it? ('sys', 'data' or '?' for help) [?]
enter 'sys' to install Alpine on disk
WARNING: The following disk(s) will be erased:
  dasda	(2.5 GB IBM 0.0.04c0)
WARNING: Erase the above disk(s) and continue? [y/N]:
enter 'y'

Installation is complete. Please reboot.

the installation is finished

At this point, don't poweroff the installer right away. Go to "Copying SSH keys to new Alpine system" (below) to have SSH access to your new Alpine system.


Copying SSH keys to new Alpine system

By default, Alpine disables root login with a password via SSH. SSH keys are used instead.

After the installer's done running (Installation is complete. Please reboot.), there are 2 ways to copy your SSH key into the new Alpine system:

  • Option 1: mount the installed disk and copy the SSH keys while still at the installer's terminal
  • Option 2: poweroff the installer, start the new Alpine system and directly add the SSH keys
if you install on KVM, boot the new Alpine system on qemu, and copy the SSH keys
if you install on z/VM, use the 3270 client to ipl the new Alpine system and copy the SSH keys

Option 1

If you use 'lvm' + 'sys' installation (like in above example), do:

# mount /dev/vg0/lv_root /mnt
# cp -ar /root/.ssh /mnt/root

If you use 'sys' (without LVM) installation, do:

# mount /dev/dasda3 /mnt (change dasda to dasdb or dasdc, etc. for whichever DASD you chose)
# cp -ar /root/.ssh /mnt/root
Note: In 'sys' installation, 1st partition is boot, 2nd partition is swap, 3rd partition is root


Then run # poweroff.

Go to "Login to new Alpine system"

Option 2

Run # poweroff.

If you use KVM, start qemu with new Alpine system (removing -kernel, -initrd, -append options)

If you use z/VM, open the 3270 client, login with your z/VM username and password. You may need to run ipl cms. Then run ipl 04c0 (or whichever DASD device you chose as root disk in earlier steps).


Wait for new Alpine system go up, then login as root user while in the qemu console (on KVM) or 3270 client (on z/VM). Then run:

# mkdir /root/.ssh
# wget https://your-website.com/your-ssh-key.pub -O /root/.ssh/authorized_keys
# chmod 700 /root/.ssh
# chmod 600 /root/.ssh/authorized_keys

Go to "Login to new Alpine system"


Login to new Alpine system

On your workstation/laptop, use SSH client to login new Alpine system: # ssh root@192.168.1.2 (or whichever ip address you used)


Extending LVM volume

After logging in to your new Alpine system, run: 

# apk add -q util-linux e2fsprogs-extra

# lsblk
NAME            MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
dasda            94:0    0  2.3G  0 disk 
├─dasda1         94:1    0  100M  0 part /boot
└─dasda2         94:2    0  2.2G  0 part 
  ├─vg0-lv_swap 254:0    0  588M  0 lvm  [SWAP]
  └─vg0-lv_root 254:1    0  1.6G  0 lvm  /
dasdb            94:4    0  2.3G  0 disk

# dasdfmt -b 4096 -d cdl -yp /dev/dasdb

# fdasd -a /dev/dasdb

# pvcreate /dev/dasdb1

# vgextend vg0 /dev/dasdb1

# lvextend -l +100%FREE /dev/vg0/lv_root

# resize2fs /dev/vg0/lv_root

# lsblk
NAME            MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
dasda            94:0    0  2.3G  0 disk 
├─dasda1         94:1    0  100M  0 part /boot
└─dasda2         94:2    0  2.2G  0 part 
  ├─vg0-lv_swap 254:0    0  588M  0 lvm  [SWAP]
  └─vg0-lv_root 254:1    0  3.9G  0 lvm  /
dasdb            94:4    0  2.3G  0 disk 
└─dasdb1         94:5    0  2.3G  0 part 
  └─vg0-lv_root 254:1    0  3.9G  0 lvm  /



Tips

If you want to disable swap partition, after finishing NTP client step, quit the installer by pressing Ctrl + C. Then run following command to complete remaining steps:

# setup-disk -s 0

Note: If you do 'sys' installation, there will be no swap partition anymore and 1st partition will be boot, 2nd partition will be root


Known Issues

1. Installation on 2 or more DASDs (either ECKD and FBA) on z/VM is not supported in the installer script at the moment. If you want to install/extend on more than 1 DASD, see "Extending LVM volume". However, installation on 2 or more virtio (SCSI) disks on KVM are supported just like other architectures.

2. The iso image is not supported yet (it is meant for LPAR installation). Use netboot images (kernel and initramfs) for now.


See more

Running Alpine s390x containers on Docker: https://wiki.alpinelinux.org/wiki/S390x/Docker

Retrieved from "https://wiki.alpinelinux.org/w/index.php?title=S390x&oldid=15271"