This procedure documents the full storage provisioning workflow performed on an IBM z/Architecture
Linux guest (lxzlin02, SUSE Linux Enterprise Server 11 on s390x) to bring a new Lotus Domino
mail instance (intranet27) into service.
The workflow covers:
dasdfmt) and partition creation (fdasd)intranet25 and intranet26 instances:
one data volume, one translog volume, and one mail volume (spanning three DASDs).WARNING:
dasdfmtis a destructive operation. All data on the target DASD is erased. Confirm device Bus-IDs carefully before proceeding.
DASD (Direct Access Storage Device) is the IBM term for a disk storage unit attached to a mainframe or IBM Z system. Unlike conventional block devices (e.g. SCSI or SATA disks) that communicate over a PCI or SAS bus, DASDs are connected through IBM's channel subsystem - a dedicated I/O fabric that offloads data transfer from the CPU entirely.
Each DASD is identified by a Bus-ID in the format <css>.<ssid>.<devno> (e.g. 0.0.0317),
where css is the channel subsystem, ssid is the subchannel set, and devno is the
four-hex-digit device number allocated by the system administrator.
The most common DASD type in IBM Z environments is ECKD (Extended Count Key Data),
a track-and-cylinder geometry format that requires a low-level format (dasdfmt) before
the device can be used. This is fundamentally different from conventional disks, where the
physical format is done at the factory. ECKD geometry is expressed in cylinders × heads
(tracks per cylinder), and the block size must be explicitly chosen - 4096 bytes is
standard for modern Linux workloads.
The other type visible in this environment is FBA (Fixed Block Architecture), which
presents a flat sector-addressable surface similar to a conventional disk. FBA devices do
not require dasdfmt.
Under Linux on IBM Z (zLinux), the kernel exposes each active DASD as a block device
under /dev/dasdX (where X is an alphabetic suffix assigned in discovery order).
A DASD must be explicitly brought online via dasd_configure before the kernel
recognises it - devices allocated to the LPAR but not yet activated appear as offline
in lsdasd -a output.
The toolchain used in this procedure maps directly to the DASD lifecycle:
| Tool | Role |
|---|---|
lsdasd |
List all DASDs known to the LPAR and their online/offline state |
dasd_configure |
Set a DASD online or offline by Bus-ID |
dasdfmt |
Low-level format: writes track headers and sets block size |
fdasd |
Partition a DASD by writing a VTOC (Volume Table of Contents) |
Once partitioned, a DASD partition (e.g. /dev/dasdv1) is a standard Linux block device
and can be used directly with LVM (pvcreate, vgcreate, etc.) like any other disk.
sudo) on lxzlin02./intranet27/.lxzlin01) must have a valid SSH client config entry for lxzlin02.The first connection attempt from lxzlin01 to lxzlin02 failed with
Could not resolve hostname lxzlin02: Name or service unknown. The ~/.ssh/config file
was owned by root instead of the connecting user, preventing writes.
# On lxzlin01 - fix ownership of the SSH config file
sudo chown jdoe /home/jdoe/.ssh/config
Then add the lxzlin02 host alias block to ~/.ssh/config (hostname/IP mapping) and
re-attempt the connection:
ssh lxzlin02
Accept the RSA host key fingerprint on first connection. Subsequent connections will
use the cached key from ~/.known_hosts.
Log in to lxzlin02 as root and list all offline DASD devices assigned to this LPAR:
lsdasd -a | grep offline
Devices confirmed offline and ready for provisioning in this session:
| Bus-ID | Purpose |
|---|---|
| 0.0.0316 |
softdomvg extension (2.3 GB) |
| 0.0.0317 |
intranet27datavg (23 GB) |
| 0.0.0321 |
intranet27translogvg (7 GB) |
| 0.0.0318 |
intranet27mailvg member 1 (23 GB) |
| 0.0.0319 |
intranet27mailvg member 2 (23 GB) |
| 0.0.0320 |
intranet27mailvg member 3 (23 GB) |
Use dasd_configure to set each device online. Repeat for every Bus-ID listed above.
dasd_configure 0.0.0316 1 0
dasd_configure 0.0.0317 1 0
dasd_configure 0.0.0321 1 0
dasd_configure 0.0.0318 1 0
dasd_configure 0.0.0319 1 0
dasd_configure 0.0.0320 1 0
The arguments are: <bus-id> <online=1> <ro=0>. Verify each device appears active:
lsdasd
Each device will be assigned a kernel block device name (dasdu, dasdv, dasdw, …).
DASD devices on z/Architecture require a low-level format before use. Use dasdfmt with
a 4096-byte block size and compatible disk layout (-p). The command prints geometry
and prompts for confirmation.
# Format each device - replace /dev/dasdX with the actual device name
dasdfmt -b 4096 -f /dev/dasdu -p # 0.0.0316 - softdomvg extension
dasdfmt -b 4096 -f /dev/dasdv -p # 0.0.0317 - intranet27 data
dasdfmt -b 4096 -f /dev/dasdw -p # 0.0.0321 - intranet27 translog
dasdfmt -b 4096 -f /dev/dasdx -p # 0.0.0318 - intranet27 mail (1/3)
dasdfmt -b 4096 -f /dev/dasdy -p # 0.0.0319 - intranet27 mail (2/3)
dasdfmt -b 4096 -f /dev/dasdz -p # 0.0.0320 - intranet27 mail (3/3)
Type yes at the confirmation prompt for each device. Formatting a 23 GB ECKD DASD
(32 759 cylinders) takes several minutes.
fdasd -a auto-creates a single partition spanning the entire DASD and writes the
VTOC (Volume Table of Contents), which is the IBM equivalent of a partition table.
fdasd -a /dev/dasdu
fdasd -a /dev/dasdv
fdasd -a /dev/dasdw
fdasd -a /dev/dasdx
fdasd -a /dev/dasdy
fdasd -a /dev/dasdz
This produces partition devices: /dev/dasdu1, /dev/dasdv1, /dev/dasdw1,
/dev/dasdx1, /dev/dasdy1, /dev/dasdz1.
Initialize each partition as an LVM Physical Volume:
pvcreate /dev/dasdu1
pvcreate /dev/dasdv1
pvcreate /dev/dasdw1
pvcreate /dev/dasdx1
pvcreate /dev/dasdy1
pvcreate /dev/dasdz1
NOTE: Always use the partition device (
dasdX1), not the raw device (dasdX). LVM on DASD requires the VTOC partition, not the whole disk.
softdomvg)The DASD 0.0.0316 (/dev/dasdu1, 2.3 GB) is used to add a fifth physical volume to
the existing softdomvg group, which hosts IBM software mount points:
vgextend softdomvg /dev/dasdu1
Create the new logical volume within softdomvg:
lvcreate -L 2342M --name optibm4lv softdomvg
NOTE: LVM rounds the requested size up to the nearest physical extent boundary.
2342Mwas rounded to2.29 GB. Use--name(two dashes), not-name.
intranet27
Three new Volume Groups are needed, following the same naming convention as intranet25
and intranet26:
# Data VG - single 23 GB DASD
vgcreate intranet27datavg /dev/dasdv1
# Translog VG - single 7 GB DASD
vgcreate intranet27translogvg /dev/dasdw1
# Mail VG - three 23 GB DASDs (total ~67 GB, matching intranet25mailvg and intranet26mailvg)
vgcreate intranet27mailvg /dev/dasdx1 /dev/dasdy1 /dev/dasdz1
Verify the new VGs are visible:
vgs
Create one LV per VG, sized to match the existing instances:
# Data LV - 22 GB
lvcreate -L 22G --name intranet27datalv intranet27datavg
# Translog LV - 7000 MB (VG capacity is ~6.88 GB; requesting 7 GB or 7042 MB exceeds
# the available extents - use 7000M as a safe value)
lvcreate -L 7000M --name intranet27transloglv intranet27translogvg
# Mail LV - 66 GB
lvcreate -L 66G --name intranet27maillv intranet27mailvg
NOTE: The translog VG (
intranet27translogvg) has ~6.88 GB total. Requesting7G,6.88G,7043M, or7042Mall fail because LVM requires a few extents for metadata. Use7000Mto stay within the available extent count.
Verify all LVs:
lvs
Format each new logical volume with ReiserFS (the filesystem used by existing instances on this server):
mkreiserfs /dev/softdomvg/optibm4lv
mkreiserfs /dev/intranet27datavg/intranet27datalv
mkreiserfs /dev/intranet27translogvg/intranet27transloglv
mkreiserfs /dev/intranet27mailvg/intranet27maillv
Type y at the confirmation prompt for each. mkreiserfs will display the UUID assigned
to each volume.
/etc/fstab
Back up the current fstab before editing:
cp /etc/fstab /etc/fstab-20110714
Add mount entries for each new logical volume. Use the LVM device mapper path
(/dev/<vgname>/<lvname>). Example entries following the pattern of existing instances:
/dev/intranet27datavg/intranet27datalv /intranet27/data reiserfs defaults 0 0
/dev/intranet27translogvg/intranet27transloglv /intranet27/translog reiserfs defaults 0 0
/dev/intranet27mailvg/intranet27maillv /intranet27/mail reiserfs defaults 0 0
NOTE: Adjust mount point paths to match the actual Domino directory layout under
/intranet27/before saving.
After completing all steps, confirm the final storage layout:
# All target DASDs active
lsdasd
# All VGs present with correct sizes and 0 free space (fully allocated)
vgs
# All LVs present with expected sizes
lvs
Expected final VG state for the new instance:
| VG | PVs | LVs | Size | Free |
|---|---|---|---|---|
intranet27datavg |
1 | 1 | 22.49 G | ~504 MB |
intranet27mailvg |
3 | 1 | 67.48 G | ~1.48 G |
intranet27translogvg |
1 | 1 | 6.88 G | ~40 MB |
| Symptom | Cause | Resolution |
|---|---|---|
Could not resolve hostname on SSH |
Host alias missing or ~/.ssh/config unreadable |
Fix file ownership (chown user ~/.ssh/config) and add host block |
Device 0.0.0XXX is already online |
dasd_configure called twice |
Harmless; device is online - continue |
Insufficient free extents on lvcreate
|
Requested size exceeds VG capacity after metadata | Reduce size by a few MB (e.g. use 7000M instead of 7042M) |
Volume group "X" not found on lvcreate
|
Used -name (one dash) instead of --name
|
Always use --name for the LV name option |
pvcreate on raw device (/dev/dasdX) |
Tab-completion selected the whole disk | Use the partition (/dev/dasdX1) - the one with the 1 suffix |