☼solaris.pro
ORACLE SOLARIS 11.4 • FIELD GUIDE

Know what your
Solaris can do.

A practical reference for administrators and engineers: commands, concepts, working examples, recovery paths and the mental models that make Solaris easier to operate.

Explore the deep dives →
Getting startedZFSZonesLDomsSMFDTrace
root@solaris
# uname -a
SunOS solaris 5.11 11.4 sun4v sparc

# ldm list
NAME     STATE   FLAGS   CONS   VCPU  MEMORY
primary  active  -n-cv-  UART   16    32G
web01    active  -n----  5000   8     16G

# zoneadm list -cv
  ID NAME       STATUS     BRAND
   0 global     running    solaris
   1 app-zone   running    solaris
FIELD GUIDEZFSZONESLDOMSSMFIPSDTRACEIPADMRBACFMABOOT ENVIRONMENTS
THE DEEP DIVES · 31 FIELD MANUALS

Start with the practical guides.

Detailed walkthroughs, real-world scenarios, commands, interpretation and precautions. Explore LDoms, DTrace, ZFS, Zones and more.

Open complete guide library →
Logical DomainsLogical Domains: architecture and designUnderstand what the SPARC hypervisor isolates, how domain roles fit together, and when to use LDoms rather than Zones.Read guide →Logical DomainsBuild a guest LDom from scratchA guided SPARC guest build: services, disk backend, networking, CPU, memory, bind, boot and console.Read guide →Logical DomainsLDoms virtual disks and ZFS backendsTrace I/O from guest vdisk to service-domain vds to backing zvol or disk, and debug common failures.Read guide →Logical DomainsLDoms virtual networking, VLANs and resiliencyUnderstand vnet/vsw, service-domain uplinks, VLAN isolation and fault domains.Read guide →Logical DomainsLDom migration: prerequisites and runbookPlan guest migrations with a preflight, execution window, verification and rollback decision.Read guide →Logical DomainsLDoms troubleshooting playbooksDiagnose failed binds, unbootable guests, missing disks, broken virtual networking and service-domain dependencies.Read guide →DTraceDTrace foundations: probes, clauses and safetyLearn the D language, probe descriptions, predicates, actions, privilege requirements and safe observation.Read guide →DTraceDTrace cookbook: syscalls and latencyInvestigate slow system calls, high syscall rates, failed opens and unexpected process activity.Read guide →DTraceDTrace cookbook: CPU profiling and hot stacksUse sampling and stacks to discover where CPU time goes.Read guide →DTraceDTrace cookbook: filesystem and disk I/OCorrelate filesystem calls, I/O completion times, ZFS behavior and block-device symptoms.Read guide →DTraceDTrace cookbook: network diagnosisTrace network-related syscalls and combine kernel observations with Solaris networking tools.Read guide →DTraceDTrace cookbook: applications and processesMeasure process startup, scheduling and user-level activity without changing application code.Read guide →DTraceDTrace advanced: aggregations, predicates and overheadBuild robust scripts with per-thread state, histograms, filtered probes and bounded output.Read guide →DTraceDTrace investigations: six end-to-end scenariosWalk from an operational symptom to a falsifiable hypothesis and a trace that can prove or reject it.Read guide →Getting startedSolaris 11.4 first-day checklistA safe order of operations for a freshly installed Solaris hostRead guide →Getting startedSolaris 10 to 11: what changedTranslate legacy administration habits to IPS, SMF, ZFS and modern networkingRead guide →AdministrationSMF service operations and debuggingDiagnose maintenance states, dependencies, logs and restartsRead guide →ZFSZFS pools: design and operationsUnderstand mirrors, RAID-Z, vdevs, scrubs and capacity planningRead guide →ZFSZFS datasets, properties and quotasBuild manageable filesystem boundaries and prevent uncontrolled growthRead guide →ZFSZFS snapshots, send/receive and recoveryUse snapshots and replication for recoverability, not as a substitute for tested backupsRead guide →ZFSZFS encryption and storage securityProtect data at rest while keeping key handling separate from data copiesRead guide →ZonesNative Zones: build and operateProvision and troubleshoot isolated Solaris application environmentsRead guide →ZonesKernel Zones: isolation and lifecycleCompare Kernel Zones with native Zones and LDomsRead guide →NetworkingIP administration with ipadmConfigure persistent addresses and debug routes and interface stateRead guide →NetworkingDatalinks, VLANs and aggregationUnderstand physical links, virtual NICs and VLAN mappingRead guide →SoftwareIPS packages and publishersFind, verify, install and manage packages with IPSRead guide →SoftwareBoot Environments and rollbackStage updates with bootable ZFS environmentsRead guide →SecurityRBAC, privileges and auditingApply least privilege and inspect rights without relying on shared root credentialsRead guide →ObservabilityFault Management Architecture (FMA)Interpret hardware and software fault reportsRead guide →ObservabilitySolaris performance triageDetermine whether CPU, memory, storage or network is limiting throughputRead guide →ReferenceTroubleshooting decision treesSymptom-driven isolation paths for common Solaris incidentsRead guide →

QUICK ORIENTATION

Explore Solaris by subject.

Use this as a runbook, refresher, or jumping-off point into Oracle's primary documentation.

01 / START

Getting Started

Installation through the first production checks.

  • Solaris 10 → 11 mindset
  • First 30 minutes
  • First-day checklist
Open guide →
02 / ADMIN

Administration

Users, processes, filesystems, services and scheduled work.

  • RBAC and users
  • SMF and logs
  • Processes and cron
Open guide →
03 / STORAGE

ZFS

Pools, datasets, integrity, snapshots and replication.

  • Send / receive
  • Encryption & compression
  • Recovery patterns
Open guide →
04 / VIRT

Zones

OS-level isolation with native and Kernel Zones.

  • Lifecycle
  • Resources & networking
  • Migration concepts
Open guide →
05 / SPARC

Logical Domains

SPARC hypervisor virtualization with Oracle VM Server for SPARC.

  • Control/service/guest roles
  • Virtual disk & network
  • SP configs & migration
Open guide →
06 / NETWORK

Networking

Modern Solaris network configuration and diagnosis.

  • ipadm & dladm
  • VLAN, aggregation, IPMP
  • Routing, DNS, NFS, SMB
Open guide →
07 / OBSERVE

Observability

Measure the system before changing it.

  • DTrace
  • prstat / vmstat / iostat
  • kstat & StatsStore
Open guide →
08 / LIFECYCLE

Software & Recovery

IPS, publishers, updates and reversible change.

  • pkg
  • Boot Environments
  • Unified Archives
Open guide →
09 / SECURITY

Security

Least privilege and auditable administration.

  • RBAC & privileges
  • Auditing & SSH
  • Encryption & immutable zones
Open guide →
MENTAL MODEL

Solaris virtualization is layered.

Pick the boundary you actually need. LDoms partition SPARC hardware resources; Zones virtualize at the OS level; Kernel Zones add a separate kernel boundary.

SPARC hardwareCPU threads, memory, PCIe, storage, NICs
HypervisorPartitions resources into Logical Domainsldm
Guest LDomIndependent OS instance and identity
Zones / KZonesFurther workload isolation inside Solariszoneadm
ApplicationsDatabase, middleware, web, services
Example architectureA SPARC host can dedicate an LDom to a database and another LDom to application infrastructure. The application LDom can then host several lightweight native Zones, while a Kernel Zone can isolate a workload that needs its own kernel/SRU lifecycle.
01 / GETTING STARTED

Your first hour on Solaris 11.4.

Inventory first. Change second.

First 30 minutes

  1. Identify release, platform and virtualization layer.
  2. Check service health and known faults.
  3. Inspect storage pools before touching filesystems.
  4. Understand network links and IP objects.
  5. Check current boot environment and package publishers.
Coming from Solaris 10?Expect SMF instead of traditional service scripts, IPS instead of SVR4 package workflows, ZFS-centric boot environments, and ipadm/dladm for persistent network configuration.
uname -a
cat /etc/release
virtinfo -a
svcs -xv
fmadm faulty
zpool status
zfs list
ipadm show-if
ipadm show-addr
dladm show-link
beadm list
pkg publisher

First-day checklist

02 / ADMINISTRATION

Operate the system, not just the shell.

Solaris administration is built around explicit service, privilege and resource models.

Users & RBAC

Prefer roles, rights profiles and explicit authorizations over routine all-powerful access.

profilesrolesauths
Processes

Start with process identity and resource use before tracing deeper.

ps -efprstatpfiles
Logging

Correlate application logs with SMF state, faults and system messages.

svcs -xvfmadm faultydmesg
03 / ZFS

Storage built around pools and datasets.

Health, integrity, properties, snapshots and replication belong in the same operational model.

The mental model

A pool aggregates storage. Datasets consume pool space and carry properties. Snapshots capture dataset state; send/receive turns those snapshots into a replication primitive.

Recovery ruleA snapshot in the same pool is useful rollback history, not an independent backup. Keep separate copies and test restores.
zpool list
zpool status -v
zpool iostat -v 2
zfs list -o name,used,avail,refer,mountpoint
zfs get compression,recordsize,atime tank/data
zfs snapshot tank/data@pre-change
zfs send -R tank/data@pre-change | \
  zfs receive backup/data

Use case: safe application change

ZFS

Take a recursive snapshot immediately before a risky application change. If the application stores data across several child datasets, snapshot at the common parent so the recovery point is easy to identify.

zfs snapshot -r tank/app@before-upgrade
# perform change
zfs list -t snapshot -r tank/app
04 / ZONES

Fast isolation with Solaris-native tooling.

Native Zones share the global zone kernel; Kernel Zones provide a separate kernel environment.

Native Zone lifecycle

zonecfg -z web01
create
set zonepath=/zones/web01
set autoboot=true
verify
commit
exit

zoneadm -z web01 install
zoneadm -z web01 boot
zlogin web01

Where Zones fit

Use native Zones when you want lightweight Solaris workload isolation and shared-kernel efficiency. Use Kernel Zones when a separate kernel environment is important. On SPARC, either can also live inside a suitable LDom.

Use caseConsolidate several application tiers into separate native Zones inside one guest LDom, while retaining the LDom as the stronger hardware-resource and OS-instance boundary.
05 / LOGICAL DOMAINS

LDoms: SPARC virtualization from the hardware up.

Oracle VM Server for SPARC Logical Domains are independent virtual machines managed by the Logical Domains Manager in the control domain.

Control domain

Runs Logical Domains Manager and controls domain configuration.

Service domain

Provides virtual services such as disk and network I/O to guests.

I/O domain

Owns direct access to physical I/O resources and can provide them to other domains.

Guest domain

Consumes assigned CPU, memory and virtual/physical I/O and runs its own OS.

Physical SPARC→Hypervisor→primary / service domains→virtual disk + network→guest LDom

Inventory before changing

Start by identifying the domain topology, constraints and services. Remember that changes can involve both live state and the configuration selected for the next chassis power cycle.

ldm -V
ldm list
ldm list -l
ldm list-services
ldm list-bindings primary
ldm list-constraints
ldm list-spconfig
Persistence mattersAfter reaching a stable configuration, save it as an SP configuration and keep an XML copy. Treat that configuration as recoverable infrastructure state.

Core lifecycle

# Example skeleton — size to your system
ldm add-domain app01
ldm add-vcpu 8 app01
ldm add-memory 16G app01

# Add virtual I/O after defining services
ldm bind-domain app01
ldm start-domain app01
ldm list app01

# Graceful lifecycle
ldm stop-domain app01
ldm unbind-domain app01

LDom commands are hardware/configuration-sensitive. Treat examples as workflow patterns and verify supported resources, firmware and device names on the target system.

Use case: virtual disk backed by ZFS

LDOM + ZFS

A common pattern is to create storage in the service domain, export it through a virtual disk service, then attach it to a guest. This keeps guest storage lifecycle separate from physical device naming.

# Illustrative names — inspect your existing VDS first
zfs create -V 40G rpool/ldoms/app01-root
ldm add-vdsdev /dev/zvol/dsk/rpool/ldoms/app01-root \
  app01-root@primary-vds0
ldm add-vdisk vdisk0 app01-root@primary-vds0 app01
ldm list-bindings app01

Use case: virtual networking

LDOM + VNET

Guests normally connect through virtual network devices backed by a virtual switch/service domain. Inspect existing service names and the physical link before adding devices.

ldm list-services
ldm list-bindings primary
# Example only; service names vary
ldm add-vnet vnet0 primary-vsw0 app01
ldm list-bindings app01
CPU & memory

Allocate virtual CPUs and memory according to workload needs; understand whether a requested change is dynamic or requires rebinding.

I/O resilience

A production design can use multiple service domains and redundant virtual I/O paths to reduce single points of failure.

Migration

Domain migration has source/destination compatibility, storage and networking prerequisites. Validate the Oracle matrix before treating it as a generic live-migration feature.

Use case: LDom + Zones consolidation

Layered virtualization

Give a workload family an independent Solaris instance in an LDom, then run several native Zones inside it. This separates host-level resources and OS lifecycle from lightweight application isolation. A Kernel Zone can be used where a separate kernel/SRU boundary is desirable.

06 / NETWORKING

Links first. IP objects second.

dladm answers link-layer questions; ipadm answers IP configuration questions.

dladm show-phys
dladm show-link
dladm show-vlan
dladm show-aggr
ipadm show-if
ipadm show-addr
netstat -rn
route -p show
getent hosts example.com

Isolation path

physical link→datalink/VLAN→IP interface→address→route→DNS/app
LDom noteWhen troubleshooting a guest, include the vnet → virtual switch → service-domain physical link path. A healthy IP stack inside the guest does not prove the backing service path is healthy.
07 / SMF

Services have state, dependencies and history.

Don't just restart a broken service: ask SMF why it is broken.

svcs -xv
svcs -l svc:/network/ssh
svcs -d svc:/network/ssh
svcs -D svc:/network/ssh
svccfg -s svc:/network/ssh listprop
svcadm restart svc:/network/ssh
svcadm clear svc:/example/service
08 / OBSERVABILITY

Measure before you tune.

Use cheap broad tools first, then narrow the question with DTrace or MDB.

CPU / processesprstat -mL 2ps -efmpstat 2
Memory / VMvmstat 2echo ::memstat | mdb -kkstat
I/Oiostat -xn 2zpool iostat -v 2fsstat

DTrace: count system calls by process

DTrace
dtrace -n 'syscall:::entry { @[execname] = count(); }'

Start with aggregation rather than printing every event. It reduces noise and makes the first question—“who is doing this?”—easier to answer.

09 / SECURITY

Least privilege is a Solaris feature, not a slogan.

Combine RBAC, privileges, auditing, service minimisation and encryption according to the workload.

profiles
roles
auths
ppriv $$
auditconfig -getpolicy
svcs -a
pkg verify
Administrative patternDefine the task an operator needs to perform, then grant a suitable rights profile/authorization rather than distributing unrestricted root access by default.
10 / SOFTWARE & RECOVERY

Make change reversible.

IPS and ZFS Boot Environments turn system updates into a lifecycle rather than a one-way event.

Before updating

pkg publisher
pkg list -u
pkg update -nv
beadm list
beadm create pre-maintenance

After updating

pkg history
pkg verify
beadm list
svcs -xv
fmadm faulty
Operational habitName manual boot environments after the change window or purpose so the rollback choice is obvious from the boot menu.
SYMPTOM-FIRST

Troubleshooting decision paths.

Start with evidence that separates layers instead of jumping directly to configuration changes.

QUICK REFERENCE

Command index.

Filter by command, task or subsystem.

CommandAreaUse
DEEP REFERENCE

Things worth bookmarking.

Fast reminders for the concepts that tend to matter during incidents.

SMF states

online, offline, disabled, maintenance, degraded, legacy_run.

ZFS properties

Check inheritance before changing a property. A value may be local, inherited, received or default.

LDom persistence

Distinguish current domain configuration from the SP configuration selected for the next power cycle.

PRIMARY SOURCES

Go deeper at Oracle.

solaris.pro is the practical field guide; Oracle documentation remains the authority for platform support, prerequisites and exact release behaviour.

Oracle Solaris 11.4 Documentation

Administration, ZFS, Zones, networking, security, observability and reference manuals.

Oracle VM Server for SPARC 3.6

LDoms installation, administration, migration, virtual I/O, security and release notes.