Cleanup LDom Memory Fragmentation

Over time after many online memory changes, the memory may become
fragmented like this example. The small blocks prevent the system 
from allocating large pages which causes degraded performance.

The fragments typically arise when you reduce memory
while the domain is running. It is recommended to reduce
memory while the guest domain is down and use start-reconf
for the control domain.

# ldm list-devices -a memory

MEMORY
    PA                   SIZE            BOUND
    0x0                  60M             _sys_
    0x3c00000            32M             _sys_
    0x5c00000            94M             _sys_
    0xba00000            1030M           _sys_
    0x4c000000           64M             _sys_
    0x50000000           2560M           primary
    0xf0000000           256M            marcel
    0x100000000          2560M           primary
    0x1a0000000          256M            marcel
    0x1b0000000          5888M           primary
    0x320000000          256M            marcel
    0x330000000          768M            primary
    0x360000000          256M            marcel
    0x370000000          3584M           primary
    0x450000000          11G             marcel
    0x710000000          15G             g0075
    0xad0000000          6G              g0095
    0xc50000000          8G              g0081
    0xe50000000          1792M
    0xec0000000          2304M           primary
    0xf50000000          2560M
    0xff0000000          128M            _sys_
    0xff8000000          64M             _sys_
    0xffc000000          64M             _sys_
    0x400000000000       2G              g0056
    0x400080000000       1G              primary
    0x4000c0000000       1G              g0056
    0x400100000000       5888M           primary
    0x400270000000       8G              g0056
    0x400470000000       8G              g0079
    0x400670000000       1G              g0056
    0x4006b0000000       8G              g0107
    0x4008b0000000       10G             g0058
    0x400b30000000       4G              g0082
    0x400c30000000       8G              s0004
    0x400e30000000       4G              g0082
    0x400f30000000       3G
    0x400ff0000000       192M            _sys_
    0x400ffc000000       64M             _sys_

Since Solaris 11.4 SRU39 the ldm command add-spconfig offers
the new -m flag to cleanup the memory configuration.

# ldm add-spconfig -m s0003m

we need to power cycle the control domain.
shutdown all guest domains.

For users of JomaSoft VDCF you can use
gdom -c shutdown cdom=s0003

afterwards we poweroff and poweron the control domain
to activate the new configuration.

# init 0

after the system is down at the ok prompt.

{0} ok
Serial console stopped.

-> stop /SYS
Are you sure you want to stop /SYS (y/n)? y
Stopping /SYS

-> show /SYS power_state

 /SYS
    Properties:
        power_state = Off

-> start /SYS
Are you sure you want to start /SYS (y/n)? y
Starting /SYS


this improved the memory allocation of all the domains

# ldm list-devices -a memory

MEMORY
    PA                   SIZE            BOUND
    0x0                  60M             _sys_
    0x3c00000            32M             _sys_
    0x5c00000            94M             _sys_
    0xba00000            1030M           _sys_
    0x4c000000           64M             _sys_
    0x50000000           24G             primary
    0x650000000          15G             g0075
    0xa10000000          8G              g0081
    0xc10000000          8G              g0082
    0xe10000000          6G              g0095
    0xf90000000          1536M           marcel
    0xff0000000          128M            _sys_
    0xff8000000          64M             _sys_
    0xffc000000          64M             _sys_
    0x400000000000       12G             g0056
    0x400300000000       10G             g0058
    0x400580000000       8G              g0079
    0x400780000000       8G              g0107
    0x400980000000       10752M          marcel
    0x400c20000000       8G              s0004
    0x400e20000000       7424M
    0x400ff0000000       192M            _sys_
    0x400ffc000000       64M             _sys_

New postgres packages 08/2026 are ready for Solaris 11.4 SPARC

New postgres packages are ready for Solaris 11.4 SPARC

You don't need to download source and compile yourself for Solaris.

We provide compiled postgres as Solaris 11.4 SPARC IPS pkgs.
Current versions are 18.6, 17.11, 16.15, 15.19, 14.24

Free download and no registration
https://www.jomasoft.ch/downloads/#js-opensource

New Features in Solaris 11.4 SRU93 (June 2026)

Another quarterly Solaris SRU including new features

Kernel Crypto Framework SMF Policy
NFS Sockets

FOSS: Perl 5.42, PHP 8.5, bash5.3
EOF: mySQL 8.0, IKEv1

New postgres packages are ready for Solaris 11.4 SPARC

You don't need to download source and compile yourself for Solaris.

We provide compiled postgres as Solaris 11.4 SPARC IPS pkgs.
Current versions are 18.4, 17.10, 16.14, 15.18, 14.23

Free download and no registration
https://www.jomasoft.ch/downloads/#js-opensource

Oracle changes Solaris 11.4 SRU delivery

In the past Oracle provided Solaris SRUs (Support Repository Updates)
each month. This does change now. The types of SRUs are reduced
from 3 types to 2 types. Instead of 12 SRUs per year we will see 8 SRUs
per year.

a) Feature SRU (NEW)

This will be available early in the months 
March, June, September, December
Around 6 weeks after the CPUs.

This SRU includes new features, tools, packages, version updates
and removal of old, outdated/insecure software.

I would recommend to install this SRU on test systems to verify
application compatibility and identify required configuration changes.


b) CPU (Critical Patch Updates)

No changes here. 

Available January, April, July, October
The release dates are predefined and are published on
https://www.oracle.com/security-alerts/#CriticalPatchUpdates

This SRU includes critical fixes and security updates.
Most of our enterprise customers update to this SRU.


For critical issues Oracle will release individual 
IDRs (Interim Diagnostic or Relief).

The Next Oracle Solaris 11.4.90 CBE is available for Developers

CBE is the Common Build Environment

The first release 11.4.42 was made available by Oracle 03/2022.
Last year 05/2025 the version 11.4.81 was published.
And now 03/2026 11.4.90 is already available.

This release can be used by Developers to develop, test and demonstrate
their software for free. It includes current compilers and tools.

Download from 
https://www.oracle.com/solaris/solaris11/downloads/solaris-downloads.html

This is great news and confirms Oracle continues investing into Oracle Solaris.

 

New Features in Solaris 11.4 SRU90 (Feb 2026)

Another quarterly Solaris SRU including new features

devnm -s
beadm preserve policy
ls -sh output change
bootadm GRUB unrestricted property
ikev2cert PIN from SMF
nscfg check
new smf svc:/system/check/name-services:default
ipadm _forward_src_routed IPv6 property removed

FOSS: BIND 9.20
EOF: GCC 12

New postgres packages are ready for Solaris 11.4 SPARC

You don't need to download source and compile yourself  for Solaris.

We provide compiled postgres as Solaris 11.4 SPARC IPS pkgs.
Current versions are 18.2, 17.8, 16.12, 15.16, 14.21

Free download and no registration
https://www.jomasoft.ch/downloads/#js-opensource

The State of Oracle Solaris - Jan 2026

Oracle Solaris is not much in the press anymore, but used
for critical systems around the world.

Solaris 11.4 was launched in 2018. Oracle provides updates/patches
every month. Every quarter (Feb,May,Aug,Nov) the update includes new features.
Oracle will provide this monthly updates till at least November 2037.

Documented in:
http://www.oracle.com/us/support/library/lifetime-support-hardware-301321.pdf

Oracle is working to make sure Solaris will run properly after 2037 for example
with changes in SRU72 delivered in August 2024. It would be no surprise
if the updates will continue after November 2037.
https://blogs.oracle.com/solaris/announcing-oracle-solaris-114-sru72

For developers a pretty current version (CBE) is available
for free to develop and test software.

Solaris CBE Download:
https://www.oracle.com/solaris/solaris11/downloads/solaris-downloads.html

Oracle even provides patches for the older releases Solaris 10
and Solaris 11.3. The last patches are expected by January 2027.
We highly recommend to migrate to Solaris 11.4 and we are
happy to support you with our very experiences engineers.

Oracle's Management and Monitoring tools (OpsCenter and VM Manager for SPARC) 
reached End-Of-Life. No patches are provided anymore. 

We at JomaSoft continue to enhance our VDCF management and monitoring tool 
for Oracle Solaris and Linux
https://www.jomasoft.ch/vdcf/

If you don't have a SPARC server yourself, maybe the JomaSoft
SPARC cloud could be helpful for you.
https://www.jomasoft.ch/services/ 

New Features in Solaris 11.4 SRU87 (Nov 2025)

Another quarterly Solaris SRU including new features

useradd -N / -U 
cableinfo
fmthard -c / -e
ldm add-/list-/remove-devalias


FOSS: GCC 15
EOF: PHP 8.1, Python 3.9

JomaSoft provides OpenJDK 17.0.16 package for Solaris 11.4 SPARC

Since three years JomaSoft produces Solaris 11.4 IPS packages
of opensource software. After postgres openjdk was added.
After 17.0.13 in December 2024, we provide now a package
with version 17.0.16.

Based on the work done by Peter Tribble. Thanks Peter!
https://github.com/ptribble/jdk-sunos-builder

Download Page
https://www.jomasoft.ch/downloads/#js-openjdk

Direct p5p Download Link
https://bit.ly/js-openjdk17-sparc

The package is provided without support or warranty.
Use at your own risk.

New Features in Solaris 11.4 SRU81 (May 2025)

Another quarterly Solaris SRU including new features

GRUB 2.12
Zone Migration requires NTP
ipadm set-ifprop -p send-unreachables
dladm show-transceiver
s10 brand compressed install archives
ldm set-logctl =func =v (verbosity)

OVM Manager SPARC Agent does not work anymore

FOSS: PHP 8.4
EOF: OpenSSL 1.0.2, Ruby 3.1, GCC 7, stackdb SMF

More details on the Solaris Blog

Solaris 11.4.78 / iostat -b and wildcards

If you are using Kernel Zones

-bash-5.2$ zoneadm list -v
  ID NAME             STATUS      PATH                         BRAND      IP
   0 global           running     /                            solaris    shared
   1 v0182            running     -                            solaris-kz excl

It is now easier on the global zone to analyze
the I/O of only the disks used by kernel zones

using the wildcard (v0182*) and the new -b (backend) flag

-bash-5.2$ iostat -xnzb -Td v0182* 5 3
Tuesday, April 29, 2025 at  7:58:29 AM EDT
                    extended device statistics
    r/s    w/s   kr/s   kw/s wait actv wsvc_t asvc_t  %w  %b device
   0.3    7.5   18.2   53.7  0.0  0.0    0.0    1.7   0   0 v0182/disk1
   0.0    0.0    0.0    9.3  0.0  0.0    0.0    5.1   0   0 v0182/disk2
   0.0    0.0    0.0    0.0  0.0  0.0    0.0    0.2   0   0 v0182/disk3
Tuesday, April 29, 2025 at  7:58:34 AM EDT
                    extended device statistics
    r/s    w/s   kr/s   kw/s wait actv wsvc_t asvc_t  %w  %b device
   0.6   14.4   25.8   68.4  0.0  0.0    0.0    1.8   0   1 v0182/disk1
   0.4   68.2    0.3 58301.9  0.0  0.4    0.0    6.5   0  24 v0182/disk2
Tuesday, April 29, 2025 at  7:58:39 AM EDT
                    extended device statistics
    r/s    w/s   kr/s   kw/s wait actv wsvc_t asvc_t  %w  %b device
  11.0    0.0   37.1    0.0  0.0  0.0    0.0    0.3   0   0 v0182/disk1
   0.0   87.2    0.0 52939.6  0.0  0.4    0.0    5.1   0  22 v0182/disk2


And we have similar output on control domains for disks
used in logical domains (Ldoms) on SPARC servers.

-bash-5.2$ iostat -znxMb -T d vds* 5 5
Tuesday, April 29, 2025 at  2:20:19 PM CEST
                    extended device statistics
    r/s    w/s   Mr/s   Mw/s wait actv wsvc_t asvc_t  %w  %b device
 356.5  120.3   16.7   11.1  0.0  0.3    0.0    0.7   0   8 60002AC000000000000004320001507B@primary-vds0
 187.3    5.1   10.2    0.1  0.0  0.1    0.0    0.5   0   9 60002AC0000000000000048F0001507B@primary-vds0
Tuesday, April 29, 2025 at  2:20:24 PM CEST
                    extended device statistics
    r/s    w/s   Mr/s   Mw/s wait actv wsvc_t asvc_t  %w  %b device
 135.2  195.0   80.3  118.0  0.0  0.9    0.0    2.8   0  50 60002AC000000000000004320001507B@primary-vds0
Tuesday, April 29, 2025 at  2:20:29 PM CEST
                    extended device statistics
    r/s    w/s   Mr/s   Mw/s wait actv wsvc_t asvc_t  %w  %b device
 178.6  275.6  115.6  103.2  0.0  1.1    0.0    2.3   0  55 60002AC000000000000004320001507B@primary-vds0
Tuesday, April 29, 2025 at  2:20:34 PM CEST
                    extended device statistics
    r/s    w/s   Mr/s   Mw/s wait actv wsvc_t asvc_t  %w  %b device
 236.6  306.4  100.2  109.9  0.0  1.1    0.0    2.1   0  61 60002AC000000000000004320001507B@primary-vds0
Tuesday, April 29, 2025 at  2:20:39 PM CEST
                    extended device statistics
    r/s    w/s   Mr/s   Mw/s wait actv wsvc_t asvc_t  %w  %b device
 196.2  282.0   95.4  117.2  0.0  1.1    0.0    2.4   0  60 60002AC000000000000004320001507B@primary-vds0

Oracle Systems TechDay 13.05.2025 in Prague

Tuesday, May 13th 2025

Onsite event in Prague, Czech Republic

About
- Oracle Systems/Server/SPARC/ZFS/Solaris
- AI, Cloud
- Hybrid Storage
- Secure Desktop

As usual Oracle Product Managers, Engineering and JomaSoft
are there.

Registration at:
https://eventreg.oracle.com/profile/web/index.cfm?PKwebID=0x912115abcd

system-repository SMF access_log cleanup

The system-repository SMF delivers IPS packages to zones.

This SMF produces an access_log file which is not rotated
by default. If you run many zones this logfile grows
to multiple 100 MB. On a customer system with more
than 20 zones this file was over 3 GB in size.

This is a sample on one of our test systems

# ls -lh /var/log/pkg/sysrepo/access_log
-rw-r--r--   1 pkg5srv  bin         148M Jan  8 10:52 /var/log/pkg/sysrepo/access_log

You can manually rotate and compress this logfile using
the following command

# time logadm -c -z 0 /var/log/pkg/sysrepo/access_log

real    0m4.989s
user    0m4.164s
sys     0m0.815s

# ls -lh /var/log/pkg/sysrepo/access_log*
-rw-r--r--   1 pkg5srv  bin            0 Jan  8 10:54 /var/log/pkg/sysrepo/access_log
-rw-r--r--   1 pkg5srv  bin        26.7M Jan  8 10:52 /var/log/pkg/sysrepo/access_log.0.gz

New Features in Solaris 11.4 SRU75 (Nov 2024)

Another quarterly Solaris SRU including new features

zfs mount/unmount -r
pgrep/pkill -Z
cups SMF disabled by default

FOSS: gcc14
EOF: snort, Node.js 18

 

Cleanup Oracle Solaris Audit files

Auditing is enabled by default on Solaris 11.4 and records security-related system events,
like logins, reboots, etc.

The audit files are stored in /var/share/audit.
Overtime this files use quiet some disk space

# cd /var/share/audit/
# du -sh
18.3G   .

For later analysis a backup of this files is recommended.
Maybe you have a central archive or security tools for this.

On the Solaris server there is no automatically cleanup
of this files active. This should be done manually from time to time.

For example to delete files older than 3 years.
find /var/share/audit -mtime +1095 -exec rm {} \;

If you are using JomaSoft VDCF you can execute this on all your Solaris 11 Nodes

# su
Password:
# echo "# cleanup audit files older than 3 years" >/var/opt/jomasoft/vdcf/config/script/cleanup_audit
# echo "find /var/share/audit -mtime +1095 -exec rm {} \;" >>/var/opt/jomasoft/vdcf/config/script/cleanup_audit

# exit
-bash-5.2$ config -c add type=SCRIPT name=cleanup_audit script=cleanup_audit os=11
Configuration SCRIPT successfully added.

-bash-5.2$ serverconfig -c exec servertype=node type=SCRIPT name=cleanup_audit
INFO: Servertype <node> selected, the following 'running' server are processed:
....

To learn more about Solaris Auditing take a look into the
Doc "Managing Auditing in Oracle Solaris 11.4"
https://docs.oracle.com/cd/E37838_01/html/E61027/index.html

Checkout what you can do with JomaSoft VDCF
https://www.jomasoft.com/vdcf/

New Features in Solaris 11.4 SRU72 (Aug 2024)

Another quarterly Solaris SRU including new features

System Account Check Service (svc:/system/check/user)
iostat sstore collection
Per-disk kstats for vds/zvblk
WebUI - LDom sheets
ZFS clonedir

FOSS: +suricata (IDS)

EOF: Python 3.7, Tomcat 8.5, Perl 5.36, Snort


Find details on the Oracle Solaris Blog

 

How to send files to Oracle Support SR using ASR Manager

Most of the time it is required to attach additional files to Oracle Support Cases,
for example ILOM/XSCF snapshots or Logfiles.

If you use the ASR Manager on Solaris it is very easy to attach
additional files. It includes a transport tool. No need to copy files to
your local machine and to upload using a Browser.

Check if you use the ASR Manager

# svcs |grep asrm
online         2024-05-08T17:29:45 svc:/application/management/asrm:default

You can just copy your files into this directory

/var/opt/asrmanager/sftransport/transfer/

The files need to start with your SR-Number, for example
3-31234567890.mylogfile

The files are transported to Oracle Support and attached
to your SR every 10 minutes.

You can check the state using

/opt/asrmanager/bin/sftransport info

Solaris 11.4 SRU60 - fc port tabular output

The default output of fcinfo hba-port contains a lot of values,
but is not optimal if you are looking for a specific attribute

# fcinfo hba-port 2100000e1ef99410
HBA Port WWN: 2100000e1ef99410
        Port Mode: Initiator
        Port ID: 10200
        OS Device Name: /dev/cfg/c5
        Manufacturer: Marvell Technology, Inc
        Model: 7023303
        Firmware Version: 8.08.04
        FCode/BIOS Version:  BIOS: 3.19; fcode: 4.02; EFI: 5.36;
        Serial Number: 463916R+1729340085
        Driver Name: qlc
        Driver Version: 230206-5.12
        Type: N-port
        State: online
        Supported Speeds: 4Gb 8Gb 16Gb
        Current Speed: 8Gb
        Node WWN: 2000000e1ef99410
        Max NPIV Ports: 253
        NPIV port list:

Solaris 11.4 SRU60 (August 2023) introduces the new -T, -O and -o flags

default tabular format

# fcinfo hba-port -T 2100000e1ef99410
PORT_WWN         PARENT_WWN       MODE      CTRL    STATE   SPEED
2100000e1ef99410 NA               Initiator c5      online  8Gb

Output for admins

# fcinfo hba-port -o port_wwn,driver_name,link_failure_count,state,speed 2100000e1ef99410
PORT_WWN         DRIVER_NAME         LINK_FAILURE_COUNT STATE   SPEED
2100000e1ef99410 qlc                 1                  online  8Gb

Output for scripts

# fcinfo hba-port -O port_wwn,driver_name,link_failure_count,state,speed 2100000e1ef99410
2100000e1ef99410:qlc:1:online:8Gb