Zones jumpstart

is there a possibility to jumpstart zones? Say, I want to run a hosting with root access, Is it possible to create a zone, preconfigure it, install additional apps, and the to copy it? Is there a migration mechanism for zones? Can I save a zone on one machine (Flash archive maybe?) and then to restore it on the another one?

I don't think JumpStart would be applicable because Zones are so dependant on the underlying OE. But you could of course create zones in a jumpstart after script. Flash however would be closer... which you could similate by just tar'ing the zone root location, however again, most of the zone is made up of links and LOFI's back into the base OE, which would make transporting a zone from one location to another difficult.
And... I've been thinking about this too. At some point I started to wonder: Whats the point? Meaning, you can create zones damned near on the fly, and the zone is based on the underlying system which means you realy only need to backup user created content and logs. But that depends on how you use zones... I find my self wondering which is wiser: to install new software directly into a zone keeping it away from the global zone (base OE), or to install the software into the global zone but only use it in the non-global zones. In the first case you want to backup the whole enviroment, in the second you don't care.
I have a feeling when Solaris10 releases there is going to be one or several books reguarding Solaris Zones Best Practices. It's almost tempting to look at Zones like a sort of backwards provisioning service, where you can have a number of diffrent system images that can be created and destroyed at will in a commodity fashion.
Which in essence you can do, but that sort of thing would probly be better implemented using hardware partitioning and Flash images.
Just my take on it anyway.
benr.

Similar Messages

To disable a service for jumpstart/create zone

To disable a service once a system is up a running, "svcadm disable" does the job well.
What would the suggested way of disabling a service before the system is up an running, in order to guarantee a service never runs.
Clarification of this point would be from a finish script when a system is being built using a Jumpstart system (which might traditionally use something like rm /a/etc/rc3.d/S55thing), or when creating a zone (which might use rm /zoneroot/zonename/root/etc/rc3.d/S55thing), either of these not running in the true context of the end system.

I wanted to elaborate a little on Tobin's comments.
Most of the Solaris services come disabled by default, but many of them are also re-enabled by our default profile, generic_open.xml.
Brian, you're on the right track, though -- changing the manifest isn't the right way to disable a service. Manifests we deliver aren't editable -- they'll get overwritten, not preserved on upgrade. You also shouldn't modify the profiles that we ship, as they're also overwritten on upgrade.
So how do you do the disabling, when you don't have a running repository to modify?
The best way is to provide a site-specific profile. You can create your own /var/svc/profile/site.xml which will be applied to the system after the system-supplied profiles (during svc:/system/manifest-import). There you can describe services as enabled or disabled.
Just create /a/var/svc/profile/site.xml during your jumpstart finish script. See the other .xml files in /var/svc/profile for examples of the format.
You can also test your profile on a running system by using "svccfg apply".

Jumpstart installation from Solaris 10 zone

Hi,
I am using a Sol 10 zone as a jumpstart server, and have encountered a problem.
Due to not being able to use the out-of-the-box nfs server, as it does not run in a local zone on Sol 10, I have looked at using a user-space nfs server, unfsd.
unfsd runs fine, and can be used by a client to nfs-mount a directory, but when it comes to using the client's flash archive to build the client, I get this error below in my snoop output:
10:54:46.04754 js-client -> js-server MOUNT3 C Mount /flash_archives/TEST/js-client.flar
10:54:46.04779 js-server -> js-client MOUNT3 R Mount Not a directory
Which implies to me that the jumpstart client is trying to nfs-mount a file, and unfsd does nto allow this.
Any advice/suggestions on how to resolve this would be appreciated.
Richard.
Edited by: Richard on Jun 22, 2012 8:51 AM

Just to provide an update.
I switched to using ftp to get the flar file from the jumpstart server, which works fine.
I do believe that the issue with NFS is because the Solaris in-built NFS will allow a file to be NFS-mounted, but the user-space NFS daemon I used ( unfsd ) does not allow this, although I've not done sufficient snooping to prove this.
Richard.

Jumpstart zones and applications

hello
Is is possible to write some rules to create zones and install aplications into those zones when automating an install with jumpstart.
Alternativly can you create a systems image of several zones and install that zone image using jumpstart.
basically what i want to do is install solaris 10, create 2 zones and install oracle 10g into those zones.
regards
neville

From within the zone, you can see what pool you're bound to by simply using
the -q argument to poolbind(1M) with a valid pid, such as "poolbind -q $$".
Alternatively, you can use the pooladm(1M) command with no arguments.
Note that if you don't have pools active, this will result in a "Facility is not active"
message but otherwise you'll see the details about the pool this zone is bound
to.
From the global zone, you can see the actual pool the zone is currently bound
by doing something like "zlogin myzone 'poolbind -q $$'". And you can see
which pool the zone will attempt to bind to the next time it reboots by using
the "zonecfg -z myzone info pool" command.
Does this help?

Is it possible to setup Jumpstart server with zones

Hi ALL
is it possible to deployee images with solaris zones
if it is possible,how to deployee,can you plz proivde any docs
cheer's
Murali Muppa

Check out JetZONES, which is a module in the standard JET distribution. The documentation is included in the bundle as a PDF file. Should be just what you're looking for.

Unattended Solaris 10 installation with jumpstart server issue

Hello,
I have setup a jumpstart server and am finding it difficult to get a fully unattended solaris 10 sparc 10/09 install on a t1000 client. The console connection to the client goes into interactive mode at the point in system identification when the installer asks if the system is networked or non-networked, and runs a setup of default time zone and date/time. Using esc-2 to advance through the interactive menu it asks if ithe system is on a subnet, for the netmask (which is there as default), asks if the current time is correct, then asks me to confirm that my selections are correct. When I confirm it goes on it's merry way to finish the installation unattended.
My guess is it does not like one of the values in the sysidcfg file being used, but can't seem to see where the problem is. I don't see anything in the table of sysidcfg keywords from the online docs that specify networked or not, and I am using the network_interface keyword to define hostname, ip, default router, netmask and ipv6, and thought that should take care of the networking portion of system identification, also have timezone and time server set.
Could it be that I am configuring an ethernet port that is not connected (bge0)? This computer will be deployed on a DMZ type network, and the jumpstart server is located on a different network our internal private network. The jumpstart installation via the server is done by using a temporary ethernet connection on a second interface of the client. One other note the defined time server is on the dmz so the installation program won't see it.
Any suggestions on what might be causing the interactive installer to kick in?
Here is the sysidcfg used for the client:
more sysidcfg_vlan3/sysidcfg
keyboard=US-English
terminal=vt100
security_policy=NONE
system_locale=en_US
install_locale=en_US
timeserver=service-time
timezone=US/Eastern
network_interface=bge0 {hostname=D-Wade.xxx.xxxx.xxx default_route=xx.xx.xxx.1 ip_address=xx.xx.xxx..3 netmask=255.255.255.0 protocol_ipv6=no}
name_service=NONE
nfs4_domain=dynamic
root_password=someEncryptedValue
service_profile=limited_net
Thanks,
John
Edited by: jmck407 on Aug 6, 2010 11:42 AM

update. Fixed the problem with date/time by using abrante's suggestion of setting the time server to localhost. Don't get prompted for date/time information in the interactive installer any more.
If I setup the sysidcfg file to configure the ethernet port being used to communicate with the jumpstart server (bge1), using an ip and netmask that is on the jumpstart server network, the installation goes through fully un-attended.
Is there any way to get a second interface (in this case bge0) configured with ip/netmask values that are not part of the jumpstart server's network, using a non interactive jumpstart installation? Using these values in sysidcfg, the only things prompted for during the installation are, Is the system on a subnet, and what is the netmask. Anyone know what the sysidcfg part of the installation does to check the validity of a netmask value given in the sysidcfg file on the jumpstart server?
If need be, I guess the interface could be configured in the finish script.
more sysidcfg_vlan3/sysidcfg
keyboard=US-English
terminal=vt100
security_policy=NONE
system_locale=en_US
install_locale=en_US
timeserver=localhost
timezone=US/Eastern
network_interface=bge0 {hostname=D-Wade.xxx.xxxx.xxx default_route=xx.xx.xxx.1 ip_address=xx.xx.xxx..3 netmask=255.255.255.0 protocol_ipv6=no}
name_service=NONE
nfs4_domain=dynamic
root_password=someEncryptedValue
service_profile=limited_net

Solaris 10 Zone for Solaris 11.1 creation with template fails.

Hi,
We are trying to use the Oracle VM Template for Oracle Solaris 10 Zone to create a solaris 10 zone on a solaris 11.1 GZ
root@exsolh0005:/opt/scripts# ./solaris-10u11-sparc -v
This is an Oracle VM Template for Oracle Solaris Zones.
Copyright Â© 2011, Oracle and/or its affiliates. All rights reserved.
This software and related documentation are provided under a license
agreement containing restrictions on use and disclosure and are
protected by intellectual property laws. Except as expressly
permitted in your license agreement or allowed by law, you may not
use, copy, reproduce, translate, broadcast, modify, license,
transmit, distribute, exhibit, perform, publish, or display any
part, in any form, or by any means. Reverse engineering,
disassembly, or decompilation of this software, unless required by
law for interoperability, is prohibited.
Version: 1.0.9.12
root@exsolh0005:/opt/scripts# ./solaris-10u11-sparc -p /rpool -a 10.83.128.35/24 -i net6 -z testzone
This is an Oracle VM Template for Oracle Solaris Zones.
Copyright Â© 2011, Oracle and/or its affiliates. All rights reserved.
This software and related documentation are provided under a license
agreement containing restrictions on use and disclosure and are
protected by intellectual property laws. Except as expressly
permitted in your license agreement or allowed by law, you may not
use, copy, reproduce, translate, broadcast, modify, license,
transmit, distribute, exhibit, perform, publish, or display any
part, in any form, or by any means. Reverse engineering,
disassembly, or decompilation of this software, unless required by
law for interoperability, is prohibited.
IMAGE: ./solaris-10u11-sparc
ZONE: testzone
ZONEPATH: /rpool/testzone
VNIC: vnicZBI83273795
IP ADDR: 10.83.128.35
NETMASK: 255.255.255.0
DEFROUTER: 10.83.128.1
TIMEZONE: Europe/Copenhagen
Checking disk-space for extraction
Ok
Extracting in /opt/scripts/bootimage.8Ham6r ...
100% [===============================>]
Checking data integrity
Ok
Checking platform compatibility
The host and the image do not have the same Solaris release:
host Solaris release: 5.11
image Solaris release: 5.10
Will create a Solaris 10 branded zone.
Checking disk-space for installation
Ok
Installing in /rpool/testzone ...
100% [===============================>]
Attaching testzone
Problem attaching testzone.
More details in /var/sadm/install/logs/solaris-10u11-sparc.log
Cleaning up and exiting.
Friday, May 10, 2013 03:18:01 PM CEST: cleaning up
root@exsolh0005:/opt/scripts#
root@exsolh0005:/opt/scripts# cat /var/log/zones/zoneadm.20130510T131756Z.testzone.attach
[Friday, May 10, 2013 03:17:56 PM CEST] ==== Starting: /usr/lib/brand/solaris10/attach testzone /rpool/testzone -c /opt/scripts/bootimage.8Ham6r/sysidcfg.testzone ====
[Friday, May 10, 2013 03:17:56 PM CEST] Progress being logged to /var/log/zones/zoneadm.20130510T131756Z.testzone.attach
[Friday, May 10, 2013 03:17:56 PM CEST] Pinning datasets under rpool/testzone
[Friday, May 10, 2013 03:17:56 PM CEST] Pinning rpool/testzone
[Friday, May 10, 2013 03:17:56 PM CEST] Pinning rpool/testzone/rpool
[Friday, May 10, 2013 03:17:56 PM CEST] Pinning rpool/testzone/rpool/ROOT
[Friday, May 10, 2013 03:17:56 PM CEST] Pinning rpool/testzone/rpool/ROOT/zbe
[Friday, May 10, 2013 03:17:56 PM CEST] Log File: /var/log/zones/zoneadm.20130510T131756Z.testzone.attach
[Friday, May 10, 2013 03:17:57 PM CEST] Converting detached zone boot environment 'zbe'.
[Friday, May 10, 2013 03:17:57 PM CEST] Unmounting /rpool/testzone/root
[Friday, May 10, 2013 03:18:00 PM CEST] setting ZFS property zoned=on on rpool/testzone/rpool
[Friday, May 10, 2013 03:18:00 PM CEST] setting ZFS property canmount=noauto on rpool/testzone/rpool/ROOT
[Friday, May 10, 2013 03:18:00 PM CEST] setting ZFS property mountpoint=legacy on rpool/testzone/rpool/ROOT
[Friday, May 10, 2013 03:18:01 PM CEST] Mounting boot environment in rpool/testzone/rpool/ROOT/zbe-0 at /rpool/testzone/root (including child datasets)
cannot open 'ERROR: Error: Command <zfs list -H -o name -t filesystem -r rpool/testzone/rpool/ROOT/zbe-0> exited with status 1': invalid dataset name
[Friday, May 10, 2013 03:18:01 PM CEST] ERROR: refresh of ERROR: Error: Command <zfs list -H -o name -t filesystem -r rpool/testzone/rpool/ROOT/zbe-0> exited with status 1 failed
[Friday, May 10, 2013 03:18:01 PM CEST] ERROR: Error: rpool/testzone/rpool/ROOT/zbe-0: No such dataset.
[Friday, May 10, 2013 03:18:01 PM CEST] Unpinning datasets under rpool/testzone
[Friday, May 10, 2013 03:18:01 PM CEST] Unpinning rpool/testzone
[Friday, May 10, 2013 03:18:01 PM CEST] Unpinning rpool/testzone/rpool
[Friday, May 10, 2013 03:18:01 PM CEST] Unpinning rpool/testzone/rpool/ROOT
[Friday, May 10, 2013 03:18:01 PM CEST] Unpinning rpool/testzone/rpool/ROOT/zbe
[Friday, May 10, 2013 03:18:01 PM CEST] Result: Attach Failed.
[Friday, May 10, 2013 03:18:01 PM CEST] Exiting with exit code 254
[Friday, May 10, 2013 03:18:01 PM CEST] ==== Completed: /usr/lib/brand/solaris10/attach testzone /rpool/testzone -c /opt/scripts/bootimage.8Ham6r/sysidcfg.testzone ====
Looks like a script bug..
Regards
Claus

Hello All,
I would love to know if someone has a solution to this as I am seeing a similar issue with a slightly different setup.
Solaris 8 and Netra 240. I get the same result with and without patches and finish scripts. After the jumpstart and the failed boot attempt from disk. I can boot from cd, copy the glm driver from the cd to disk, and then everything works as it should. Obviously this is not a solution but only a work around.
Is it something to do with the SCSI hardware architecture used in the *240s?
Edited by: Gareth_Mann on Oct 2, 2007 7:41 PM

Non-global zone network configuration

Hi,
Zones are a new thing for me so please excuse me if this is a basic query... I have recently jumpstarted a system using a jumpstart script that was developed by somebody else. It creates two non-global zones and configures their network interfaces.
I have unplumbed one of the virtual interfaces for a particular zone because the IP address it was using is actually being used by another system on the network. However, when I reboot the zone, the interface is re-assigned the same IP address again. The IP address in question is not in /etc/hosts on any of the zones, and in the non-global zones the "hostname.<interface>" files do not exist at all. Also, the IP address is not in sysidcfg in any of the zones.
So basically, interface e1000g0:2 is being assigned an IP address that was configured by the jumpstart script, so perhaps the jumpstart script has placed that IP address in some file that is read when the zone is booting. I have even checked rc scripts just in case but I cannot find the IP address anywhere. Would anybody please be able to tell me where the configuration information could be coming from in this scenario (nsswitch.conf specifies only files).
Thank you in advance...

its in the zone config.
zonecfg -z <zone in question> info
it should list a net address and physical device. you can then use:
zonecfg -z <zone in question>
from here you can remove the net statements, or change the address if you want to keep using the net card in your zone.

Probem in creating Solaris 10 branded zone on Solaris 11.0

Hello,
I'm creating a zone on HP Blade with Solaris 11.0 using VM template Solaris 10u10 (x86).
I haven't /etc/defaultrouter file, my Solaris 11 is 11/11 and not Express version.
The error shows that variable used for temporary directory is lost. It's very strange because I completed 5 installations in the same way without any problem. But two last just are stopping with this error:
./setupSysidcfgStaticIP[50]: America/Montreal/sysidcfg.VL-VM-LD104: cannot create [No such file or directory]
cp: cannot access America/Montreal/sysidcfg.VL-VM-LD104
chmod: WARNING: can't access America/Montreal/sysidcfg.VL-VM-LD104
The "setupSysidcfgStaticIP" script is one of the many in the image which runs the creation of sysidcfg-file for the zone. It has 9 arguments to run:
CONT=$1
ZONEPATH=$2
NIC=$3
IP_ADDR=$4
NETMASK=$5
DEFROUT=$6
PASSWD=$7
TZONE=$8
TMPDIR=$9
The sysidcfg-filename is calculated as SYSIDCFG=$TMPDIR/sysidcfg.$CONT <- here is the error, the $TMPDIR fails
After that the script creates the file and renames it:
/usr/bin/rm -rf $SYSIDCFG
/usr/bin/sed -e s/RzoneR/$CONT/ $_tktHome/conf/sysidcfgStaticIP |\
/usr/bin/sed -e s/RnicR/$NIC/ |\
/usr/bin/sed -e s/RnetmaskR/$NETMASK/ |\
/usr/bin/sed -e s/RdefaultrouterR/$DEFROUT/ |\
/usr/bin/sed -e s/Rip_addrR/$IP_ADDR/ |\
/usr/bin/sed -e s]RtimezoneR]$TZONE] |\
/usr/bin/sed -e s]RpasswdR]$PASSWD] >> $SYSIDCFG
/usr/bin/cp $SYSIDCFG $ZONEPATH/root/etc/sysidcfg
/usr/bin/chmod 777 $SYSIDCFG
The tmpdir is created and used in main script but it fails in the subroutine. I don't understand why it was used before without any problem and now it fails. I spent yesterday about 6 hours to debug it but still don't have any idea how to fix it.
I use the image solaris-10u10-x86.bin .
I cannot use the newer image solaris-10u11-x86.bin (which works perfectly) because it isn't certified for our production and the set of servers needs to be installed with the same patch level.
Any help will be appreciated!
Vlad
Edited by: 919435 on May 24, 2013 5:56 AM

Problem found. To jumpstart last two HP Blades we used the copies of AI templates. When we did it from scratch and re-installed the Solaris 11.0 we have no more errors.

Trouble w/installing Solaris 10 branded zone on solaris 11.

Having issues creating a policy that works installing solaris 10 u10 branded nfs zone on sol11 in opscenter 12c u1. Maybe i'm just overlooking something basic or it isn't supported in opscenter. I'm able to create the policy but seem to get a very non-informative error message when deploying it.
Error Message:
"The DeploymentPlan execution job failed because the DeploymentProvider ZoneDeploymentProvider for Step Create Solaris Zones failed to generate tasks for the job: Cannot prepare zone tasks: java.lang.NullPointerException. Contact My Oracle Support if the problem persists. (10445)"
Here is the OpsCenter Profile:
Name Prefix:      hous
Starting Number:      1
Zone Description:      solaris 10 update 10
ZoneType:
Branded Zone
Branded Zone Image:
s10-update10-flar
Automatic Recovery:
Yes
Priority of Recovery:      0
CPU Shares:      1
CPU Cap:      0
Physical Memory Cap:      0
Locked Memory Cap:      0
Virtual Memory Cap:      0
Language:      en_US.ISO8859-15
Time Zone:      US/Central
Terminal Type:      xterm
NFSv4 Domain Name:      dynamic
Automatically boot zone when the global zone is booted:
Yes
Automatically boot zone after creation:
Yes
Storage for the metadata Library:      NAS, zone-prod1

Problem found. To jumpstart last two HP Blades we used the copies of AI templates. When we did it from scratch and re-installed the Solaris 11.0 we have no more errors.

Cannot create solaris 10 zone with flar archive

Hello,
i'm trying to configure a solaris zones with a flar archive of a solaris 10 system.
i've followed the procedure described here : http://www.sun.com/bigadmin/content/submitted/branded_zones_flash.jsp
whenever i try to create a zone and use: set brand=solaris10 the commit will fail with Unknown brand Missing document.
solaris version is Solaris 10 5/08 s10s_u5wos_10 SPARC and i have both the Solaris 8 and Solaris 9 container software installed.
is there another way to install a zone with the native brand or solaris 10 using an archive of some sort?
kind regards,

Thanks,
I've figured as much as using a whole root zone but i still struggle with the configuration of a solaris10 brand and the creation of a zone config with that brand.
I might be bettor of installing the zone without the flar image and try to copy the needed files and applications to the zone.
I only need the solaris10 brand zone to test the flar image so i can use it to restore the system with flar and jumpstart.

Solaris 10 8/07 on sparc jumpstart issue

hi,
I just tried to jumpstart via 'boot net:dhcp - install' the 8/07 release. No changes on my servers, only a link to the new repository. Though it throws me into the interactive install pretty early. I remeber from the earlier days that the 'question' that is asked at that point is a hint for the missing entry in the sysidcfg, doesn't seem to be this time:
─ The Solaris Installation Program ─────────────────────────────────────────────
The Solaris installation program is divided into a series of short sections
where you'll be prompted to provide information for the installation. At
the end of each section, you'll be able to change the-selections you've----
made before continuing.
About navigation... back to a confirmation screen you've already passed,
[ snip ]
This is my sysidcfg file:
keyboard=US-English
system_locale=en_US
timezone=MET
timeserver=localhost
terminal=vt100
root_password=MYROOTPWD
nfs4_domain=MY.DOMAIN.DE
network_interface=primary {dhcp protocol_ipv6=no}
name_service=DNS {domain_name=my.domain.de
name_server=my-ip-address-for-dns}
security_policy=kerberos {default_realm=MY>REALM
admin_server=MY.KERBEROS.SERVER
kdc=MY.KDC}
Any hints someone ???
cheers
~christoph

hi,
last thing I see is
'starting rpc...'
─ The Solaris Installation Program ─────────────────────────────────────────────
The Solaris installation program is divided into a series of short sections
where you'll be prompted to provide information for the installation. At
the end of each section, you'll be able to change the selections you've
made before continuing.
About navigation...
- The mouse cannot be used
- If your keyboard does not have function keys, or they do not
respond, press ESC; the legend at the bottom of the screen
will change to show the ESC keys to use for navigation.
────────────────────────────────────────────────────────────────────────────────
F2_Continue F6_Help
<F2>
─ Identify This System ─────────────────────────────────────────────────────────
On the next screens, you must identify this system as networked or
non-networked, and set the default time zone and date/time.
If this system is networked, the software will try to find the information
it needs to identify your system; you will be prompted to supply any
information it cannot find.
> To begin identifying this system, press F2.
────────────────────────────────────────────────────────────────────────────────
F2_Continue F6_Help
cheers
christoph

Zone create not installing all zone root files

I'm installing a new T5240 Solaris 10 u6 10/08 with the 10_Recommended patch kit from 09 Dec 2008. The system is jumpstarted with /, /usr, /var, /export/home on SVM mirrored internal 146g drives. When I first got the server up and running, I tested zone creation with a default zone setup:
zonecfg -z test
create -b
set zonepath=/export/home/zones
set autoboot=true
add net
set physical=nxge0
set address=192.168.103.254
end
verify
commit
exit
zoneadm -z test install
Preparing to install zone <test>.
Creating list of files to copy from the global zone.
Copying <139100> files to the zone.
Initializing zone product registry.
Determining zone package initialization order.
Preparing to initialize <1164> packages on the zone.
This worked, so I removed it and continued with the server configuration:
The nxge0 interface has been configured with 802.1q vlan (nxge103000) and is successfully on the network. I configured MPX/IO on the pair of fp interfaces. I've added a disk from our HP EVA SAN array and created a ZFS pool and filesystem for the zoneroot, mounted at /opt/zones/zoneroots/oraprod1.
Now, when I try to recreate the test zone it now fails to copy all of the required files to the root, puts many errors into the install log about files changed in the global-zone root after installation and won't boot the zone past single-user maintenance mode.
The zone create now shows: (after adjusting the phys net i/f name)
bash-3.00# zoneadm -z test2 install
Preparing to install zone <test2>.
Creating list of files to copy from the global zone.
Copying <78> files to the zone.
Initializing zone product registry.
Determining zone package initialization order.
Preparing to initialize <1165> packages on the zone.
When I zlogin to the zone after it's booted into maint, it's obvious that many root fs files are missing - including
most of the SMF service manifests.
I've searched for forums, mailing lists and sun support knowledgebase without finding anything similar.
What am I missing?

Yes, I've reproduced the problem with and without the '-b' option on create, on UFS as well as ZFS filesystem and on /export/home/zones on the local drive.
My plan now is to walk thru the configuration process one step at a time, trying to create the zone at each step. I re-jumped the server last night and created the zone correctly. I then removed it, installed patches in single user, rebooted with "-r" and was able to recreate the zone. I've enabled mpxio and i can still create the zone. Next up tonight is to add the 802.1Q configuration, add the ZFS rootzone and run our custom scripts that to server hardening, user setup, etc.
There was one bug report that sounded "interesting" having to do with a patch set that wasn't zone-friendly as it left patch files in place w/o fixing the manifests so when the zone create ran it found files that didn't match the checksums or whatever is used to validate file correctness when copying files into the zone.
I'll keep updating this with my progress and findings.

Pkgmap files missing in global zone, can't build non-global zone

My solaris 10 server is missing the pkgmap files for the packages. As a result, I can't build a non-global zone. Is there a way to recreate the pkgmap files?
The OS on the Solaris 10 server was installed via jumpstart (initial install). However, the Jumpstart process used a Solaris 9 boot server which seems to have caused the missing pkgmap problem.
Does anyone know of any other problems which would result from a version mismatch between a boot and installation server during the jumpstart process?

Hi, i have problems with building transmission from svn too:
$ versionpkg
==> retrieving latest revision number from svn... 3730
==> newer revision detected: 3730
==> Entering fakeroot environment
==> Making package: transmission-svn 3730-1 (Di 6. Nov 08:28:38 CET 2007)
==> Checking Runtime Dependencies...
==> Checking Buildtime Dependencies...
==> Retrieving Sources...
==> Validating source files with md5sums
==> Extracting Sources...
==> Removing existing pkg/ directory...
==> Starting build()...
Fetching external item into 'Transmission/third-party/libevent'
Checked out external at revision 477.
Checked out revision 3730.
==> SVN checkout done or server timeout
==> Starting make...
./autogen.sh: line 16: autoreconf: command not found
Creating aclocal.m4 ...
Running glib-gettextize... Ignore non-fatal messages.
Copying file mkinstalldirs
Copying file po/Makefile.in.in
Please add the files
codeset.m4 gettext.m4 glibc21.m4 iconv.m4 isc-posix.m4 lcmessage.m4
progtest.m4
from the /aclocal directory to your autoconf macro directory
or directly to your aclocal.m4 file.
You will also need config.guess and config.sub, which you can get from
ftp://ftp.gnu.org/pub/gnu/config/.
Making aclocal.m4 writable ...
Running intltoolize...
PKGBUILD: line 33: ./configure: No such file or directory
make: *** No targets specified and no makefile found. Stop.
==> ERROR: Build Failed. Aborting...
==> ERROR: Reverting pkgver...
i dont know whats up with the autoreconf
i hope anyone can help me!
greez

Interactive service at first boot after jumpstart

I have a lot of issues/questions with smf, so this posting is abit long. Sorry for that.
So far we have installed our own software by jumpstarting entire Solaris, copying an rc script to rc2.d , rebooting, then have that rc script to the rest of the installation. This was necessary as some important parts of the installation required daemons that are installed and started at the early stages of that rc script. It cannot be done from the finish script with everything mounted under /a. And, to make things more complex, at the end of the installation I have some interaction prompting certain passwords.
This whole mechanism blew up with the introduction of smf. So I started reading about smf and am trying to convert all this to a service.
So far I have found that I need to setup a transient single instance service, with a :true stop method, my initial rc script in /lib/svc/method/oluinstall as the start method, the service enabled by default and put the manifest in /var/svc/manifest/site/oluinstall.xml. I have a manifest that I believe to be OK. All files in /var/svc and /lib/svc would be copied there by the finish script.
What is not yet clear to me is:
a) how to get all the output to be displayed on the console instead of (or in addition to) redirecting it all to a smf logfile and
b) how to become interactive, despite this stronlgy being discouraged?
Are any of these possible? Or is there another solution to my initial installation problem?
I haven't had a chance to test my manifest as I have hit another problem that is mentioned in the Beta 7 Bugs of the Release Notes: autofs does not start (5084183). I have adjusted my finish script to create /a/etc/defaultdomain to make keyserv happy, but autofs does still not come up. svcs reports various services as offline or unitialized and the system does not get to the multi-user-server milestone:
offline        15:41:34 svc:/network/inetd-upgrade:default
offline        15:41:34 svc:/network/inetd:default
offline        15:41:34 svc:/milestone/multi-user-server:default
offline        15:41:36 svc:/network/nfs/client:default
offline        15:41:36 svc:/system/filesystem/autofs:default
offline        15:41:36 svc:/system/dumpadm:default
offline        15:41:37 svc:/system/fmd:default
offline        15:41:38 svc:/application/print/ipp-listener:default
offline        15:41:39 svc:/system/zones:default
offline        15:41:41 svc:/network/nfs/server:default
offline        15:41:44 svc:/network/smtp:sendmail
uninitialized 15:41:36 svc:/network/rpc/gss:ticotsord
uninitialized 15:41:37 svc:/network/security/ktkt_warn:ticotsord
uninitialized 15:41:37 svc:/network/telnet:default
uninitialized 15:41:37 svc:/network/rpc/smserver:default
uninitialized 15:41:40 svc:/application/print/rfc1179:default
uninitialized 15:41:40 svc:/network/nfs/rquota:ticlts
uninitialized 15:41:41 svc:/network/nfs/rquota:udp
uninitialized 15:41:41 svc:/network/ftp:default
uninitialized 15:41:42 svc:/network/finger:default
uninitialized 15:41:42 svc:/network/login:rlogin
uninitialized 15:41:42 svc:/network/rpc/rstat:udp
uninitialized 15:41:43 svc:/network/rpc/rusers:udp
uninitialized 15:41:43 svc:/network/shell:tcp
uninitialized 15:41:43 svc:/network/shell:tcp6onlyThis impacts my oluinstall script as it depends on automounter running. I tried a lot of things but I don't know how to have the autofs service up by default after a fresh jumpstart with no NIS/NIS+ configured and an exsiting /etc/defaultdomain. Can you help here?
I also tried to validate my manifest with svccfg validate and all I got was "couldn't parse document". Not very helpful. I used a toll called "rxp" that I got fro ma colleague to help find me the problems. When I fixed that I now get the following:
# svccfg validate oluinstall.xml
oluinstall.xml:17: element service: validity error : Element service content does not follow the DTD, expecting (create_default_instance? ,
single_instance? , restarter? , dependency* , dependent* , method_context? , exec_method* , property_group* , instance* , stability? ,
template?), got (create_default_instance single_instance dependency dependent exec_method method_context exec_method property_group stability
template )
svccfg: Document is not valid.
Well, I guess I need to include my manifest now as well ...
<?xml version="1.0"?>
<!DOCTYPE service_bundle SYSTEM "/usr/share/lib/xml/dtd/service_bundle.dtd.1">

<service_bundle type='manifest' name='ubsjpstrt:oluinstall'>
<service
     name='site/oluinstall'
     type='service'
     version='1'>
     <create_default_instance enabled='true' />
     <single_instance />
     <dependency
          name='oluinstall_multi-user'
          type='service'
          grouping='require_all'
          restart_on='none'>
          <service_fmri value='svc:/milestone/multi-user' />
     </dependency>
     <dependent
          name='oluinstall_multi-user-server'
          type='service'
          grouping='require_all'
          restart_on='none'>
          <service_fmri value='svc:/milestone/multi-user-server' />
     </dependent>
     <exec_method
          type='method'
          name='start'
          exec='/usr/bin/rcwrapper /lib/svc/method/oluinstall'
          <method_context>
               <method_credential user='root' group='root' />
          </method_context>
          timeout_seconds='0' />
     <exec_method
          type='method'
          name='stop'
          exec=':true'
          timeout_seconds='3' />
     <property_group name='startd' type='framework'>
          <propval name='duration' type='astring' value='transient' />
     </property_group>
     <stability value='Unstable' />
     <template>
          <common_name>
               <loctext xml:lang='C'>
                    OLU Base jumpstart installation
               </loctext>
          </common_name>
          <documentation>
               <manpage title='oluinst' section='1M'
               manpath='/usr/share/man' />
          </documentation>
     </template>
</service>
</service_bundle>Finally, at the end of my install script, I need to remove the service as I need it to run only once. Before smf the rc script simply removed itself. How do I do this now?
svcadm disable svc:/site/oluinstall
svccfg delete svc:/site/oluinstall
Is that correct?
Thanks for any help!
- Tom

Tom:
Great questions. I'll try to answer them in order.
What is not yet clear to me is:
a) how to get all the output to be displayed on the console instead of
(or in addition to) redirecting it all to a smf logfile and
b) how to become interactive, despite this stronlgy being discouraged?Becoming a console interactive service requires two parts:
- You need to redirect your input and output to /dev/console. Example
services which do this are kdmconfig (x86 platforms) and the
sysidtools (all platforms)--that is, your fellow interactive services.
Please attempt to keep your error messages going to the logfile by
making the redirection only for specific commands, and not the entire
start method.
- More importantly, you also have to insert yourself before any of the
console-occupying login services and after the last of the configuration
services. On workstations, this set of services includes GDM (for
GNOME/JDS) and possibly CDE, depending on build.
(We're going to make this a little easier and then much easier in the
future. But you should only write interactive services if you are the
systems deployer, and not a software vendor.)
You can use svcs -l to list the dependencies on console-login, like so
$ svcs -l console-login
fmri         svc:/system/console-login:default
name         euro-friendly console login
enabled      true
state        online
next_state   none
restarter    svc:/system/svc/restarter:default
contract_id 133
dependency   optional_all/none svc:/platform/i86pc/kdmconfig:default (online)
dependency   exclude_all/none svc:/application/gdm-login:default (disabled)
dependency   require_all/none svc:/system/sysidtool:net (online) svc:/system/sysidtool:system (online)
dependency   require_all/none svc:/system/utmp:default (online)
dependency   require_all/none svc:/system/identity:node (online)
dependency   require_all/none svc:/system/filesystem/local (online)
I haven't had a chance to test my manifest as I have hit another
problem that is mentioned in the Beta 7 Bugs of the Release Notes:
autofs does not start (5084183). I have adjusted my finish script to
create /a/etc/defaultdomain to make keyserv happy, but autofs does
still not come up.I would actually disable keyserv, as what's happening is nfs/client is
probably enabled and getting stuck. The best way to report on this is
to run "svcs -xv".
I also tried to validate my manifest with svccfg validate and all I
got was "couldn't parse document". Not very helpful. I used a toll
called "rxp" that I got fro ma colleague to help find me the problems.
When I fixed that I now get the following:
# svccfg validate oluinstall.xml
oluinstall.xml:17: element service: validity error : Element service
content does not follow the DTD, expecting (create_default_instance? ,
single_instance? , restarter? , dependency* , dependent* ,
method_context? , exec_method* , property_group* , instance* ,
stability? , template?), got (create_default_instance single_instance
dependency dependent exec_method method_context exec_method
property_group stability template ) svccfg: Document is not valid. We ship xmllint(1), but I agree that svccfg validate should be more
helpful. Anyways, the point is that your method_context element needs to
precede your exec_method elements. If you want to override a single
method's context, you embed the method_context element.
/var/svc/manifest/network/ldap/client.xml has a non-trivial example.
Finally, at the end of my install script, I need to remove the service
as I need it to run only once. Before smf the rc script simply removed
itself. How do I do this now?
svcadm disable svc:/site/oluinstall
svccfg delete svc:/site/oluinstallWe have use a similar approach for singly-run upgrade services, like
network/inetd-upgrade. I discourage deleting the service (and deletion
after disable is a race between the restarter and the repository anyway).
I hope these answers are helpful.
- Stephen

Zones jumpstart

Similar Messages

Maybe you are looking for