Solaris 10 Containers and OS Compatibility

I was wondering what OS's can be run as of today in a Sparc based Solaris 10 container? Can you run instances of Linux, and previous versions of Solaris such as 8 and 9?

Zones do not "boot" another OS. They instantiate an application environment that is a sort of clone of the global zone. You'll gain a better understanding of why this isn't really the right question to ask if you review the Solaris Zones paper we published at LISA '04. The paper is at http://blogs.sun.com/roller/page/dp/20041120#lisa_04_solaris_zones_operating
Hope this helps. As for your Solaris 9 apps-- Solaris 10 is guaranteed to be binary compatible with Solaris 9. Your apps should migrate and run without a problem, and without needing to be recompiled.

Similar Messages

Solaris 8 Containers and x86 compatability

My client relies on a program for the meat and potatoes of our processing. This program currently only works on Solaris 8. We are using Sun Blade 2500 workstations. However we are moving offices and to build up the new lab, the client wants to get away from the Sun workstations and move towards Linux or Windows machines. He is however willing to migrate to Solaris 10 provided the the program that we depend on is still usable. Enter my solution: Solaris 8 containers.
The documentation for containers was not very clear. I am interpreting two different messages from the Sun website. One says that Solaris8 Containers will work on SPARC systems and/or Systems running Solaris 10 10/08 or later. Other documentation says Solaris Containers are SPARC system dependent.
I installed Solaris 10 on a HP x86 based desktop via VirtualBox. I installed all the SUNWs8brandX patches that came with the containers bundle and followed the install setup directions. I cannot verify as I am getting a "Cannot Execute Brand Specific Error" message.
Does anyone else have a suggestion as to how I would get a Solaris 8 dependent program running on a Linux or Windows based machine? Is my solution viable?

agquint wrote:
What I mean was the post below your first one. But you correctly interpreted it in your next post. So "Solaris 8 containers" only works for SPARC. I don't get why Sun would only have containers support for SPARC systems. That is in total contradiction to why I would think they would develop containers in the first place: backwards compatibility to support programs that aren't yet migrated to the current version of Solaris regardless of whether they are x86 or SPARC. They have a solution but it only works at %50? hmph.There is probably just very little demand for it. During the Solaris8/Solaris9 days, Sun was killing off the x86 line. There is very little installed base of Solaris 8 x86 compared to Solaris 8 SPARC. (It wouldn't help you out, it appears, since you have SPARC binaries).
Darren

Solaris Express 11 and Mac compatibility

Hi all,
I would want to use Solaris with my Mac so I tried to boot on the DVD and install the Developer version, the Mac boot on it correctly and started to install Solaris. But I have a problem with the screen (APPLE 23' with a NVidia 7300GT) when it wants to switch in the window mode, my screen freeze and the system returns to the console mode...
I tried with another screen (samsung 17') and the installation in a window mode worked !
So the problem is the compatibility with the screeen or my graphic card ? (it seems that the 7300GT is not yet supported, it doesn't appear in the listed supported devices report)
how can I do to work with my Apple screen with the high resolution (1920*1200) ?
thanks a lot...

Cindy,
Thanks for the quick reply. If I'm understanding correctly, the SCSI unmap command was implemented in last November's Solaris 11.1 release but has since been disabled due to performance problems. I'm running the original November release with no updates so can I assume that SCSI unmap is enabled on my system? If so, do I run any risk of data corruption by staying with my current version? I'm currently running Solaris 11.1 from a SATA SSD.

Solaris containers and Fiber connection

Hi guys,
I have sucsessfully migrated solaris 8 system to solaris 10 containers. and also mounted some RAID enabled partition on non global zone without any issue. now i need to connect a fiber channel to my host OS (solaris 10) and add the SAN partition to my container (solaris 8). if i run cfgadm command it says cfgadm: Library error: Device library initialize failed: Invalid argument.
there are some other ways to mount the SAN (mount the SAN on host and export the partition to container). but i need the above method . is it possible to mount the SAN partition from the non-global zone ? if so how should i do it .
thanks,
Tharanga

As far as i know you can't give the non-global zones the required permissions which is needed to mount a FC LUN directly into the non-global zone.
I think you have to mount the LUN in the global zones, and then loop-back-mount them into the non-global zone.
.7/M.

Grid 11gR2 (11.2.0.2) on Solaris 10 Containers getting PRVF-5150 error

I am attempting to install Oracle DB RAC Grid Infrastructure 11gR2 on Solaris 10 Containers and am getting ASM PRVF-5150 error. After doing some search internally/externally, I found something like: .." the prerequisite check for Oracle ASM devices reports an error (with PRVF-5150) that the given disk is not a valid path. Workaround: This error can be ignored."
I have 2 T2000's, Oracle Solaris 10 9/10 s10s_u9wos_14a, 144488-12, Solaris Whole-Root Zone, and Sun StorEdge 3320 (SCSI shared disk array). I have installed Grid 11.2.0.1 in the same environment and I didn't get PRVF-5150 error. Attempting to install 11.2.0.2 would consistently produce the error.
My shared device can be seen from both nodes (and as mentioned 11gR1 was successfully installed/configured/deconfigured). cluvfy.sh also passed.
crw-rw---- 1 grid asmadmin 32, 4 Jul 18 14:58 /dev/rdsk/c2t0d0s4
If I were to ignore it and continue, I get the following error:
ACFS-9200: Supported
ACFS-9300: ADVM/ACFS distribution files found.
ACFS-9307: Installing requested ADVM/ACFS software.
ACFS-9349: Unable to configure driver 'oracleadvm'.
ACFS-9310: ADVM/ACFS installation failed.
USM driver install actions failed
Oracle DB and RAC 11gR2 has been certified on Solaris 10 Containers.
Any suggestion is appreciated. Thanks in advance.

Hi,
You can use this tech note on MOS.
*Troubleshooting 11.2 Grid Infastructure Installation Root.sh Issues [ID 1053970.1]*
Put here alertlog file of grid
Please use formatting options using tags output between tags[/code ]
Regards,
Levi Pereira
Edited by: Levi Pereira on Jul 19, 2011 12:24 PM

SUN Solaris: Containers and Virtualization

Hi Everyone,
Just starting on a new project which will deployed on SUN Mxxxx servers.
We have opted to virtualize the systems, does anyone have any experience on Solaris Virtualization, containers, etc?
Any white papers, how to guide or experience would be greatly appreciated.
Kind Regards
Michael

We make use of containers/zones heavily (together with ZFS as filesystem) and it works like a charm. We also use on one box Solaris Cluster together with containers.
Check the following notes:
Note 1246022 - Support for SAP applications in Solaris Zones
Note 870652 - Installation of SAP in a Solaris Zone
Note 724713 - parameter settings for Solaris 10
What database will be used?
Markus

Compatibility of Solaris WBEM servises and SBLIM wbemcli utility

There is a strange situation, when Solaris wbem server is running on port *:5988, access to CIM server are granted for users, firewall is disabled, however there is no possibility to connect solaris CIM server from linux hosts using wbemcli.
So the question is: Is there a compatibility between Solaris WBEM realization and SBLIM realization?

From what I have read Sun Cluster 3.2 will not be supported with RAC 11g R2 untill September time frame.
The functionality will be released via a cluster core patch.
Currently Sun Cluster is looking down in /etc/init.d/init.crs to start CRS (11g R1). With 11g R2 the start script for GRID is now /etc/init.d/ohasd.
SC does support HA 11g R2, but not RAC at this point in time.

10g RAC on Solaris 10 containers - Question on Approach

Hi All
To introduce a little, I am trying to set-up 10g RAC (active-active) on Solaris 10 Containers. For the disaster recovery site, I was considering an option for storage replication across datacenters. As I look at the situation where the database could be highly transactional, I wanted to go for this option hoping I might be better able to recover db on dr site if a failure happens and also alleviate some load from the prod servers. In this regard I have a set of questions for which if you could share your experiences, thoughts .. I am sure will help my preperation.
1 Is Oracle Supported on Solaris 10 containers? I believe it is not supported and ofcourse am getting into the certification matrix for this with oracle. Anyways, any work around tried out by anyone ?
2 Which product is better in terms of performance while co-existing and co-functioning with Oracle? Sun Cluster!!! Vertas Cluster!!! any other? I am unable to link to any recommendations from Oracle and hence trying to understand the generic or say standard approach to this.
3 I understand on Prod site, Oracle remains active-active on two node cluster while I will mirror my SAN between Prod site and DR site on a timely basis, while DB is up and running. This mirroring has to be only an asynchronous mirroring for some limitations. I am unsure about my recoverability if we need to switch over to DR. Will I be able to recover the Database to a point when my last mirroring has been completed successfully when I say the mirroring happened while the Prod db is up and running (hot backup type on disk level). Until now I used a cold backup for Storage replication and I was sure to bring it up there. This hot backup kind of attempt is raising unrest and questions about the abilities.. :-(
4 I also have come across a term called federated databases? Is it similar to what I am trying to achieve here? Storage level replocation across data centres?
I know it takes some time to get to the end of it and finally implement it, but your inputs will certainly help me get things straight before I hit this first click. Thanks for all your inputs, advices, suggestions. :-)
Regards!
Sarat

According to Oracle "Both Oracle 9i R2 and 10g R2 databases have been certified to run in a Solaris Container."
There are references here :- http://forums.sun.com/thread.jspa?threadID=5397695
"Oracle databases (RAC and non-RAC) have been certified to run in a global Solaris 10 Container"
What storage are you using ? (filesystem, ASM ?) Personally I prefer ASM now, but it does depend on the storage platform and backup strategy.
What backup infrastructure do you have ? RMAN backup directly to tape is better than file copy.
For DR standby purposes I would use Dataguard instead of disk mirroring / storage replication. It is much more flexible (e.g can switch direction of replication, do incremental backups from the standby etc).
I would use CRS instead of VCS. CRS is built for the job, and integrates with netca, dbca, etc for service management, and also integrates with Dataguard Broker for standby switchover.

Different X behaviour between Solaris 2.6 and 8

Hi,
We have recently upgraded a development box to Solaris 8 from 2.6. We have recompiled (same compiler) an application with no changes... and we get different behaviour. There are two problems that we are tracking so far:
1. Buttons don't seem to be responding to their callbacks so we never leave our modal event loops.
2. It looks like we're getting truckloads of X expose events causing lost of flicker on every dialog. The CPU is maxing out the whole time, redrawing windows takes forever after hiding it, etc. It seems to be mainly text boxes, but I can't be sure.
We've copied a binary compiled on 2.6 onto the box and it runs fine so it seems. Is anyone aware of any issues/patches around this area, or any other info we can get to assist tracking down where the exposes are coming from.
Thanks,
Rhys

Progress!
If I compile with -DMOTIF12_HEADERS and then link with -lMrm12 (which resolves to libMrm.so.3 instead of libMrm.so.4 )and -lXm12 (which resolves to libXm.so.3 instead of libXm.so.4) my application behaves as expected.
So it would seem that something has changed in Motif that my application does not like, or we were aleays doing something wrong and it just got worse, or... any ideas on where I can start hunting for such a beast?
Thanks,
Rhys
Moazam,
Just rebuilt using 6U2 both with and without -compat=4
and it still exhibits the same behaviour... lots of
expose events, flicker all over the place, high CPU
usage (its around 60% and Xsun is around 30%).
Any further ideas?
Rhys
To my knowledge, 4.2 is not supported on Solaris 8.
You should try compiling your code with Forte
Developer 6.2, and if you need, you can use the
-compat flags.
-Moazam

Solaris 10 with Solaris 9 containers

A customer has Solaris 10 servers. They are interested in making solaris 9 containers on this system. Is that doable? I thought you could only build solaris 9 containers on Solaris 9 systems....

Solaris 10 runs both Solaris 8 and 9 Container software. yust download the correct software from : http://www.sun.com/software/solaris/containers/getit.jsp
There is lots of documentation on how to configure but its no different then a native zones/container.
You can even create a Solaris 10 zone (not the same as native) using the solaris 9 container as a template.
I've used it very succesfully to migrate a number of old solaris 8 and 9 servers and had no problems what so ever.

Migrating Legacy Applications to Solaris 10 Containers

Hi, Good Day,
We have some customers who are keen to migrate their legacy banking systems to run on Solaris 10 containers. We intend to assess the feasibility of the project.
Kindly advice the considerations/potential pitfalls(if any) of running legacy Solaris 8 applications on Solaris 10 containers. Kindly advice those types of applications that cannot run on Solaris 10 containers?
What kind of information should we extract from the customer to understand these considerations?
Thank-you and Best Regards,
Siong Guan

Oracle migration workbench solve only the technical point of the migration. That won't solve the data mapping you have to define on your own.
Data mapping is the most important task of a migration, totally depending of the source and target systems. You need expertise on both of them and discuss with the business and functional experts people.
Then, and only after defining the data mapping, you can go for the technical migration.
Nicolas.

Beginner setting up solaris 10 container and having network card problems

Hi I am just learning about solaris containers and am having an issue of getting my new zone to be seen on the network.
I have installed the global zone server name GPSAPP02 network address of 192.168.200.45 works fine.
I created :-
a pool called "ellipse-pool" using pooladm/poolcfg
a zone called "ellipse-zone" using zoneadm/zonecfg
I used the followijng commands...
zoneadm list -cv
zonecfg -z ellipse-zone
create               --
set zonepath=/zfspool/zones/ellipse-zone
set autoboot=true
     add net
set address=192.168.200.46/24
set physical=eri0
set defrouter=192.168.200.254
end
set pool=ellipse-pool
verify
commit
exit
The output from the "zonecfg -z ellipse-zone info" command is below:-
zonename: ellipse-zone
zonepath: /zfspool/zones/ellipse-zone
brand: native
autoboot: true
bootargs:
pool: ellipse-pool
limitpriv:
scheduling-class:
ip-type: shared
inherit-pkg-dir:
dir: /lib
inherit-pkg-dir:
dir: /platform
inherit-pkg-dir:
dir: /sbin
inherit-pkg-dir:
dir: /usr
net:
address: 192.168.200.46/24
physical: eri0
defrouter not specified
I then installed the zone and rebooted using zoneadm.
I can login to the zone and setup the timezone and terminal etc.
I gave this a hostname of GPSEL01 as this is what I want it to be known as on the network.
So far so good but from the network I cannot ping the IP address I assigned to the zone 192.168.200.46
What have I missed or what should I check to troubleshoot this?
I ran a ifconfig -a and received the following information...
ifconfig -a
lo0:1: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1
inet 127.0.0.1 netmask ff000000
# ifconfig lo0:1 192.168.200.46 netmask 255.255.255.0 up
ifconfig: SIOCSLIFNETMASK: lo0:1: permission denied

Hi
I notice in your output from ifconfig -a in the global zone that the NIC card is nge1.
However in your zoneconfig you have set the NIC card to eri0 (physical = eri0). I cannot see this NIC card in your ifconfig -a.
When a zones is of IP-type shared-ip it needs to use a network interface that is already plumbed into the global zone. In your case you need to change the NIC in your zone config to nge1 :
zonecfg -z ellipse-zone
select net address=192.168.200.46
set physical=nge1
end
exit
Then reboot your zone.
A couple of other minor comments.
1) As it appears the zone is on the same subnet as the global zone, if the global zone already has the default route of 192.168.200.254 setup you don't need to specify it in the zone config. You are most likely to use the defrouter parameter if the zone was on a different subnet to the global zone (but shared the same physical NIC).
2) You have specified the netmask in the address field which is fine. But again this should not be necessary as long as your /etc/netmasks file is correct in the global zone.
Hope this helps

I have a serious (and bizarre!) issue with my novation impulse (Although i've tried it with two other keyboards and i still have the same problem) and its compatibility with mainstage 3

i have a serious (and bizarre!) issue with my novation impulse (Although i've tried it with two other keyboards and i still have the same problem) and its compatibility with mainstage 3.
the problem is best explained on the following one - page thread:
https://discussions.apple.com/thread/3951518?start=0&tstart=0
(Clearly i'm not alone in this problem, although i think i figured out what's going wrong a little more than he did...read on!)
his solution, to put mainstage in jump mode, is very unsatisfactory to me, as it bounces all of a sudden to drastically different settings.
basically, my analysis is that my controller is NOT receiving MIDI date from mainstage. in other words, mainstage knows what my controller is doing, but my controller doens't know what mainstage is doing.
let's say i turn the knob all the way to the right ... 127...and the virtual fader goes to the right like it's supposed to.
now...next...let's say i change to a different patch, where that same VIRTUAL fader is not at the max clockwise position..maybe it's only at 1pm. now when i turn the physical knob to the RIGHT, the midi data is still at 127 on the controller! it didn't "reset" to sync up with the new level (say 80 or so) setting on the new patch. so i can't increase that new setting of 80 by continuing to turn the knob to the right. i have to turn it all the way to zero,...and then continue PAST zero until the controller thinks that IT is at 0...at that point the controller and mainstage are in agreement, and things work fine....so bascially, the keyboard thinks the level is at max...but mainstage thinks the level is at 1pm.
i am using Logic 9, and i have a macbook pro 2.9 Ghz I7 with 8 gigs of memory and OS X 10.8.4

Hi Josh,
Thanks for taking the time to contact us here a Novation for technical support. Lets continue to correspond via email so we can get your issue resolved.
Thanks.
Mike Towns

Should I use Standard C++ Containers and Stuff for games and other real time activities?

Hello,
I'm a C++ developer using VS 2012 and VS 2010 for developing AAA titles. I have read about not using STL and other stuff provided in the standard headers that come with VS. I read most of the stuff on the websites based on game programming and some
are really from the people well known in the industry. I have seen cases where they wont even use vector, list, map and others and not even use utility functions and algorithms. In such cases they write those containers and stuff themselves which has almost
the same interface and so much of debug and implementation time spent on such huge code.
I have two questions:
1: Isn't the C++ implementation that comes with VS optimized for the platform for better performance? Isn't it using some intrinsic functions that people on the client end doesn't know about and supplying their own implementation would indeed be
more slower in basic container operations such as insert, remove, find, swap, copy? Lets assume that we supply our own custom allocators for faster memory management to every container that we use. Also, they take care of fragmentation, alignment and stuff.
Why develop custom containers with almost same interface, why not spend that time on writing allocators and other stuff that might actually help?
2: There are times when we include a lot, a lot, of unnecessary stuff through the standard headers in a huge code base. Unnecessary, because we only needed a thing or two from such huge headers including other huge headers and so on. Now, I know templates
aren't instantiated unless used, same goes for the members functions inside them and blah blah. Since, these are precompiled headers, it is safe to assume that there is no compile time hit for that unnecessary stuff. My question is, is there any hidden effect
of such inclusions on code size (executable) that grows with the huge growth of the code base? In my opinion there shouldn't be, but I wanna know I'm not mistaken just in case.

Hi
I can tell you a story about one of our projects. We had to implement a block management system for harddisks (like the parts of file systems that allocate, free blocks and do the bookkeeping stuff). Now, one guy was convinced, that we should use the STL,
because it's easy to use and very well optimized and things like that. An other one said, that he would implement it by hand. So we decided to do both for a little test. The result was, that the hand implemented thing was much easier to write and much faster...
and the one using STL saw, that some classes were so slow, that he couldn't believe it... he then tryed to build some classes by hand and exchange them (e.g. array classes and stuff like that) and it was way faster than the STL and he wrote them in minutes
(!)... since then nobody wanted to use STL again... we don't even try it. It ended that we wrote our own string and array classes and stuff like that... easy little things that we understand. Now it's easy to use and fast and still flexible, because we know
what it's doing...
... but, that's just a story about our development and of course, we're often programming very close to the hardware and also using assembler languages often... so, I don't know what this tells us... you have to decide :-)
Rudolf

How can I get the Microsoft enhanced Point and Print compatibility driver installed on Win7 LC master?

We recently migrated from a Windows Server 2003 to Windows Server 2012 print server. When recomposing a linked clone pool with the old 2003 server, drivers were taken care of for the printer when the user printed for the first time after desktop recompose. With 2012, any printer using the Point and Print driver doesn't seem to obtain the driver again from the 2012 server. All of our clients are running Windows 7 which means the Microsoft enhanced Point and Print compatibility driver is not installed in the OS natively.
I apparently need to figure out a way to get this installed on the host image so it trickles down to snapshots when recomposing the linked clone pool but I haven't been able to dig much information on doing this. Does anyone out there have any insight on how to accomplish this? If I recompose a pool now for a user who is using a new Point and Print driver, they get the following message that is resolved by removing the printer from their account and re-adding it:
"The 'Microsoft enhanced Point and Print compatibility driver' printer driver is not installed on this computer. Some printer properties will not be accessible unless you install the printer driver. Do you want to install the driver now?"
I figured the easy solution would be to just install a printer using the Point and Print driver and then remove it...leaving the driver behind (unless it cleaned up after itself). The problem with this is in my testing, even after removing and re-adding a printer that set off installation of this Point and Print driver on the client, any other printer that was using it still required the remove/re-add process in order to work, else I got the same message. That tells me that the driver may be custom for each printer installed?
Color me a little confused. I do NOT like the way that Server 2012 handles printers. I've come across a few other issues as well (such as duplex capabilities not being remember on the print server for a printer) that I never had with the 2003 server.
Thanks!!

I was able to get it to work using your instructions off of here and your other thread
Thank you!
Chris
http://social.technet.microsoft.com/Forums/en-US/864553c2-c8ff-49d2-bd48-eb7b47381111/windows-2012-print-server-problem-with-xp-clients?prof=required

Solaris 10 Containers and OS Compatibility

Similar Messages

Maybe you are looking for