Upgraded to OpenSolaris Build 133 from Build 132

Upgrade from Build 132 to 133 was without problems on my system.
The intel X11 driver still does not work (  last working version was 129 ) - only vesa driver.

Announcement mail is here.

The common-agent-container service also goes into maintenance.

root@solewer:~# beadm list                                                     
BE       Active Mountpoint Space   Policy Created
--       ------ ---------- -----   ------ -------
OSOL_129 -      -          164.07M static 2009-12-12 11:11
OSOL_132 NR     /          13.33G  static 2010-02-05 21:19
root@solewer:~# pfexec pkg image-update --be-name OSOL_133


DOWNLOAD                                  PKGS       FILES    XFER (MB)
Completed                              2254/2254 18603/18603  627.1/627.1

PHASE                                        ACTIONS
Removal Phase                            14563/14563
Install Phase                            71828/71828
Update Phase                             26051/26051
PHASE                                          ITEMS
Reading Existing Index                           8/8
Indexing Packages                          2254/2254
Optimizing Index...
PHASE                                          ITEMS
Indexing Packages                          2506/2506
pkg: unable to unmount /tmp/tmpFdKLOD
Deleting content cache

---------------------------------------------------------------------------
NOTE: Please review release notes posted at:

http://opensolaris.org/os/project/indiana/resources/relnotes/200906/x86/
---------------------------------------------------------------------------

using SSH protocol within python -> paramiko

For a script, I had to connect to a remote host using the ssh protocol.

The ssh implementation for python is in paramiko.

Paramiko needs also the Crypto Module.

For Solaris 10 I downloaded both packages, compiled and installed both
without problems. The command sequence for both packages are:

python setup.py build 

python setup.py install

After this python can be started and paramiko imported:

>>> import paramiko

A nice and helpful tutorial can be found here by Jesse Noller.

A simple start would be:

>>> ssh=paramiko.SSHClient()                                    
>>> ssh.connect( "127.0.0.1", username="ewer", password="test" )  

There is also an SFTP client for file transfer available.

remove fading of background during logout in a SunRay-Session

When you logout from the Solaris 10 System, the background is faded darker.
In a SunRay-Environment this causes multiple screen redraw with network-
traffic.

To deactivate the animation, this blog from Chris Gerhard was very useful for
me.

Insert into /etc/gtk-2.0/gtkrc for the system or ~/.gtkrc for the user:

gtk-enable-animations = 0

cssh - cluster SSH -> Solaris 10 Update 8 fails to start

Now I'm upgrading the SunRay Infrastructure from Sun Fire T1000
to Fujitsu Server ( Xeon CPU's ).

For managing both servers at once, I use cssh ( Cluster SSH ) . I installed
als pdsh but this has the same problem as cssh in the initial version.

cssh claims about a problem with the configuration option ConnectTimeout.

SSH on Solaris does not recognize this.

So I manually adapted the cssh-Script

   #$config{ssh_args} = " -x -o ConnectTimeout=10"
    $config{ssh_args} = " -x"

Removing the ConnectTimeout=10 option sovled the problem, and
I can now log into the systems.

Upgrade to OpenSolaris Dev Build 132

Upgrade to OpenSolaris Build 132 finished without problems.

root@solewer:~# pkg image-update --be-name OSOL_132                           
DOWNLOAD                                  PKGS       FILES    XFER (MB)
Completed                              1212/1212 41381/41381  870.2/870.2

PHASE                                        ACTIONS
Removal Phase                            12708/20400
Warning - directory usr/share/gdm/applications not empty - contents preserved in /tmp/tmptVgzps/var/pkg/lost+found/usr/share/gdm/applications-20100205T212121Z
Removal Phase                            13074/20400
Warning - directory usr/X11/share/man not empty - contents preserved in /tmp/tmptVgzps/var/pkg/lost+found/usr/X11/share/man-20100205T212122Z
Removal Phase                            20400/20400
Install Phase                            37175/37175
Update Phase                             37330/37330
PHASE                                          ITEMS
Reading Existing Index                           8/8
Indexing Packages                          1212/1212
Optimizing Index...
PHASE                                          ITEMS
Indexing Packages                          1397/1397
pkg: unable to unmount /tmp/tmptVgzps
Deleting content cache

---------------------------------------------------------------------------
NOTE: Please review release notes posted at:

http://opensolaris.org/os/project/indiana/resources/relnotes/200906/x86/
---------------------------------------------------------------------------

First start issued message about failing to start the
common agent container.

Feb  5 22:36:29 solewer cacao: [ID 702911 daemon.error] Error: Fail to start cacao agent. default
Feb  5 22:36:29 solewer svc.startd[9]: [ID 652011 daemon.warning] svc:/application/management/common-agent-container-1:default: Method "/usr/lib/cacao/lib/tools/scripts/cacao_smf start default" failed with exit status 1.
Feb  5 22:36:31 solewer cacao_launcher[5759]: [ID 302520 daemon.error] SUNWcacaort launcher : Common Agent Container exited with code <1>
Feb  5 22:36:31 solewer cacao_launcher[5759]: [ID 314456 daemon.error] SUNWcacaort launcher : no retries available
Feb  5 22:36:31 solewer cacao: [ID 702911 daemon.error] Error: Fail to start cacao agent. default
Feb  5 22:36:31 solewer svc.startd[9]: [ID 652011 daemon.warning] svc:/application/management/common-agent-container-1:default: Method "/usr/lib/cacao/lib/tools/scripts/cacao_smf start default" failed with exit status 1.
Feb  5 22:36:31 solewer svc.startd[9]: [ID 748625 daemon.error] application/management/common-agent-container-1:default failed: transitioned to maintenance (see 'svcs -xv' for details)                              

But the Xorg GUI fails to start with the intel driver.

/var/adm/messages contains a lot of this messages: 

Feb  5 22:41:15 solewer gdm-binary[5816]: [ID 702911 daemon.warning] WARNING: GdmDisplay: display lasted 1.542882 seconds
Feb  5 22:41:18 solewer gdm-binary[5816]: [ID 702911 daemon.warning] WARNING: GdmDisplay: display lasted 2.790445 seconds

Compiled and installed CUPS 1.4.2 on Solaris 10 U8

Today I downloaed the CUPS 1.4.2 release.

After setting some Environment variable to /usr/local/bin and /usr/local/lib ( installed packages from http://www.sunfreeware.com ) as decribed in
the INSTALL.txt file:

       CFLAGS="-I/some/directory" \
        CPPFLAGS="-I/some/directory" \
        CXXFLAGS="-I/some/directory" \
        DSOFLAGS="-L/some/directory" \
        LDFLAGS="-L/some/directory" \

I could compile and install CUPS, when I disable gss-api during the configure with --disable-gssapi ( Compile-Problems with Kerberos ).

Other options where --prefix and --enable-dbus --enable-dbus.

The source tree contains a SMF Descriptor in /init/cups.xml which 
can be imported with

svccfg import /init/cups.xml

After enabling the service, the browser can be pointed to http://localhost:631.

The new feature is, a searching button for printers in the network.
If you have the PPD files, the can also be supplied during the configuration.


Java-Printing with JDK 1.6.0 U18 in jEdit  results in a InvocationTragetException.

The solution was to add the cups-Library directory into LD_LIBRAY_PATH.
After this printing in jEdit with JDK worked.

New installed Solaris 10 fails to start rusers, rquota

I installed a new Solaris 10 U8 ( 10/09 ) on a x86 Server.

The SMF-Framework set the rusers-Service into maintenance mode.

The log-File contained the following message:

Property 'name' of instance svc:/network/rpc/rusers:default is missing, inconsistent or invalid

I had to just take the data for rpc from the local file and not via nis or
anything else.

-> Modified /etc/nsswitch.conf  and the service was online.

This link was helpful in my case.

AIX problem after moving into a different network

I love the solaris boxes. They are so simple to use!!!

This week I changed the IP Adress wit smitty on our p520 AIX 5.3.

After moving the server into the new network, it started up, but I
couldn't login over the serial console.

There were just message that the NIS server could not be reached.

To get into the "single user mode", I had to insert the DVD with the OS,
jump into the menu on the serial console, select the DVD as boot device
and start from there.

During this run I had to say, that I want a maintenance boot and the
disk to use.

After this, I had a shell, where I could manually modify the files, because
smitty didn't work.


Now the system needs about 3/4 hour to boot fully. Think it was the
automount, which let the system hang, but now I do not have more
time to try, why this is the case.

The system also doesn't boot automatically. Have to boot it from the
console.

Installing Solaris 10 U8 on a Fujitsu RX200 S5

This week I attempted to install Solaris 10 on a
Fujitsu RX200 S5.

I attached a USB-CDROM and started to install.

After the base config, the installer does not find a Hard disk to
install to.

The System has SAS-Disks inside.

The solution was, to enter the Raid-Bios and create a
virtual disk.

After the virtual Disk was available, the installer found the
hard disk to install to.

Mercurial 1.4.3 is available for download -> no change for solaris necessary

Mercurial 1.4.3 is available.

In the previous releases, I had to manually change some c-files to get it
compiled on OpenSolaris.

This version just worked with a "gmake local".

[ewer@solewer:/]$~/mercurial-1.4.3/hg --version
Mercurial Distributed SCM (version 1.4.3)

Copyright (C) 2005-2010 Matt Mackall and others
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.