Mar 31, 2009

CMX Stingray 888 FM 4GB -> volume automatically lowers after some time

As I wrote in one of my last entries, I own a CMX Stingray 888 FM MP3-Player with 4 GB

The last day's I experienced the situation, that I adjusted the volume during playing a podcast.
During the following time period the volume automatically decreased and kept steady at 22.

This is a problem, when listening when strolling beside a road where buses and trucks drive
by and you nearly hear nothing at this volume level.

Today I found, what I have changed to cause the behavior.

I have activated in System / ear protection, the protection level. With this active, the
volume is automatically adjusted.

After deslecting the featuer, the volume keeps constant.

accidentially stopped my OpenSolaris notebook with mdb

Today I read an article about ZFS performance, where somebody had
problems with the performance.

After disabling ZFS prefetch, his performance was better. So I did the following:

mdb -kw

did some measuring with ptime and set the value back.

mdb -kw

and the fatal error was ending mdb without thinking and just typing exit


This was a fatal error, because it executes the last action and wrote 0x0 as
the address of the exit function.

As I had typed, I remembered the that I read the following blog:
How to shoot yourself in the foot with mdb

All I had to do, was a hard power reset.

Mar 28, 2009

Acrobat Reader 9.1 on OpenSolaris Build 110 runs now.

In my former blog entry, I wrote about not being able to start the
Acrobat Reader 9.1 on my OpenSolaris Build 110 system without using
the profile shell.

Now after some investigation I found the problem.

bash-3.2$ ppriv -e -D acroread
acroread[2668]: missing privilege "proc_priocntl" (euid = 2558, syscall = 112) needed at secpolicy_setpriority+0x16
terminate called after throwing an instance of 'RSException'

After modifying my user account to set this to the default privileges I can now start the Acrobat Reader 9.1 in my bash
usermod -K \ defaultpriv=basic,dtrace_kernel,dtrace_proc,dtrace_user,proc_priocntl \

A logout/login was necessary to activate the login.

Acrobat Reader 9.1 on OpenSolaris Build 110

I installed the new Acrobat Reader 9.1 on my OpenSolaris Build 110 system.
Starting it as role root worked, but running as normal user didn't work.
I have a bash as standard shell.
The following profile is active for the user:
bash-3.2$ profiles
Console User
Suspend To RAM
Suspend To Disk
CPU Power Management
Network Wifi Info
Primary Administrator
Zone Management
Basic Solaris User

When I start a pfsh and start acoread, the application works.

Here is a truss excerpt when starting with bash:

6663: lwp_create(0x0804712C, 0x0080, 0x08047128) = 3
6663/1: priocntlsys(1, 0x080471AC, 2, 0x080472CC, 0) Err#1 EPERM [proc_priocntl]
6663/1: lwp_detach(3) = 0
6663/1: lwp_continue(3) = 0
6663/3: lwp_create() (returning as new lwp ...) = 0
6663/3: context(3, 0xC83C0A60)
6663/3: schedctl() = 0xC8093010
6663/3: lwp_sigmask(3, 0xFFBFFEFF, 0x0000FFF7) = 0xFFBFFEFF [0x0000FFFF]
6663/3: lwp_exit()
6663: brk(0x0BEF1E38) = 0
6663: brk(0x0BF09E38) = 0
6663: brk(0x0BF09E38) = 0
6663: brk(0x0BF41E38) = 0
6663: fstat64(2, 0x080470A0) = 0
6663: write(2, 0xCCAAFE30, 48) = 48
6663: t e r m i n a t e c a l l e d a f t e r t h r o w i n g
6663: a n i n s t a n c e o f '

Seems as if the priocntlsys-call is the problem.

There is also a discussion on the adobe forum, but nothing other helps on
the Solaris platform.

Upgraded OpenSolaris 2008.11 on my Laptop to Build 110

After the HOME PC also the Laptop

bash-3.2$ prtdiag
System Configuration: FUJITSU SIEMENS LIFEBOOK S6120
BIOS Configuration: Phoenix/FUJITSU Version 1.25 01/29/2004

is now upgraded to Build 110.

Didn't have any problem.

Mar 27, 2009

Upgraded HOME PC to OpenSolaris Build 110

This evening I upgraded the installatoin to Build 110.
For me the upgrade worked without any problem

D-LINK USB WLAN access still works.
SiS Xabre 300 works out of the box.
Audio does not work. ( never worked here on the PC )

Mar 24, 2009

terminator - Terminal : Switching the size of active shell

On my OpenSolaris box I'm using terminator instead of gnome-terminal ( which is used internally ).

You can open new shells in terminator with -- or --.

When pressing -- the actual shell with the cursor is
increased to the whole window size and pressing again shows all open

Mar 22, 2009

Home-PC upgraded to OpenSolaris Build 109

Today I started to try the upgrade of my home pc. An old PC with Pentium 4 CPU with
3.06 GHz and 512 MB RAM.

The GUI is a SIS Xabre card:

pci bus 0x0001 cardnum 0x00 function 0x00: vendor 0x1039 device 0x0330
Silicon Integrated Systems [SiS] 330 [Xabre] PCI/AGP VGA Display Adapter
Somwhere from Build 93, I couldn't update because of regular problems on accessing
the hard disk. Upgrade went fine, but after reboot the system was not able to
access the hard disk.

Today I attempte to upgrade to the last Build 109 on dev.

pkg set-authority -P -O

to set the default on the development branch.

After this the update took some long time, because of the small memory footprint.

But after this, the reboot was ok and I could boot to build 109.

The network connection was done with the interated Ethernet controller:

pci bus 0x0000 cardnum 0x04 function 0x00: vendor 0x1039 device 0x0900 Silicon Integrated Systems [SiS] SiS900 PCI Fast Ethernet
Therefore I had to use the sfe-Package from

The nwam also did the automatic connection. But to my surprise the WLAN-USB
stick which is an D-Link G122 ( as it looks on the webpages ) is now also recognized
by the system.

Disconnecting the Ethernet-LAN cable switched over to the USB WLAN.

Because of problems getting the intel driver on the Notebook to run with WPA,
I have changed to WEP.
After selecting the WLAN in nwam and logging in:

I got the connection to my AP and can log into the Internet via WLAN :-)

The D-Link uses the rum driver, which worked out of the box for me.

Also with the 512 MB memory, Surfing does not seem very slow.

Mar 19, 2009

Filesystem contains large files, but lacks LARGE_FILE flag in superblock.

This morning, the fileserver was not online. It is a linux server hosting some
filesystems which are shared via NFS and samba.

After the tape backup the server was hanging in the reboot sequence, because
of requesting a manual file system check.

The error message seen above was

Filesystem contains large files, but lacks LARGE_FILE flag in superblock.
After logging in with the root-password and issuing the fsck command manually,
there was one fix and the system could reboot.

The strange thing was, that this filesystem contains since some months large
files, but only this morning this was a problem.

dumpe2fs -h is a help to investigate the filesystem.

server:/root # dumpe2fs -h /dev/sdc1
dumpe2fs 1.28 (31-Aug-2002)
Filesystem volume name:
Last mounted on:
Filesystem UUID: 3a693ec2-aac3-4207-9282-060677ef3a2d
Filesystem magic number: 0xEF53
Filesystem revision #: 1 (dynamic)
Filesystem features: has_journal filetype needs_recovery sparse_super large_file
Filesystem state: clean
Errors behavior: Continue
Filesystem OS type: Linux
Inode count: 8978432
Block count: 17932548
Reserved block count: 179321
Free blocks: 3058549
Free inodes: 8678021
First block: 0
Block size: 4096
Fragment size: 4096
Blocks per group: 32768
Fragments per group: 32768
Inodes per group: 16384
Inode blocks per group: 512
Last mount time: Thu Mar 19 08:11:23 2009
Last write time: Thu Mar 19 08:11:23 2009
Mount count: 1
Maximum mount count: 26
Last checked: Thu Mar 19 08:09:25 2009
Check interval: 15552000 (6 months)
Next check after: Tue Sep 15 09:09:25 2009
Reserved blocks uid: 0 (user root)
Reserved blocks gid: 0 (group root)
First inode: 11
Inode size: 128
Journal UUID:
Journal inode: 8
Journal device: 0x0000
First orphan inode: 0

As you can see, it's already an old installation.
The interesting thing is, that this also shows the number of free inodes, which
can be shown by df -i /Directory

server:/root # df -i /DIR
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/sdc1 8978432 300418 8678014 4% /DIR
This was long time ago a problem, that the system responded with a full
filesystem, but df -h showed still free space. The problem was, that there were
no more free inodes available.

Mar 16, 2009

synergy & OpenSolaris Build 109 bring the @ back

I had problems running the intel driver on build 107/108/109. On the xwin-discuss
list, I got from a Sun employee a new intel driver. This one works now.
On Build 108, there was a problem with the kernel and the X-server. The
kernel didn't properly supply the keyboard.

So I had to configure it by hand in the xorg.conf.

With Build 109, I removed the keyboard configuration again and it works now.

I get the german keyboard.

The other thing is:

I'm using synergy for sharing the keyboard and mouse between my SunRay and
my OpenSolaris laptop.
During the usage of the keyboard configuration in the xorg.conf, I didn't get
the extra characters like @ and \, which are only reachable with the Alt-Gr on a
German keyboard.
But with Buiild 109 and the keyboard reporting of the kernel the characters
are back again.

Now I see also in the Xorg-log the used parameters:

(**) Option "CoreKeyboard"
(**) Keyboard0: always reports core events
(II) Keyboard0: Opened device "/dev/kbd"
(--) Keyboard0: Keyboard type: USB (6)
(--) Keyboard0: Keyboard layout: 9
(--) XKB: model: "sun_type6_euro_usb"
(--) XKB: layout: "de"
(--) Keyboard0: XkbModel = sun_type6_euro_usb
(--) Keyboard0: XkbLayout = de
(**) Option "AutoRepeat" "500 30"
(**) Option "XkbRules" "xorg"
(**) Keyboard0: XkbRules: "xorg"
(**) Option "XkbModel" "sun_type6_euro_usb"
(**) Keyboard0: XkbModel: "sun_type6_euro_usb"
(**) Option "XkbLayout" "de"
(**) Keyboard0: XkbLayout: "de"
(**) Option "Panix106" "off"
(**) Option "CustomKeycodes" "off"
(**) Keyboard0: CustomKeycodes disabled
I used just a simple rule of

Section "InputDevice"
Identifier "Keyboard0"
Driver "kbd"
# Option "XkbRules" "xorg"
# Option "XkbModel" "pc105"
# Option "XkbLayout" "de"
Option "BellDevice" "/dev/audio"
which used a different parametrization.

Mar 13, 2009

Upgrade OpenSolaris to Build 109

The upgrade to Build 109 was no problem for me.

pkg image-update
init 6

Everything worked fine. Nothing other to do. X11 came back after the reboot.

bash-3.2$ uname -a
SunOS solewer 5.11 snv_109 i86pc i386 i86pc Solaris

Mar 6, 2009

Mount of ext3 partition of an old linux system hangs on OpenSuSE 11.1

A colleague attempted to mount an ext3 partition from an old SuSE installation on
the new installed OpenSuSE 11.1, but the mount kept hanging for a long time.
The mount command was killed with SIGKILL.

To investigate the problem, I started with strace :

lxreka:~ # strace mount -t ext3 /dev/sdb2 /mnt
execve("/bin/mount", ["mount", "-t", "ext3", "/dev/sdb2", "/mnt"], [/* 87 vars */]) = 0
brk(0) = 0x805f000
access("/etc/", R_OK) = -1 ENOENT (No such file or directory)
open("/etc/", O_RDONLY) = 3
getpid() = 3470
open("/etc/mtab~3470", O_WRONLY|O_CREAT|O_LARGEFILE, 0600) = 3
close(3) = 0
gettimeofday({1236338198, 546933}, NULL) = 0
link("/etc/mtab~3470", "/etc/mtab~") = -1 EEXIST (File exists)
open("/etc/mtab~", O_WRONLY|O_LARGEFILE) = 3
gettimeofday({1236338198, 547779}, NULL) = 0
alarm(30) = 0
fcntl64(3, F_SETLKW64, {type=F_WRLCK, whence=SEEK_SET, start=0, len=0}^C
As can be seen, the mount kept in the fcntl hanging to set a write lock.

There were a lot of /etc/mtab~* files there. I removed every file and after this
the mount finished successful.

Mar 4, 2009

Solaris 10 could not NFS-mount from new installed Linux NFS server

A new installed NFS-Server based on Linux shared over NFS some directories.
On the Solaris 10 boxes the directories should be mounted via automount, but
the failed.

Also a manual mount fails:
mount -F nfs linux:/dir /dir

The reason here is, that both, the new linux server and the solaris 10 box, use
NFS Version 4, but Linux and Solaris are not compatible.

What I did to solve this, was setting vers=3 option in the on the NIS server and Solaris was able to mount again.

automounted directry was not accessible.

Yesterday a colleague transfered an NFS shared directory on a linux box.
Our two sunray servers are running without reboot. On one sunray server
the directory mounted via automount ( map's are shared via NIS ) was available,
on the other server it was not accessible.

Today I checked it again.
fuser /sdk

didn't reveal any user, so I did a
umount -f /sdk

and after this an
ls /sdk

showed the content again.

On the second box where still some users logged in. I suppose, that the directory was still in use somewhere.

Mar 2, 2009

Learning English - English as a Second Language

Yesterday, I had time and zapped through the podcast on ma iRadio from Noxon.

One interesting Podcast I found was ESL - English as a Second Language
It contains an Audio Podcast with a short story. After this the phrases are describe in english.

There is also a transcript for each episode from the website available.

Give it a try, if you want to learn a little bit more english.