Thursday, January 31, 2008

dicom standard where?

Where are the different VR value representations defined?
PS 3.5 means section 5 - page 24-33 in 2006 then some encoding examples follow
where are general image attributes
PS 3.3 talks about each type of image type
I think CP14 is the supplement that helps to understand data elements.

Tuesday, January 29, 2008

minicom error message Phonelist garbled. What to do?

I could never figure out how minicom stores its state information. Where
does it put the phone numbers anyway???? Not it /etc/minicom or such.

Now ordinarily this is not a problem: let minicom worry itself
However today there was a problem.
When I startedup minicom:
error message:
Phonelist garbled:
and lo and behold:
i can't dial my list of numbers -
it apparently doesn't remember any of them!!!

error message:
Phonelist garbled.

I dont even know where minicom keeps them. This is irritating!

Well I did this
strings /usr/bin/minicom

and i looked around to see what it had in there: and looked for Phonelist garbled :)

minicom: cannot setreuid(%d, %d)
Error writing to ~/.dialdir!
Manually entered number
Phonelist corrupted
Phonelist garbled (?)
cp %s %s.%hd
Old dialdir copied as %s.%hd
No connection: %s.

so clearly there is something about .dialdir

so now I did
strings .dialdir (it is a binary file :))

and I found my old modem target phone numbers.
so I copied it to a new file minicom-old-dialdir :)
and I re-entered the minicom dialto numbers and
on restarting minicom it has regenerated a new .dialdir.

Sneaky program. Resoureful user!

raid 1 loses a disk how to add back in

cat /proc/mdstat
Personalities : [raid1]
md6 : active raid1 sda1[1]
488383936 blocks [2/1] [_U]

unused devices:

what to do?
if can find the drive on line
(fdisk -l )
then do
mdadm /dev/md6 -r /dev/sdb1

mdadm /dev/md6 -a /dev/sdb1

to add back in and then monitor with
at /proc/mdstat2:/home/wustl# cat /proc/mdstat
Personalities : [raid1]
md6 : active raid1 sdb1[1] sda1[0]
488383936 blocks [2/1] [U_]
[>....................] recovery = 0.0% (279680/488383936) finish=145.4min speed=55936K/sec

unused devices:

Sunday, January 27, 2008

postgresql 7.4 obsolte

installation of postgresql-8,3 will automatically create cluster postgresql-8.3
need to port old to this new.
however the upgrade automatically creates 8.3 cluster which must be dropped
before the transfer of old data to 8.3 version.
thus need to run

exim4 mail configuration files /etc/exim4/exim4.conf.template

there is sometimes a problem with verizon mail authentication
with the default setup of exim4

in file

need to set


see my posting of may 29 2007

giant Xsession error logs fill drive. /etc/X11/Xsession

I had noticed a bug a while ago that there were hidden error log files in my home directory as
large as 25G in size!

the solution to that was
to modify the file


#old contensts were
#exec >>"$ERRFILE" 2>&1
#my modification
exec > "$errfile" 2>&1

1. who needs these X error logs anyway?
2. even if we want em, let up should _overwrite them_ not add to the tail!
they can get really large over months!

see this post at


This problem just should be fixed. A running Debian system just
kills itself by filling the hard disk with .xsession-errors
files. Unless, of course, the administrator or all users
know how to prevent it.

I fixed it by editing my /etc/X11/Xsession:

Changed this:
exec >>"$ERRFILE" 2>&1

to be like this:
exec >"$ERRFILE" 2>&1

Now the .xsession-errors file is initialized everytime when
a user logs in.

Best regards,

-- Harri --

Wednesday, January 23, 2008

how to keep track of index in perl grep

well the idea is that
map { express ? $_:() } @list;
is equivalent to
grep exp @list;

so therefore

#!/usr/bin/perl -w
use strict;

my @blug = qw (apple bug orange twist app);

# now create anonymous array reference including the array index

my $i=0;
my @newarray;
foreach (@blug) {
push ( @newarray, [$_,$i] );

print "now this certainly works \n";
foreach (@newarray) {
print $$_[0] ."\n";

my @out= map {$$_[0]=~/app/ ? $_:() } @newarray;

print "and we are done \n";
foreach (@out) {
print $$_[0] ."\n";


Saturday, January 19, 2008

secure add more space to server to store stuff

ok a standard process.

1 added 2 new drives and moved out old one
have now
mdo-md6 thus 7 raids = 8 sata drives + 1 pata drive
i used up all 8 places on the sata300 cards (4+4).

then when i put in new drives and booted up it did not find /dev/md5 (1 drive in place).
why not because it was now the device /dev/sdh1 while the
/etc/mdadm/mdadm.conf file did not have /dev/sdh1 in the first line.
so when i added it in to fix the line to say
DEVICES /dev/sd[abcdefgh]1
then we were ok and
all the old raids were found and created on bootup.

next we needed to create new raid /dev/md6.
so steps
fdisk -l
found the two new drives that did not have any formating

fdisk /dev/sde
and selected p 1 whole drive type fd
fdisk /dev/sdg
and selected p 1 whole drive type fd

then i did
mdadm -C /dev/md6 --level=1 --raid-devices=2 /dev/sd[eg]1

error device /dev/md6 not found
so i did
mknod b 9 6 /dev/md6

and then i did
mdadm -C /dev/md6 --level=1 --raid-devices=2 /dev/sd[eg]1
and it worked fine to start creating it
mkfs.ext3 /dev/md6
then we wait for the sync of the raid and the formating to happen

then we do the
creation of the new
we need the new UUID for the raid...

so i did
cp /etc/mdadm/mdadm.conf /root/mdadm.conf

mdadm -E /dev/sde1 |grep UUI >>mdadm.conf

edit the mdadm.conf so that we now use the UUID to build the raid1 for /dev/md6
made sure that the devices line had /dev/sd[abcdefgh]1 so we can build all of them
then I edited the
/etc/init.d/raid_build_mount line to assemble and mount the new raid as
`mount /dev/md6 /home/big6 `
and then we are done.

Tuesday, January 15, 2008

browse over ssh

in addition to
ssh -ND

set up in
about:config in iceweasel
network.proxy.socks_remote_dns to true!

then all dns requests go out over ssh instead of over nework directly.

Friday, January 11, 2008

freeglut-dev (something like htat
libs needs -lGLU -lGL

for gtlglext

./configure --prefix=/usr

for gtkglextmm
./configure --prefix=/usr
and we are ok

use more space on the drive

df -h
shows you are not using whole drive
tune2fs -l /dev/md0
shows 5% of space reserved
can do
tune2fs -m 1 /dev/md0 drops it down to 1% wasted space. could do 0.
not on system drive.

Sunday, January 6, 2008

Getting Citrix for Carecast working on debian sid amd64

ok how did i get citrix working on the a debian64 machine.

download the latest citrix application from the website.

Now citrix uses 32 bit libs not the 64 bit libs you have native so you need to use


i actually don't know which of those. maybe just first one?

so then untar it and run the application


which installs to

/usr/lib/ICAClient$ ldd wfcmgr => (0xffffe000) => not available (or something like that...) => /usr/lib32/ (0xf7d35000) => /usr/lib32/ (0xf7d25000) => /usr/lib32/ (0xf7d1d000) => /usr/lib32/ (0xf7d04000) => /usr/lib32/ (0xf7cef000) => /lib32/ (0xf7ceb000) => /lib32/ (0xf7cd4000) => /lib32/ (0xf7b8d000) => /usr/lib32/ (0xf7b3d000) => /usr/lib32/ (0xf7a50000) => /usr/lib32/ (0xf7a42000) => /usr/lib32/ (0xf7a3f000)
/lib/ (0xf7fa2000) => /usr/lib32/ (0xf7a3a000)

thus notice that we need
(now this is what it shows after i did copied to /usr/lib32/

which as a 32 bit binary
does not seem to exist in debian sid at this moment? or i couldnt find it

so i went to a 32 bit debian sid machine and copied that file and put it in

and I did this linking...

ln -s /usr/lib32/ /usr/lib32/

now when i run

ldd wfcmgr

/usr/lib/ICAClient$ ldd wfcmgr => (0xffffe000) => /emul/ia32-linux/usr/lib/ (0xf7d3c000) => /usr/lib32/ (0xf7d35000) => /usr/lib32/ (0xf7d25000) => /usr/lib32/ (0xf7d1d000) => /usr/lib32/ (0xf7d04000) => /usr/lib32/ (0xf7cef000) => /lib32/ (0xf7ceb000) => /lib32/ (0xf7cd4000) => /lib32/ (0xf7b8d000) => /usr/lib32/ (0xf7b3d000) => /usr/lib32/ (0xf7a50000) => /usr/lib32/ (0xf7a42000) => /usr/lib32/ (0xf7a3f000)
/lib/ (0xf7fa2000) => /usr/lib32/ (0xf7a3a000)
now it seems to find it, although I certainly did NOT put it in


I put it in /usr/lib32/ which is bizarre! But hell it works. who knows what magic?

and I was done.

( thus notice thus we need to have the
/lib32 and /usr/lib32 directories
it looks like the correct ones are /usr/lib32

then to run citrix. go to the web site you want to go to
log in to citrix application center
then when you click:

it tells you that :

"you have chosen to open
which is a ICA file

open with "Browse for application"

and browse for the file


and select that and you will be ok.

then: Go to the Box:

"Citrix presentation server client for linux"

you will see the
Carecast Pool in the middle of box.
and go to the Connections menu
and select Connect
and you will open up a Carecast client

and you can log in!
Good Luck!

Saturday, January 5, 2008

where to get audio codecs?

i am suprised that i didnt have this link here before.

marillat repositories