<$BlogRSDUrl$>

Wednesday, June 16, 2010

Perl30Ex.dll + IIS 

POST isn't working with Perl30Ex.dll.  Here is what I've checked:
  1. Check install, permissions, etc (everything seems configured correctly)
    http://blogs.iis.net/wadeh/archive/2009/04/13/running-perl-on-iis-7.aspx
  2. Look through CGI doc, try different options (no apparent effect)
    http://perldoc.perl.org/CGI.html
  3. Enable failed request tracking (no failures logged)
    http://mvolo.com/blogs/serverside/archive/2007/07/26/Troubleshoot-IIS7-errors-like-a-pro.aspx
  4. Compare source pages between working (apache) and non-working (iis) systems
    They are the same
  5. Disable windows authentication in IIS


Saturday, June 12, 2010

Radeon+Debian (again) 

This morning's update of Debian testing borked my X again.  If I try to run startx, the system hangs (black screen, unresponsive (?) keyboard).  Luckily, when booting from an older kernel, X seems to run fine.

Here is my card:
gsharp@wormwood:~$ lspci -v |grep VGA
01:00.0 VGA compatible controller: ATI Technologies Inc RV710 [Radeon HD 4350] (prog-if 00 [VGA controller])

Here are the pertinent bits from /var/log/Xorg.0.log

Older kernel (working):
(II) [KMS] Kernel modesetting enabled.
(WW) Falling back to old probe method for vesa
(WW) Falling back to old probe method for fbdev
(II) LoadModule: "fbdevhw"
(II) Loading /usr/lib/xorg/modules/linux/libfbdevhw.so
(II) Module fbdevhw: vendor="X.Org Foundation"
        compiled for 1.7.7, module version = 0.0.2
        ABI class: X.Org Video Driver, version 6.0
(II) RADEON(0): Creating default Display subsection in Screen section
        "Default Screen" for depth/fbbpp 24/32
(==) RADEON(0): Depth 24, (--) framebuffer bpp 32
(II) RADEON(0): Pixel depth = 24 bits stored in 4 bytes (32 bpp pixmaps)

Newer kernel (not working):
(II) Primary Device is: PCI 01@00:00:0
(II) [KMS] drm report modesetting isn't supported.
(WW) Falling back to old probe method for vesa
(WW) Falling back to old probe method for fbdev
(II) Loading sub module "fbdevhw"
(II) LoadModule: "fbdevhw"
(II) Loading /usr/lib/xorg/modules/linux/libfbdevhw.so
(II) Module fbdevhw: vendor="X.Org Foundation"
        compiled for 1.7.7, module version = 0.0.2
        ABI class: X.Org Video Driver, version 6.0
(EE) open /dev/fb0: No such file or directory
(II) RADEON(0): TOTO SAYS 00000000fe9f0000
(II) RADEON(0): MMIO registers at 0x00000000fe9f0000: size 64KB

(II) RADEON(0): PCI bus 1 card 0 func 0
(II) RADEON(0): Creating default Display subsection in Screen section
        "Default Screen" for depth/fbbpp 24/32
...
(EE) RADEON(0): Timeout trying to update memory controller settings !
(EE) RADEON(0): You will probably crash now ...
(EE) RADEON(0): Timeout trying to update memory controller settings !
(EE) RADEON(0): You will probably crash now ...
...
The new (not working) kernel goes on and loads module "vgahw" as well, which the older (working) kernel doesn't.

Originally, I had these three packages installed:
xserver-xorg-video-ati
xserver-xorg-video-radeon
xserver-xorg-video-radeonhd
So I removed the radeon one (presumably not needed for my card?):
apt-get remove xserver-xorg-video-radeon
This worked (in the sense that I get video), but didn't work (in the sense that X reverted to VESA).  The Xorg.log yields:
(II) LoadModule: "ati"
(WW) Warning, couldn't open module ati
(II) UnloadModule: "ati"
(EE) Failed to load module "ati" (module does not exist, 0)
So, I manually add radeonhd to the /etc/X11/xorg.conf
Section "Device"
    Identifier     "Device0"
    Driver         "radeonhd"
EndSection
However, this causes X not to start with the following error:
(II) [DRM] Kernel mode setting enabled
(EE) FATAL: RadeonHD presently does not work with kernel modesetting (KMS).
Please disable KMS in your kernel.
(II) UnloadModule: "radeonhd"
(EE) Screen(s) found, but none have a usable configuration.
Apparently the best way to deal with this is to tell the kernel to disable kernel modesetting, by adding radeon.modeset=0 (or is it radeonhd.modeset ??).  See e.g. this link: http://lists.opensuse.org/radeonhd/2009-11/msg00175.html

To test this, I need to reboot, and specify the kernel parameters to grub.  This method is described in this link http://grumpymole.blogspot.com/2007/05/ubuntu-how-to-edit-grub-boot-parameters.html:
  • Press "Esc" to intercept boot countdown
  • Go to the grub menu entry
  • Press 'e' to start editing.
  • Scroll down to the "linux..." line. Add "radeon.modeset=0" to the end.  (Yes, radeon, not radeonhd).
  • Press CTRL-x to boot using that kernel and those parameters.
The method for making these changes permanent are not well described, but here is how to do it:
  • Edit /etc/default/grub
  • Change GRUB_CMDLINE_LINUX to include "radeon.modeset=0"
  • grub-mkconfig -o /boot/grub/grub.cfg
  • grub-install /dev/sdb (or whichever drive has the relevant MBR)
http://forum.nginx.org/read.php?26,90217,90217
References:
http://wiki.debian.org/AtiHowTo
http://www.x.org/wiki/radeonhd

















Friday, June 11, 2010

Cluster file synchronization 

I like unison well enough, but maybe something better needed for the cluster.  For example csync2.  Look at the post by Vitek, with his recommendations:

http://pooteeweet.org/blog/0/909#m909


Wednesday, June 09, 2010

kondemand problem 

I am getting sluggish performance out of twofish, with the following output from top.  Sometimes %si (software interrupts) goes very high, though it is not shown in the current top dump.
top - 18:21:05 up 35 days,  3:30,  1 user,  load average: 0.71, 0.48, 0.39
Tasks: 250 total,   1 running, 249 sleeping,   0 stopped,   0 zombie
Cpu(s):  0.0%us,  0.0%sy,  0.0%ni, 99.7%id,  0.0%wa,  0.0%hi,  0.3%si,  0.0%st
Mem:  20557640k total,   865564k used, 19692076k free,   159480k buffers
Swap: 23446828k total,        0k used, 23446828k free,   532112k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
13359 gcs6      20   0 51564 3560 2560 R   18  0.0   0:03.12 top
   93 root      15  -5     0    0    0 S    2  0.0 305:29.57 kondemand/5
 4168 root      20   0  8908  504  376 S    2  0.0 199:42.04 irqbalance
   95 root      15  -5     0    0    0 S    2  0.0 303:43.57 kondemand/7
   50 root      15  -5     0    0    0 S    1  0.0 217:07.99 events/15
   99 root      15  -5     0    0    0 S    1  0.0 306:53.07 kondemand/11
  101 root      15  -5     0    0    0 S    1  0.0 332:41.09 kondemand/13
   89 root      15  -5     0    0    0 S    1  0.0 308:02.11 kondemand/1
   91 root      15  -5     0    0    0 S    1  0.0 319:29.00 kondemand/3
 4750 root      20   0 14608  632  476 S    1  0.0   2:11.19 cron

Not sure the problem, but I will like to try upgrading the kernel.  It seems there are 3 different kernel revisions in the 4 machines.


Sunday, June 06, 2010

id3 tools on linux 

Ugh. Every time I need to debug ID3 tags, I have to sort through a bunch of buggy, obsolete id3 tools.  Well, this time I'll try to take notes.

Here are the ones in Debian (starting with command line tools):

id3
http://code.google.com/p/id3/

Completely useless (id3v1 only).

id3tool
http://nekohako.xware.cx/id3tool/
Completely useless (id3v1 only).


id3v2
http://id3v2.sourceforge.net/

Mostly useless.  Does not, for example, show encoding information.

mp3diags
http://mp3diags.sourceforge.net/
Nice program.  Wish it would give more information about encoding.

This page is powered by Blogger. Isn't yours?