Re: hibernate vs hibernate-ram ?

2008-04-14 Thread Daniel Tryba
On Sat, Apr 12, 2008 at 03:40:39PM +0200, Hans-J. Ullrich wrote:
 Second question:  Is there a way on a multi processor system to prevent the 
 system to use both cores (yes, I know, can start as a single core machine) 
 and after start up using the second (and now empty cpu) for one process ?

My guess would be to modify /etc/init.d/rcS from
exec /etc/init.d/rc S
to
exec numactl --physcpubind=0 /etc/init.d/rc S
(or better yet to modify the initrd)

The affinity of all subprocesses is inherrited, thus should be bound to
cpu0.

This offcourse will only function on a numa arch. But I guess smp has
similar controls.

-- 

 When you do things right, people won't be sure you've done anything at all.

   Daniel Tryba


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: nvidia way versus debian way of installing nvidia 3D graphics drivers in Lenny..

2008-04-14 Thread Lennart Sorensen
On Sat, Apr 12, 2008 at 10:00:25PM +, Michael Fothergill wrote:
 I have installed Debian Lenny on my HP G6062 laptop.  Xserver works 
 reasonably using the vesa driver.  I couldn't get it to work with the nv 
 driver.
 
 It seems to see the nvidia network controller as well.
 
 There is an nvidia 7000M graphics chip in the laptop.  I assume that it could 
 do some 3D graphics processing.
 
 But if I would want to use this capability I would need to install the 
 proprietary nvidia drivers manually.
 
 I Googled this and found quite a few web pages on this.
 
 It seems that Nvidia have their own installer and guide on their own web site 
 for Linux users but Debian also has its own Debian Way of installing the 
 nvidia drivers.
 
 There also seems to be other individuals with web pages giving advice on this 
 subject.
 
 Left to my own devices I would likely choose this site to get instructions on 
 how to proceed here:
 
 http://wiki.debian.org/NvidiaGraphicsDrivers
 
 But this seems to be a fast moving subject and maybe this site is out of date 
 now.
 
 I would appreciate comments on which web sites provide the most up to date 
 and appropriate way of doing it.
 
 I will use those sites and ignore the other ones and install the drivers.

I updated mine last week. :)

http://www.tinyplanet.ca/~lsorense/debian/debian-nvidia-dri-howto.html

It works for me, and I often get emails from people saying it works for
them.  Sometimes I get emails with questions about things not working,
which I then figure out and update the howto to deal with.

-- 
Len Sorensen


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



apt-* vs aptitude vs synaptic

2008-04-14 Thread Hans-J. Ullrich
Dear maintainers,

it seems for me, that the commandline tool apt, the ncurses tool aptitude 
and the graphical tool synaptic might use different databases. 

So I need a little more background (and knowledge) about this, otherwise I 
cannot explain myself, why apt-get dist-upgrade is giving another result 
as aptitude (in my case apt-get dist-upgrade wants to deinstall some 
openoffice.org-packages, beryl* and some other, but aptitude will not want 
to deinstall those).

Are they using all different databases ? Or is it just related to some 
configuration ?

I looked into the manuals, but found no explanation...

Regards

Hans
 
  


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: apt-* vs aptitude vs synaptic

2008-04-14 Thread Lennart Sorensen
On Mon, Apr 14, 2008 at 04:21:05PM +0200, Hans-J. Ullrich wrote:
 Dear maintainers,
 
 it seems for me, that the commandline tool apt, the ncurses tool aptitude 
 and the graphical tool synaptic might use different databases. 
 
 So I need a little more background (and knowledge) about this, otherwise I 
 cannot explain myself, why apt-get dist-upgrade is giving another result 
 as aptitude (in my case apt-get dist-upgrade wants to deinstall some 
 openoffice.org-packages, beryl* and some other, but aptitude will not want 
 to deinstall those).
 
 Are they using all different databases ? Or is it just related to some 
 configuration ?
 
 I looked into the manuals, but found no explanation...

They all use dpkg's database of installed packages, and the same list of
available packages.  They do however differ in how they resolve
dependancy comflicts.  Aptitude tries much mroe complex solutions to try
and avoid uninstalling something than apt-get.  Not sure what synaptic
does since I never use it.

I believe the officially recommended tool in Debian is aptitude as of
the Etch release.  It simply does dependancy resolution better than the
other tools.  I personally tend to mostly use apt-get still, mostly out
of habit.  Of course any apt-get command can be issued with aptitude the
same way, except you get the more advanced dependancy resolution.
aptitude will offer possible solutions to conflicts and let you pick an
option, while apt-get simply makes a decision and asks if you want to
proceed.

-- 
Len Sorensen


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: Problems installing amd64 with Supermicro motherboard

2008-04-14 Thread Lennart Sorensen
On Sat, Apr 12, 2008 at 03:56:58PM -0700, Francesco Pietra wrote:
 Hi Giacomo:
 I do have a fast internet connections, though the current installation of 
 amd64 was for a RealTek. On the Supermicro there is an integrated Gigabit, 
 which is not seen by Linux. This also means that I can't scp to the machine 
 and from it, which spoils the possibility offered by running computational 
 programs. Also the CDROM is not seen (by both Linux and BIOS): might it be 
 that the 80-wire ATA 133 cable used is not compatible with the Pioneer 
 DVD-110?

No an 80 wire IDE cable should always work.

 It seems that the needed steps are:
 
 1) Fix the hardware problem of the CDROM

Or someone disabled IDE support in the BIOS.  Or linux doesn't recognize
the IDE controller.

 2) Prevent Intel Boot Agent from initializing (which also takes time)

There should be a BIOS option to turn of network boot support.

 3) Deactivate from BIOS all that is not needed, such as parallel port, 
 COM1/COM2 not to subtract resources.

Usually not a bad idea.

 4) Reinstall amd64, possibly lenny instead of etch (finally, this is not a 
 server, it is a workstation)

On a new board, lenny is much more likely to be successful.

To get in to the BIOS, try just holding you hand on lots of keys on the
keyboard while the system powers up.  Often you can generate a keyboard
error, and it will offer you something like F1 to continue, F2 to enter
setup.  Other keys to enter the BIOS are DEL, INS, F2, F10, F12.  Some
laptops only let you enter the BIOS after a power off, not a reboot, not
sure if any server or desktop systems are like that.

-- 
Len Sorensen


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: Problems installing amd64 with Supermicro motherboard

2008-04-14 Thread Lennart Sorensen
On Sat, Apr 12, 2008 at 10:17:03AM -0700, Francesco Pietra wrote:
 Hi
 I finally succeeded in entering BIOS (although the Intel Boot Agent is a pest 
 that has still to be eliminated; thanks to ael for suggesting Freedos: not 
 yet tried, I first wanted to configure BIOS at least where I am sure what I  
 am doing).
 
 In configuring the CPU (4 dual-core amd 875 Rev E1: cache L1 256KB, cache L2 
 2048KB) it is not clear to me how to set the MTRR mapping. The choice is 
 between continuous (which makes the PCI hole noncacheable) and discrete 
 (which places the PCI hole beloww the 4GB boundary).
 
 Oddly, the System Memory is reported 20480MB (which agrees with the result of 
 command
 
 $ cat /proc/meminfo
 
 when Linux is launched. Actually, I installed aside each socket 2GB + 2GB + 
 1G + 1G Kingston DDR1 ECC, so I expected a total 24GB. All these memories 
 were OK on a Tyan S2895, and insertion on the Supermicro H8QCE board seems 
 correct. I did myself all the management of the 2GB, and I was carefully 
 grounded.  Should the filling of the slots be different?

Is all the RAM the same speed?  Is it all ECC?  Is it all buffered?

The specs for the board seem to indicate it supports 32GB (16 x 2GB) of
buffered 400MHz DDR, or 64GB (16 x 4GB) of buffered 333MHz DDR.

What happens if you only install the 2GB modules and leave out the 1GB
modules?

So out of your 24GB you are only seeing 20GB?

As far as I can tell from the manual the way to install the ram would
be:



CPU3 2GB in 1B
 2GB in 1A
1GB in 2A1GB in 2B
1GB in 2B1GB in 2A
2GB in 1A
2GB in 1BCPU4

2GB in 1BCPU1
2GB in 1A
1GB in 2B1GB in 2A
1GB in 2A1GB in 2B
 2GB in 1A
CPU2 2GB in 1B

Could you post your 'dmesg' output, to see what the e820 tables and mtrr
and such show as well as the kernel reserved memory?

-- 
Len Sorensen


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



apt-mirror oddity.

2008-04-14 Thread Chris Wakefield
Greetings all:

I've finally decided to set up my own mirror of unstable with apt-mirror on one 
of the machines on my network, which is running fine getting the initial 
download of x86_64.

Running nload on the box, I notice a steady upload of about 15% of download 
bandwith as well, can someone explain this?

Thanks,
Chris W.



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: apt-mirror oddity.

2008-04-14 Thread Lennart Sorensen
On Mon, Apr 14, 2008 at 09:44:47AM -0700, Chris Wakefield wrote:
 I've finally decided to set up my own mirror of unstable with apt-mirror on 
 one 
 of the machines on my network, which is running fine getting the initial 
 download of x86_64.
 
 Running nload on the box, I notice a steady upload of about 15% of download 
 bandwith as well, can someone explain this?

My guess with be that it is sending checksums and requests back to the
site you are downloading from.

-- 
Len Sorensen


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: Problems installing amd64 with Supermicro motherboard

2008-04-14 Thread Francesco Pietra
Hi
Sorry for having posted fragmentarily. Some issues (mostly due to my faulty 
doing) have been fixed. Some are still unresolved (I'll post later about your 
second kind mail, I have to find the way to take the dmesg output as the scp to 
my desktop does not work)


--- On Mon, 4/14/08, Lennart Sorensen [EMAIL PROTECTED] wrote:

 From: Lennart Sorensen [EMAIL PROTECTED]
 Subject: Re: Problems installing amd64 with Supermicro motherboard

 To: Francesco Pietra [EMAIL PROTECTED]
 Cc: Giacomo Mulas [EMAIL PROTECTED], debian-amd64@lists.debian.org, A J 
 Stiles [EMAIL PROTECTED]
 Date: Monday, April 14, 2008, 8:30 AM
 On Sat, Apr 12, 2008 at 03:56:58PM -0700, Francesco Pietra
 wrote:
  Hi Giacomo:
  I do have a fast internet connections, though the
 current installation of amd64 was for a RealTek. On the
 Supermicro there is an integrated Gigabit, which is not
 seen by Linux. This also means that I can't scp to the
 machine and from it, which spoils the possibility offered
 by running computational programs. Also the CDROM is not
 seen (by both Linux and BIOS): might it be that the 80-wire
 ATA 133 cable used is not compatible with the Pioneer
 DVD-110?
 
 No an 80 wire IDE cable should always work.


 
  It seems that the needed steps are:
  
  1) Fix the hardware problem of the CDROM
 
 Or someone disabled IDE support in the BIOS.  Or linux
 doesn't recognize
 the IDE controller.

The cable was not perfectly fixed. Now CD-ROM OK.

 
  2) Prevent Intel Boot Agent from initializing (which
 also takes time)
 
 There should be a BIOS option to turn of network boot
 support.


While I was planning to follow Intel's instructions how to prevent their Boot 
Agent to initialize, and wondering how my new SPM system can load DOS or 
FREEDOS, Supermicro Europe has just answered that:

Regarding Intel Boot Agent:
It is not possible to remove or disable the Intel Boot agent other then 
disabling onboard LAN via Jumper.

This seems to be in contrast with what Intel says at:
http://www.intel.com/support/network/adapter/pro100/bootagent/
General Information:

To prevent the Boot Agent from executing, enter your systems BIOS 
configuration and find the boot device order settings. Move the boot agent 
further down the list, preferably after the hard drive or whatever device you 
prefer to boot from.
The Boot Agent will initialize during the system startup, even if it is not the 
first boot device. During the time the initialization screen is displayed, you 
can press Control-S to configure the Boot Agent's internal settings.
To prevent the Boot Agent from even initializing, you can turn it off 
completely. The IntelĀ® Boot Agent can be completely turned off by using the 
IBAUtil.exe DOS utility found in the PROBOOT archive, using the command line 
IBAUTIL - FD (flashdisable).

Incidentally, there is no option  in the AMI BIOS to change the order settings. 
Because of this, may be Supermicro is correct on saying that the Boot Agent 
will always appear at my boots.
 
  3) Deactivate from BIOS all that is not needed, such
 as parallel port, COM1/COM2 not to subtract resources.
 
 Usually not a bad idea.

Actually I have deactivated parallel port only. I was unsure whether 
deactivating serial prevents PS/2 or some other basic device from working.



  4) Reinstall amd64, possibly lenny instead of etch
 (finally, this is not a server, it is a workstation)
 
 On a new board, lenny is much more likely to be successful.

If I succeed in setting the older etch installation on network, I plan to 
complete the task with a debootstrap to lenny. 



 
 To get in to the BIOS, try just holding you hand on lots of
 keys on the
 keyboard while the system powers up.  Often you can
 generate a keyboard
 error, and it will offer you something like F1 to
 continue, F2 to enter
 setup.  Other keys to enter the BIOS are DEL, INS,
 F2, F10, F12.  Some
 laptops only let you enter the BIOS after a power off, not
 a reboot, not
 sure if any server or desktop systems are like that.

As it is AMI, the key is DEL. It worked and I carried out few - absolutely safe 
- modifications to the BIOS settings. Then I rearranged - correctly - the RAM 
modules with the result that all 24GB are detected and the interleaved 
(128-bit) memory conditions are satisfied. Oddly, however, the BIOS can no more 
be accessed, while debian amd64 loads correctly.


Thanks
francesco pietra

 
 -- 
 Len Sorensen
 
 
 -- 
 To UNSUBSCRIBE, email to
 [EMAIL PROTECTED]
 with a subject of unsubscribe. Trouble? Contact
 [EMAIL PROTECTED]


  

Be a better friend, newshound, and 
know-it-all with Yahoo! Mobile.  Try it now.  
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ



Re: Problems installing amd64 with Supermicro motherboard

2008-04-14 Thread Lennart Sorensen
On Mon, Apr 14, 2008 at 10:23:29AM -0700, Francesco Pietra wrote:
 Sorry for having posted fragmentarily. Some issues (mostly due to my faulty 
 doing) have been fixed. Some are still unresolved (I'll post later about your 
 second kind mail, I have to find the way to take the dmesg output as the scp 
 to my desktop does not work)

Make sure openssh-server is installed on the target machine or scp won't
work.

-- 
Len Sorensen


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: Problems installing amd64 with Supermicro motherboard

2008-04-14 Thread Francesco Pietra
Hi
reported at the bottom the output of 

dmesg  gsemd


--- On Mon, 4/14/08, Lennart Sorensen [EMAIL PROTECTED] wrote:

 From: Lennart Sorensen [EMAIL PROTECTED]
 Subject: Re: Problems installing amd64 with Supermicro motherboard
 To: Francesco Pietra [EMAIL PROTECTED]
 Cc: Giacomo Mulas [EMAIL PROTECTED], debian-amd64@lists.debian.org, A J 
 Stiles [EMAIL PROTECTED]
 Date: Monday, April 14, 2008, 8:47 AM
 On Sat, Apr 12, 2008 at 10:17:03AM -0700, Francesco Pietra
 wrote:
  Hi
  I finally succeeded in entering BIOS (although the
 Intel Boot Agent is a pest that has still to be eliminated;
 thanks to ael for suggesting Freedos: not yet
 tried, 


Se please my previous answer: no chance to remove the Boot Agent. 

I first wanted to configure BIOS at least where I am
 sure what I  am doing).
  
  In configuring the CPU (4 dual-core amd 875 Rev E1:
 cache L1 256KB, cache L2 2048KB) it is not clear to me how
 to set the MTRR mapping. The choice is between
 continuous (which makes the PCI hole
 noncacheable) and discrete (which places the
 PCI hole beloww the 4GB boundary).
  
  Oddly, the System Memory is reported 20480MB (which
 agrees with the result of command
  
  $ cat /proc/meminfo
  
  when Linux is launched. Actually, I installed aside
 each socket 2GB + 2GB + 1G + 1G Kingston DDR1 ECC, so I
 expected a total 24GB. All these memories were OK on a Tyan
 S2895, and insertion on the Supermicro H8QCE board seems
 correct. I did myself all the management of the 2GB, and I
 was carefully grounded.  Should the filling of the slots be
 different?
 
 Is all the RAM the same speed?  Is it all ECC?  Is it all
 buffered?


Like in the case of the CD-ROM, one slot was not well fixed. Now OK.

 
 The specs for the board seem to indicate it supports 32GB
 (16 x 2GB) of
 buffered 400MHz DDR, or 64GB (16 x 4GB) of buffered 333MHz
 DDR.
 
 What happens if you only install the 2GB modules and leave
 out the 1GB
 modules?
 
 So out of your 24GB you are only seeing 20GB?
 
 As far as I can tell from the manual the way to install the
 ram would
 be:
 
 
 
 CPU3 2GB in 1B
  2GB in 1A
 1GB in 2A1GB in 2B
 1GB in 2B1GB in 2A
 2GB in 1A
 2GB in 1BCPU4
 
 2GB in 1BCPU1
 2GB in 1A
 1GB in 2B1GB in 2A
 1GB in 2A1GB in 2B
  2GB in 1A
 CPU2 2GB in 1B

Ypu are quite correct. In fact this is the order that I set after having edited 
the BIOS. Changing the position of the memory slots (or because one was badly 
fixed, has prevented my access to BIOS. It seems inescapable to clear CMOS. To 
this regard I posed the question to Supermicro Europe;

You suggested to clear CMOS by removing the battery and shorting the battery 
socket. This is considered unsafe by most people. Why not shortening the JBT1 
without removing the battery?

They have just answered By shorting the battery socket you can be 100% sure 
the COMS has been cleared



 
 Could you post your 'dmesg' output, to see what the
 e820 tables and mtrr
 and such show as well as the kernel reserved memory?

dmesg  dsemd

having loaded amd64 from the raid1 disks that were in use in my previous 
two-socket (dual-core opteron 265) machine. As you can see there is no correct 
driver for the embedded Gigabit. I have just got the right drived (module) from 
Supermicro and have to find the right way to load the module fro a diskette or 
usb key with modprobe (I never loaded a module before)

Bootdata ok (command line is root=/dev/md2 ro )
Linux version 2.6.18-3-amd64 (Debian 2.6.18-7) ([EMAIL PROTECTED]) (gcc version 
4.1.2 20061115 (prerelease) (Debian 4.1.1-20)) #1 SMP Mon Dec 4 17:04:37 CET 
2006
BIOS-provided physical RAM map:
 BIOS-e820:  - 0009fc00 (usable)
 BIOS-e820: 0009fc00 - 000a (reserved)
 BIOS-e820: 000e8000 - 0010 (reserved)
 BIOS-e820: 0010 - bfff (usable)
 BIOS-e820: bfff - bfffe000 (ACPI data)
 BIOS-e820: bfffe000 - c000 (ACPI NVS)
 BIOS-e820: fec0 - fec01000 (reserved)
 BIOS-e820: fee0 - fee01000 (reserved)
 BIOS-e820: ff70 - 0001 (reserved)
 BIOS-e820: 0001 - 00062000 (usable)
DMI 2.3 present.
ACPI: RSDP (v000 ACPIAM) @ 0x000fa1b0
ACPI: RSDT (v001 A M I  OEMRSDT  0x07000631 MSFT 0x0097) @ 
0xbfff
ACPI: FADT (v002 A M I  OEMFACP  0x07000631 MSFT 0x0097) @ 
0xbfff0200
ACPI: MADT (v001 A M I  OEMAPIC  0x07000631 MSFT 0x0097) @ 
0xbfff0390
ACPI: OEMB (v001 A M I  AMI_OEM  0x07000631 MSFT 0x0097) @ 
0xbfffe040
ACPI: SRAT (v001 A M I  OEMSRAT  0x07000631 MSFT 0x0097) @ 
0xbfff5ab0
ACPI: DSDT (v001  1HQC8 1HQC8003 0x0003 INTL 

Re: Problems installing amd64 with Supermicro motherboard

2008-04-14 Thread Francesco Pietra



--- On Mon, 4/14/08, Lennart Sorensen [EMAIL PROTECTED] wrote:

 From: Lennart Sorensen [EMAIL PROTECTED]
 Subject: Re: Problems installing amd64 with Supermicro motherboard
 To: Francesco Pietra [EMAIL PROTECTED]
 Cc: Giacomo Mulas [EMAIL PROTECTED], debian-amd64@lists.debian.org, A J 
 Stiles [EMAIL PROTECTED]
 Date: Monday, April 14, 2008, 10:59 AM
 On Mon, Apr 14, 2008 at 10:23:29AM -0700, Francesco Pietra
 wrote:
  Sorry for having posted fragmentarily. Some issues
 (mostly due to my faulty doing) have been fixed. Some are
 still unresolved (I'll post later about your second
 kind mail, I have to find the way to take the dmesg output
 as the scp to my desktop does not work)
 
 Make sure openssh-server is installed on the target machine
 or scp won't
 work.

The server is correct. What is lacking if internet connection (se my mail of a 
few minutes ago) so that things do not pass through the router.

Thanks
francesco
 -- 
 Len Sorensen
 
 
 -- 
 To UNSUBSCRIBE, email to
 [EMAIL PROTECTED]
 with a subject of unsubscribe. Trouble? Contact
 [EMAIL PROTECTED]


  

Be a better friend, newshound, and 
know-it-all with Yahoo! Mobile.  Try it now.  
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: Problems installing amd64 with Supermicro motherboard

2008-04-14 Thread Lennart Sorensen
On Mon, Apr 14, 2008 at 11:13:18AM -0700, Francesco Pietra wrote:
 Hi
 reported at the bottom the output of 

So you have the CDROM, and you have all your ram.  So what else was
wrong with it other than the silly netboot agent making bootup take a
while?

-- 
Len Sorensen


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: apt-* vs aptitude vs synaptic

2008-04-14 Thread Karl Schmidt

One other tool you might consider is wajig that includes the gui front-end gjig.

Apt has finally started doing logging, but have found wajig's logging to be more useful - 
and a log you might not want to delete except when moving between releases.


The wajig log has saved me a lot of time in figuring out what update broke 
something.

If you need to do work via the command line, wajig is intuitive and MUCH easier to learn 
than memorizing all the apt and dpkg switches.






Karl Schmidt EMail [EMAIL PROTECTED]
Transtronics, Inc. WEB http://xtronics.com
3209 West 9th StreetPh (785) 841-3089
Lawrence, KS 66049 FAX (785) 841-0434

Subtle recursive jokes in signatures are funny.




--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]