January 2005 - openSUSE AMD64 - openSUSE Mailing Lists

Artificially limiting memory
by Calkins, Dennis 08 Jan '05

08 Jan '05

Hi, If I have a SUSE node with 8 GIG of memory, is there a way to trick the system into thinking it only has 4 GIG of memory? The Monta Vista's 2.4 Kernel had a kernel Command line Parameter. Does SUSE? Basically, I want to prove that my applications are benefiting from the larger memory ( they run slower with less memory ) without having to physically removing the DIMMS? Thanks in advance.

3 2

no remote X after upgrades to 9.1/9.2
by Joe Georger 07 Jan '05

07 Jan '05

I have 2 Opteron machines. Both were running Suse 9.0 amd64. I upgraded one to Suse 9.1 and the other to Suse 9.2. Now I cannot open X applications remotely when using rlogin. I get the "cannot connect to X server" even after using xhost + and setting the DISPLAY env variable on the remote machine. I am aware that ssh -X is preferable - and that does seem to work between my Suse boxes - but not all machines are configured properly as ssh servers. We are behind a firewall here, and we also have some machines on private lans, so if there is a way around this I would like to know. Thanks, Joe

6 15

9.2, hmmm
by Yoriaki FUJIMORI 07 Jan '05

07 Jan '05

Dear listers, I got SuSELinux9.2 for Japanese before the year end, and some bad experiences folloed. (1) The media is a Double layer. Unfortunately, my old DVD rom cannot read it properly. So, I had to buy new ones with DVD+R DL capabilities. One of them, NEC, however, cannot read 64bit distribution. At the beginning of installation, I see something to the effect that I am going to install 32bit linux onto 64bit machine!? On the second DVD+R DL, Plextor, I checked if I can read it. After confirming the media can be read properly, I created duplicates. A funny thing is the duplicate can be read properly on the NEC drive!? (2) I installed 9.2 on to EPoX 9NDA3+, with FX55, 2GB, and a single Seagate 160GB SATA. The machine has been running properly for a month or so as a FreeBSD5.3 diskless cluster. But, I have observed system freezes several times already with SuSE Linux 9.2. # An SATA drive is now connected to the SATA3 connector. I wonder if somebody can tell me a success story with a similar setup. Thanks for your attention. yoriaki fujimori

6 11

Reiserfs errors
by Alexandre Moutinho Santos 07 Jan '05

07 Jan '05

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I'm having a problem related to reiserfs in SuSE 9.2 x86_64. Everytime I try to compile Mozilla Firefox with the -march=k8 flag, the system hangs with kernel errors related to reiserfs. I then installed another disk (a SATA one) and formatted it with ext2 filesystem. After moving the mozilla folder to this new disk the compilation was made with absolutly no problems. I tried to copy the mozilla folder back to the disk formatted in reiserfs and the process hanged again my PC. I have a Asus A8V with a AMD3500+ (130nm) and the hard disk is a PATA Maxtor. I tried with both kernel versions and had the same problem. I'm using now the one that comes in the DVD because the new one have problems with alsasound. Alex -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.5 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFB3nJwPjVgz/mpp6YRAgaJAKCcs3l2/cRwSrrh0Fk6IMmAnLjTPACgh2BZ wXXB2UZvtsAaVeNPHF+Rv3o= =jIyD -----END PGP SIGNATURE-----

2 4

trouble with network on asus a8v
by fabrice piccini 07 Jan '05

07 Jan '05

Hello list, I'm having serious trouble with the following config: * asus A8V mobo - bios 1007 - * athlon 64 3500 (90nm) * 2*512Mb ram (3-3-3-8) * system disc on u320scsi (tekram dc390u4b/w card) maxtor 15k atlas 36GB * 3 data disks (maxtor ide 120GB) on hdb,hdc,hdd * no overclocking or special settings * suse 9.2 x86_64 with latest patches * marvell (on board) gigabit chip (sk98lin driver), connected to a dlink gigabit switch (dgs-1008d); At the present time, the purpose of this machine is to be tested before going into production. Here are the symptoms: - the system totally hangs when the vsftpd server (started thru xinetd) is called from the outside (32 bits suse9.2 boxes); reset is the only solution... - Nothing in /var/log/messages and no crash dump anywhere on disk - I've removed vsftpd and installed pure-ftpd instead -> This time, ftp transfers are ok (no more freeze) - Started nfs server on the box and created a share; when I try to mount the nfs share from a remote 32 bits box (asus A7n8x with marvell gigabit chip), the 64 bits system is again totally frozen ! Except that, the system is running like a race car (no network problem when I initiate network from the box - internet browsing is perfect)! Any suggestion, experience, clue, tip, etc... is welcome, 'cause I'm stuck ! cheers, fabrice

2 4

Kfloppy formatting
by Mike Roy 07 Jan '05

07 Jan '05

Hello everyone: I'm experiencing some problems trying to get Kfloppy to format a floppy disk. I'm running SuSE 9.1 (32-bit) on an older machine and SuSE 9.1 (64-bit) on the new machine. In either case, the floppies won't format. I've even tried the command line approach fdformat /dev/fd0 (u or h) 1440 and while the terminal seems to be working, the floppies are not formatted. When I try Kfloppy, the menu 'greys out' (after the 'destroy data' warning). Seems to me that when Kfloppy is working properly, there's a horizontal bar indicator showing the work being done (as a percentage). This 'bar' does not appear now. On the older machine, I've adjusted the fstab so that the floppy drive must be 'mounted' and 'unmounted' and not 'automatically mounted' as on the new machine. Any ideas greatly appreciated. Cheers, Mike

1 0

Re: [suse-amd64] USB storage read 64k problem
by Bernd Paysan 07 Jan '05

07 Jan '05

On Friday 07 January 2005 00:54, Matthew Dharm wrote: > Don't do it that way. Match on the SCSI ID strings. It's a lot > easier. The SCSI ID string is often quite meaningless. My USB MP3 stick claims as SCSI vendor "Generic", and as device "MusicDrive". I've a SCSI scanner that announces itself as "SCANNER" - and as far as I know, most devices from this vendor do the same, so plug&play doesn't work. The USB ID is much better. > But, you can follow the symlinks if you want. It's nominally > guaranteed to stay that way. > > Really, I'm not the hotplug expert. You should talk to them. Ok. On the other hand, I suspect that quite a number of USB storage devices have problems with 64k transfers, due to the fact that Windows doesn't do such long transfers ("what works with Windows is ok"). It's probably better to whitelist devices that can - and that would actually benefit from long transfers. You can increase the max_sectors just as easy as decrease them. -- Bernd Paysan "If you want it done right, you have to do it yourself" http://www.jwdt.com/~paysan/

1 0

Re: [suse-amd64] USB storage read 64k problem
by Bernd Paysan 06 Jan '05

06 Jan '05

On Thursday 06 January 2005 23:03, Matthew Dharm wrote: > > Since this will probably take some time to do it would be good > > if you could first put Bernd's patch in, and then later when you > > really move it move it together with all other exceptions. > > Otherwise there is a risk that it gets lost. > > The only think which it making it take any time is that people keep > insisting I put it in the kernel. Once we do that, hotplug has > nothing to do, so they never do anything. > > It's only like 3 lines of hotplug script... For me, it's not obvious where to put it. As far as I can see from the kernel log, the hotplug mechanism doesn't get involved until after the kernel has detected the SCSI drive - and then, getting at the USB vendor&device ID is guesswork (going into /sys/$DEVICE/device/../../.. - is this guaranteed to stay that way?). Maybe if you export the ids into the SCSI block device sysfs directory, we could talk about 3 lines of hotplug script. -- Bernd Paysan "If you want it done right, you have to do it yourself" http://www.jwdt.com/~paysan/

1 0

Re: [suse-amd64] USB storage read 64k problem
by Andi Kleen 06 Jan '05

06 Jan '05

On Thu, Jan 06, 2005 at 12:59:27PM -0800, Matthew Dharm wrote: > On Thu, Jan 06, 2005 at 09:53:44PM +0100, Andi Kleen wrote: > > > Since this can be fixed via sysfs at runtime, I'm not really interested in > > > a patch. > > > > > > We're actively trying to move all this stuff out to userspace where hotplug > > > can take care of it. > > > > Hmm, but who should maintain the exception lists then? > > When they are in the kernel there is at least one central place. > > The hotplug folks already want it. At least, that's my understanding. Since this will probably take some time to do it would be good if you could first put Bernd's patch in, and then later when you really move it move it together with all other exceptions. Otherwise there is a risk that it gets lost. -Andi

1 0

Re: [suse-amd64] USB storage read 64k problem
by Andi Kleen 06 Jan '05

06 Jan '05

> Since this can be fixed via sysfs at runtime, I'm not really interested in > a patch. > > We're actively trying to move all this stuff out to userspace where hotplug > can take care of it. Hmm, but who should maintain the exception lists then? When they are in the kernel there is at least one central place. -Andi

1 0