Re: [uml-user] Multiple UMLs sharing a single tap device

2006-03-11 Thread Julius Schwartzenberg
Jim Carter schreef: With many guests on one host, remember that the CPU is shared, and if several guests get CPU-intensive at the same time their individual performance will drop off. Also UMLs take a lot of memory; be sure there's enough. Do UMLs really take a lot of extra memory? I'm using a

Re: [uml-user] Multiple UMLs sharing a single tap device

2006-03-11 Thread Julius Schwartzenberg
Blaisorblade schreef: That's possible, but you need to make UML use the daemon transport, and run uml_switch beforehand, connecting it to a pre-setup tap0 device. See the main page for more info (http://user-mode-linux.sourceforge.net/ -> Virtual networking). Thanks for the tip. Seems to be wo

Re: [uml-user] Segfaulting applications inside UML

2006-03-11 Thread Julius Schwartzenberg
Blaisorblade schreef: Ah, ok, I see - you're using a 2.4 host kernel. TLS support can't work correctly then - all I can do is provide a better diagnostic for this case. You should continue using the "move /lib/tls away" solution, but on some distro (like Fedora Core 4 and beyond) it won't work

Re: [uml-user] Segfaulting applications inside UML

2006-03-10 Thread Julius Schwartzenberg
Blaisorblade schreef: Thanks a lot for the fix. It indeed solves the problem. I had actually also tried 2.6.15-bs3, but that gave me lot's of new TLS errors (and things still weren't working), so it seemed even worse. It was -bs3 or -bs1? -bs1 gave a lot of messages, but -bs3 not, and until now

[uml-user] Multiple UMLs sharing a single tap device

2006-03-10 Thread Julius Schwartzenberg
Hi, I was wondering if it was possible to have multiple UMLs share a single tap device. My goal is to have one subnet on tap0 in which all the UMLs are too, so you get one big virtual ethernet network. I now get 'TUNSETIFF failed, errno = 16' when I try to get the second UML to work though.

Re: [uml-user] Segfaulting applications inside UML

2006-03-10 Thread Julius Schwartzenberg
Blaisorblade schreef: On Friday 10 March 2006 01:06, Itamar Reis Peixoto wrote: try rm -rf /lib/tls At most mv /lib/tls /lib/tls.away - 2.6.15-bs3 supports TLS, and 2.6.17 should include officially the support. Thanks a lot for the fix. It indeed solves the problem. I had actually also tr

[uml-user] Segfaulting applications inside UML

2006-03-09 Thread Julius Schwartzenberg
Hi, I seem to have a problem were certain kinds of applications segfault when trying to run or use them. I'm using Debian Sarge. The guest kernel I'm using is version 2.6.15.6-bs2. I also had the problem with a plain 2.6.15.1 kernel. My host is Slackware 10.1 with kernel version 2.4.26. The

Re: [uml-user] Debian root_fs images

2006-01-03 Thread Julius Schwartzenberg
I've finally found some time to look into this again. Blaisorblade wrote: I haven't got any X11 stuff at all on my server, so I can't really try that. I just did xhost +serverhostname on my workstation, this should allow all connections from that host? Apart that as you say con=pts is good, b

Re: [uml-user] Debian root_fs images

2005-12-11 Thread Julius Schwartzenberg
Hi, Thanks again for your messages! Blaisorblade wrote: I do not have X11 on my server. I also tried it with the DISPLAY environent pointing to my workstation's X11 server, but I got another error then. Test if you can open an xterm directly first, it's not so easy (at least you need to use x

Re: [uml-user] Debian root_fs images

2005-12-05 Thread Julius Schwartzenberg
Thanks a lot for your messages! I still seem to have some issues though. Blaisorblade wrote: Still using devfs? I've seen some systems hanging because of that, time ago... I don't know. I believe the guide said the default kernel configuration would produce a working system, so I'm using that.

[uml-user] Debian root_fs images

2005-12-02 Thread Julius Schwartzenberg
Hi, I'm trying to set up a Debian UML system, but I can't seem to get the Debian root images to work. One of them just hangs during the initialization (after I had created a symlink manually from /dev/ubd0 to /dev/ubd/0) and another one tried to launch an xterm (?) after initialization. Is it