| --- | Log | opened Sun May 15 00:00:48 2005 |
| 00:02 | karsten | if anyone's alive: UML on Debian (Sid) host, potato=>woody client (upgrade in process), issues w/ hwclock, sshd startup, and accessing multiple disk images via 'ubd<num>'=<file> cmdline opts. |
| 00:03 | karsten | hwclock: hangs UML. I've disabled it. Question: is there a fix in a later hwclock that addresses this or should I just remove / disable. |
| 00:03 | karsten | sshd startup: in sysvinit, UML hangs at or immediately after sshd startup. |
| 00:04 | karsten | Multiple disks. ubd0=root1.img ubd1=swap.img mem=96M eth0=tuntap,tap0,192.168.0.200 con0=fd:1,fd:2 con1=xterm con=null init=/bin/bash root=/dev/ubd |
| 00:04 | karsten | ...is cmdline. However mounting /dev/ubd/[1-7] all result in the same FS as /dev/ubd/0, as verified by listing FS. |
| 00:05 | karsten | Any assistance appreciated. |
| 00:07 | --- | ---> peter [~braam@218.247.142.130] has joined #uml |
| 00:08 | --- | <<-- jvds [~rghf@office.vaserv.com] has quit (Server closed connection) |
| 00:08 | --- | ---> jvds [~rghf@office.vaserv.com] has joined #uml |
| 00:40 | Newsome | Hi, Karsten. |
| 00:41 | Newsome | is the client trying to use devfs? (that's what it looks like with paths like /dev/ubd/?) |
| 00:42 | Newsome | Are the device nodes correct in the UML's root_fs? |
| 00:42 | --- | <<-- orospakr [~orospakr@CPE0004762b7051-CM001225701f0e.cpe.net.cable.rogers.com] has quit (Read error: Connection reset by peer) |
| 00:46 | Newsome | karsten: about hwclock, I recall something about debian UMLs hanging, but I don't remember the details. I'd say just remove/disable, since the UML is never going to have that sort of access to the hardware. |
| 00:48 | Newsome | karsten: after sshd starts, are you able to ping the UML? Is port 22 open on the UML? |
| 00:50 | Newsome | karsten: also, what UML kernel version? |
| 00:52 | karsten | Newsome: Sorry, was in another workspace... |
| 00:52 | Newsome | no problem :) |
| 00:52 | karsten | Linux version 2.4.26-3um (root@mizar) (gcc version 3.3.4 (Debian 1:3.3.4-11)) #2 Wed Sep 8 18:07:31 PDT 2004 |
| 00:53 | karsten | (/proc/version) from UML. |
| 00:53 | karsten | sshd _had_ worked prior to upgrade, and I was sshing into the UML instance. Now it's hanging. |
| 00:53 | Newsome | hmm |
| 00:53 | karsten | And yeah, I figured that setting hwclock was probably extraneous, but it's still annoying. |
| 00:53 | karsten | The real annoyance is not being able to mount an additional partition. |
| 00:54 | Newsome | the hwclock is probably not a problem under 2.6.x, but I haven't been following 2.4.x for a while |
| 00:54 | karsten | OK. |
| 00:54 | karsten | $ ls |
| 00:54 | karsten | misc mp mp2 root1.img root1.img.old swap.img uml-invocation |
| 00:54 | karsten | So those are my images. root1.img is the main root image, no problems with it. |
| 00:54 | karsten | swap.img now has an ext3fs in it. |
| 00:55 | karsten | $ sudo mount -o loop,ro -t ext3 $PWD/swap.img $PWD/mp |
| 00:55 | karsten | $ ls mp |
| 00:55 | karsten | lost+found this-is-the-swap-image |
| 00:55 | karsten | Where t-i-t-s-i is an empty file. That's viewed from host. |
| 00:55 | karsten | Guest: # cat /proc/cmdline |
| 00:55 | Newsome | okay |
| 00:55 | karsten | ubd0=root1.img ubd1=swap.img mem=96M eth0=tuntap,tap0,192.168.0.200 con0=fd:1,fd:2 con1=xterm con=null init=/bin/bash root=/dev/ubd0 |
| 00:56 | karsten | So I've got udb0 and udb1 |
| 00:56 | karsten | s/db/bd/ |
| 00:56 | karsten | # mount -r -t ext3 /dev/ubd/1 /mnt/misc0 |
| 00:56 | karsten | (guest) |
| 00:56 | karsten | # ls /mnt/misc0 |
| 00:56 | karsten | bin dev home lib mnt root sys usr |
| 00:56 | karsten | boot etc initrd lost+found proc sbin tmp var |
| 00:56 | Newsome | wow, that's odd |
| 00:57 | Newsome | how about 'ls -al /dev/ubd/?' |
| 00:57 | karsten | Repeat same for ubd's 0-7 |
| 00:57 | karsten | root@(none):/# ls -al /dev/ubd/ |
| 00:57 | karsten | total 22 |
| 00:57 | karsten | drwxr-xr-x 2 root root 1024 May 14 17:53 . |
| 00:57 | karsten | drwxr-xr-x 8 root root 20480 May 14 19:44 .. |
| 00:57 | karsten | brw-r--r-- 1 root root 98, 0 Dec 11 2002 0 |
| 00:57 | karsten | brw-r--r-- 1 root root 98, 1 Dec 11 2002 1 |
| 00:57 | karsten | brw-r--r-- 1 root root 98, 2 Dec 11 2002 2 |
| 00:57 | karsten | brw-r--r-- 1 root root 98, 3 Dec 11 2002 3 |
| 00:57 | karsten | brw-r--r-- 1 root root 98, 4 Dec 11 2002 4 |
| 00:57 | karsten | brw-r--r-- 1 root root 98, 5 Dec 11 2002 5 |
| 00:58 | karsten | brw-r--r-- 1 root root 98, 6 Dec 11 2002 6 |
| 00:58 | karsten | brw-r--r-- 1 root root 98, 7 Dec 11 2002 7 |
| 00:58 | karsten | Minors line up. |
| 00:59 | --- | ---> orospakr [~orospakr@CPE0004762b7051-CM001225701f0e.cpe.net.cable.rogers.com] has joined #uml |
| 00:59 | karsten | Got a paster page as well ... |
| 01:00 | karsten | ...when my connction stops lagging.... |
| 01:00 | karsten | I'll also suddenly go offline in about an hour, for a few seconds. |
| 01:00 | Newsome | okay |
| 01:01 | karsten | Anything else you'd like to see while we're at it? dmseg? |
| 01:01 | Newsome | can't think of anything right now. |
| 01:01 | Newsome | In the upgrade, do you know if debian changes from or to a devfs setup? |
| 01:02 | karsten | If so, it's not active. |
| 01:02 | karsten | The ubd thing preexisted the upgrade. |
| 01:03 | Newsome | so that didn't work before, either. Okay. |
| 01:03 | karsten | http://channels.debian.net/paste/141 |
| 01:03 | karsten | There we go. |
| 01:03 | karsten | The mounts, dmesg, /proc/partitions, /proc/mounts |
| 01:03 | Newsome | okay, try removing /dev/ubd/1, and recreating with "mknod /dev/ubd/1 b 98 16" |
| 01:04 | karsten | Then reboot? |
| 01:04 | Newsome | (i.e. different minor) |
| 01:04 | Newsome | no, shouldn't have to |
| 01:04 | karsten | right. |
| 01:05 | Newsome | just try mounting again off /dev/ubd/1 |
| 01:05 | karsten | Mounting /dev/ubd/2 still mounts ubd1 |
| 01:06 | karsten | Bingo! |
| 01:06 | karsten | # mount -r -t ext3 /dev/ubd/1 /mnt/misc0 |
| 01:06 | karsten | kjournald starting. Commit interval 5 seconds |
| 01:06 | karsten | EXT3-fs: mounted filesystem with ordered data mode. |
| 01:06 | karsten | root@(none):/# ls /mnt/misc0/ |
| 01:06 | karsten | lost+found this-is-the-swap-image |
| 01:06 | karsten | Bad minor modes? |
| 01:06 | Newsome | yeah, the minors 1-15 are used for partitioned ubds |
| 01:07 | karsten | Ahh.... |
| 01:07 | Newsome | check out your /proc/partitions |
| 01:07 | karsten | So I want 16-32? |
| 01:07 | Newsome | "98 0" and "98 16" |
| 01:07 | karsten | Right. |
| 01:07 | karsten | Ok.... |
| 01:07 | karsten | Incidentally, I'd tried creating partitioned images as well. |
| 01:07 | Newsome | Here's a script that could create all the nodes if you'd like - http://www.theshore.net/~caker/uml/makeUBDdev.sh |
| 01:08 | karsten | So I want 16-23? |
| 01:08 | Newsome | That'll create /dev/ubda with /dev/ubda[0-15] for example |
| 01:08 | Newsome | for partitioned ubd1, I think you'd use 17-31 |
| 01:09 | karsten | for i in $( seq 2 7 ); do rm $i; mknod $i b 98 $(( 16 + i - 1 )) ; done |
| 01:09 | karsten | Sorted, I think. |
| 01:10 | Newsome | the /dev/ubd/[0-7] location is common to devfs setups |
| 01:11 | karsten | OK. |
| 01:11 | Newsome | I think the more common location (at least now) is /dev/ubd[a-h] |
| 01:12 | Newsome | and /dev/ubd[a-h][0-14] being the nodes for partitioned ubds |
| 01:12 | Newsome | I'm using udev, so it only creates the ones that really exist |
| 01:13 | karsten | got it. |
| 01:13 | Newsome | ...thinking about the ssh startup problem... |
| 01:13 | --- | <<-- nextime [~nextime@213-140-6-96.fastres.net] has quit (Server closed connection) |
| 01:13 | karsten | Analagous to /dev/hda /dev/hda[0-20] |
| 01:13 | Newsome | yes |
| 01:14 | --- | ---> nextime [~nextime@213-140-6-96.fastres.net] has joined #uml |
| 01:15 | karsten | Script wants to be run in /dev/ubd, correct? |
| 01:16 | Newsome | my nodes are just in /dev, but /dev/ubd should work (creating /dev/ubd/ubd[a-h]) |
| 01:17 | karsten | Got it. |
| 01:18 | Newsome | about the sshd startup...do you have a /lib/tls directory? |
| 01:19 | Newsome | (inside the UML root) |
| 01:21 | karsten | One minute. |
| 01:21 | karsten | Weired problem copying / running that script. |
| 01:21 | Newsome | ok |
| 01:22 | karsten | File gets truncated when copying via hostfs. |
| 01:23 | Newsome | Ah. Yeah, I haven't had too much success with hostfs, myself. |
| 01:24 | karsten | Ok... |
| 01:24 | karsten | Can't get network up... |
| 01:24 | karsten | Let's try ye olde standby... |
| 01:25 | karsten | ...X11 cut'n'paste. |
| 01:25 | karsten | OK... |
| 01:25 | Newsome | good thing it's nice and short |
| 01:26 | | * karsten makes the script more verbose.... |
| 01:27 | karsten | It was running. It's slower than I'd thought it would be. |
| 01:27 | Newsome | hmm |
| 01:27 | karsten | Done. |
| 01:29 | karsten | OK, should I nuke /dev/ubd/* then? |
| 01:29 | Newsome | I would, but make sure you update fstab too |
| 01:29 | karsten | ;-) |
| 01:31 | Newsome | you mentioned this was a potato->woody upgrade? not ->sarge? |
| 01:32 | karsten | BTW, is it possible to specify /dev/fd0 to access host's floppy? |
| 01:32 | karsten | Yeah, I started from a base2_2.tgz image (potato). |
| 01:32 | karsten | Upgrading to woody. |
| 01:32 | karsten | Then sarge, eventually. |
| 01:32 | karsten | brb |
| 01:33 | Newsome | ok |
| 01:33 | karsten | back. |
| 01:34 | karsten | I'm going to reboot the guest a sec. |
| 01:34 | karsten | Having swap is sorta nice.... |
| 01:36 | karsten | OK, rebooting. |
| 01:36 | karsten | I'm doing init=/bin/bash BTW to get by the sshd issues. |
| 01:36 | Newsome | I think it should be possible to access the host's floppy, but I don't remember trying |
| 01:37 | karsten | Right. Of course, host's floppy is a not-very-often-mounted USB drive. |
| 01:37 | karsten | Feh. |
| 01:37 | Newsome | do you have a /lib/tls directory? |
| 01:37 | karsten | Frozen |
| 01:37 | Newsome | hrm, not good |
| 01:37 | karsten | freeze on 'ifconfig' |
| 01:38 | karsten | Yes, /lib/tls/ |
| 01:38 | karsten | ...and stuff in it. What's that? |
| 01:38 | karsten | I didn't have /proc mounted, that might have been part of it.... |
| 01:38 | Newsome | tls is thread-local storage, and UML doesn't support it. During the upgrade, it probably put those libraries there |
| 01:38 | karsten | /tmp is tmpfs |
| 01:39 | karsten | Nuke it? |
| 01:39 | Newsome | a common workaround is to "mv /lib/tls /lib/tls-broken" |
| 01:39 | karsten | ...or just nudge it ... right. |
| 01:39 | Newsome | well, recently Blaisorblade mentioned he was working on UML support for TLS, so it might work sometime...who knows |
| 01:40 | Newsome | or you can use dpkg-divert |
| 01:40 | karsten | Dig. |
| 01:41 | karsten | Usage? |
| 01:41 | karsten | TFM is vague. |
| 01:42 | Newsome | sorry, I'm not really sure. I've never stuck with a .deb-oriented distro longer than a week. :) |
| 01:42 | karsten | Heh. |
| 01:42 | karsten | What's your poison? |
| 01:43 | Newsome | Fedora Core right now. I've been using RH distros since 4.something. Slackware before that. |
| 01:43 | karsten | Sec, I'm in #debain too... |
| 01:44 | | * karsten started w/ RH 4.2 but got over it about 5.4. |
| 01:45 | Newsome | I think something like this (but the #debian people would know better): dpkg-divert --divert /lib/tls.backup /lib/tls |
| 01:46 | karsten | # dpkg-divert --divert /lib/tls-broken/ /lib/tls/ |
| 01:47 | karsten | Adding `local diversion of /lib/tls/ to /lib/tls-broken/' |
| 01:47 | karsten | Looks sane. |
| 01:49 | karsten | OK, booting signle rather than init=/bin/bash. |
| 01:49 | karsten | Adding Swap: 66552k swap-space (priority -1) |
| 01:49 | | * Newsome crosses fingers |
| 01:49 | karsten | w00t! |
| 01:49 | karsten | That's working now. |
| 01:49 | Newsome | great |
| 01:50 | karsten | Serial line 0 assigned device '/dev/ptyp0' |
| 01:50 | karsten | Unimplemented ioctl in ssl_ioctl : 0x541e |
| 01:50 | karsten | A few lines like that. |
| 01:50 | karsten | Cannot get serial info: Invalid argument |
| 01:51 | karsten | Incidentally, trying to set con<num>=port:9000 seems to indicate UML wants to spawn a telnet daemon, is this correct? |
| 01:51 | karsten | ...and/or hardwired? |
| 01:52 | Newsome | I think it's hardwired, but I don't remember |
| 01:52 | karsten | OK. |
| 01:53 | karsten | INIT: Entering runlevel: 2 |
| 01:53 | karsten | Starting services, slowly... |
| 01:53 | karsten | ssh... |
| 01:53 | karsten | Disabling protocol version 1. Could not load host key |
| 01:54 | karsten | . |
| 01:54 | karsten | ...is the next line, then nothing. Still single-user mode. No ping. |
| 01:54 | Newsome | hrm... |
| 01:54 | karsten | Any way, BTW, to try connecting to /dev/ptyp0? |
| 01:55 | karsten | ...which is where my serial lines are at. |
| 01:55 | Newsome | The "Serial line 0..." stuff comes from a serial console in /etc/inittab. Did you specify a serial line on the command line? |
| 01:56 | karsten | Lesse... Um. No, looks like I took that out: umid=uml1 ubd0=root1.img ubd1=swap.img mem=96M eth0=tuntap,tap0,192.168.0.200 con0=fd:1,fd:2 con1=xterm con=null single |
| 01:57 | karsten | Still hung. |
| 01:57 | karsten | uml_mconsole also unresponsive. |
| 01:57 | karsten | I've got about three minutes to a net interrupt, FYI. |
| 01:57 | Newsome | okay |
| 01:58 | karsten | Killed the UML process. |
| 01:58 | karsten | Suggestions for a serial line and appropriate /etc/inittab entries (guest)? |
| 01:59 | karsten | T0:23:respawn:/sbin/getty -L ttyS0 9600 vt100 |
| 01:59 | karsten | T1:23:respawn:/sbin/getty -L ttyS1 9600 vt100 |
| 01:59 | Newsome | I think those inittab entries look sane |
| 01:59 | karsten | T3:23:respawn:/sbin/mgetty -x0 -s 57600 ttyS3 |
| 01:59 | karsten | OK. How can I reach 'em from host? |
| 01:59 | karsten | ssl=??? |
| 01:59 | karsten | xterm? |
| 02:00 | Newsome | I think "ssl0=pty" and then look in the UML's boot messages to see what pty it gets |
| 02:01 | Newsome | then connect to that device. I don't know of a good, automated way, though. |
| 02:06 | --- | <<-- karsten [~karsten@host-66-81-223-254.rev.o1.com] has quit (Read error: Connection reset by peer) |
| 02:06 | --- | ---> karsten [~karsten@host-66-81-223-157.rev.o1.com] has joined #uml |
| 02:07 | Newsome | wb |
| 02:07 | Newsome | that must be rather annoying |
| 02:08 | --- | <<-- karsten [~karsten@host-66-81-223-157.rev.o1.com] has quit (Quit: ) |
| 02:08 | --- | ---> karsten_ [~karsten@host-66-81-223-157.rev.o1.com] has joined #uml |
| 02:08 | karsten_ | Ping |
| 02:09 | --- | User: *** karsten_ is now known as karsten |
| 02:09 | karsten | OK. |
| 02:09 | Newsome | welcome back |
| 02:09 | karsten | Danke. |
| 02:10 | karsten | The hard part was the oftc connection. And juggling two servers on irssi. |
| 02:10 | Newsome | provider cut you off every day? |
| 02:10 | karsten | Dialup. I kill and restart the connection at midnight to ensure a full night's updates. |
| 02:10 | Newsome | ah, okay |
| 02:10 | karsten | I get 12h connections, but want to make sure it's the right 12h. |
| 02:11 | karsten | So it's actually my own cronjob ;-) |
| 02:11 | karsten | OK, we were wondering about setting up serial lines. I'd like a backdoor into the guest even if networking is down (which it appears to be doing now). |
| 02:12 | Newsome | okay |
| 02:12 | Newsome | here's the last couple things I said, in case you didn't catch them |
| 02:12 | Newsome | I think "ssl0=pty" and then look in the UML's boot messages to see what pty it gets |
| 02:12 | Newsome | then connect to that device. I don't know of a good, automated way, though. |
| 02:12 | karsten | Basically, given minicom, telnet, and netcat, is there some way I can be assured of connecting.... |
| 02:12 | karsten | Hrm. |
| 02:12 | karsten | Might just set tty13 or something. I can chvt to that. |
| 02:13 | Newsome | or, start up in screen |
| 02:13 | karsten | What's that win me? |
| 02:14 | Newsome | if you want, you can have it spawn new screen windows, rather than starting xterms |
| 02:15 | karsten | Ah. |
| 02:15 | karsten | Special setting/ |
| 02:15 | karsten | s,/,?, |
| 02:16 | Newsome | yeah, a bit more complex to get setup, though |
| 02:16 | karsten | ssl0=tty13 ssl1=xterm |
| 02:16 | Newsome | or, for now, you could just add another con=xterm |
| 02:16 | karsten | I'll try that. |
| 02:17 | karsten | Problem is when I'm booting single, I don't get the add'l cons yet. Can launch 'em manually. |
| 02:17 | Newsome | oh, right |
| 02:19 | karsten | OK, serial 1 is an xterm w/ nothing in it... |
| 02:20 | karsten | ...spawned getty. |
| 02:20 | karsten | Got a login. |
| 02:20 | Newsome | that sounds good |
| 02:21 | karsten | Can't log in though. |
| 02:22 | Newsome | any users other than root? |
| 02:23 | karsten | Trying root |
| 02:23 | Newsome | sounds like the serial line isn't in /etc/securetty |
| 02:23 | karsten | Yes, I've got another user, but /etc/nologin won't let me. |
| 02:23 | Newsome | hrm |
| 02:24 | karsten | Hrm. If you kill the xterms they don't respawn. |
| 02:25 | karsten | Ifconfig says network is up but I can't ping host => guest ... |
| 02:26 | karsten | Going to quit and start _outside_ a host screen session. |
| 02:27 | Newsome | is tap0 configured? |
| 02:27 | karsten | Nested screen on nested host/guest sessions gets confusing. |
| 02:27 | karsten | Yeah. |
| 02:27 | karsten | tap0 Link encap:Ethernet HWaddr 00:FF:FC:93:61:5B |
| 02:27 | karsten | inet addr:192.168.0.200 Bcast:192.168.0.255 Mask:255.255.255.0 |
| 02:31 | --- | <<-- Dougie_ [~Doug@shadow.idmf.net] has quit (Server closed connection) |
| 02:31 | karsten | uml1:~# Serial line 3 assigned device '/dev/ptyp0' |
| 02:31 | --- | ---> Dougie_ [~Doug@shadow.idmf.net] has joined #uml |
| 02:31 | karsten | Any idea how to connect to /dev/ptyp0? |
| 02:31 | Newsome | try minicom to /dev/tty13 |
| 02:31 | Newsome | or something like that |
| 02:32 | Newsome | or, did you say it was going to one of the VTs? |
| 02:33 | karsten | sudo minicom --ptty=ttyp0 |
| 02:33 | Newsome | okay |
| 02:34 | karsten | Color, even ;-) |
| 02:35 | Newsome | sweet! |
| 02:35 | karsten | adduser <user> tty |
| 02:35 | karsten | ...allows access as well. Or you could chown/chgrp pty. |
| 02:35 | karsten | Naturally, it's not in securtty, but that answers another question.... |
| 02:37 | karsten | Just for grins I'm purging ssh. |
| 02:37 | karsten | I think I goofed the install what with hangs. |
| 02:37 | --- | <<-- Dougie [~Doug@shade.idmf.net] has quit (Server closed connection) |
| 02:37 | --- | ---> Dougie [~Doug@shade.idmf.net] has joined #uml |
| 02:38 | Newsome | does it ping before sshd tries to start up? |
| 02:38 | karsten | Nope |
| 02:38 | karsten | # ping 192.168.0.1 |
| 02:38 | karsten | PING 192.168.0.1 (192.168.0.1): 56 data bytes |
| 02:38 | karsten | <sits there> |
| 02:38 | karsten | 68 packets transmitted, 0 packets received, 100% packet loss |
| 02:39 | Newsome | doesn't sound like an ssh-specific problem, then |
| 02:39 | karsten | No, it's not. |
| 02:39 | karsten | Unrelated burp. |
| 02:39 | karsten | Networking *had* been working fine. Might be host-side. |
| 02:40 | Newsome | is the UML running as root, or as a user? |
| 02:40 | karsten | User. |
| 02:40 | karsten | Which is what I have been doing. |
| 02:40 | Newsome | check the permissions on /dev/net/tun |
| 02:40 | karsten | 666 |
| 02:40 | karsten | crw-rw-rw- 1 root root 10, 200 2003-04-19 21:21 /dev/net/tun |
| 02:40 | karsten | The mode of the beast. |
| 02:41 | Newsome | yeah, that looks fine |
| 02:41 | karsten | cat /proc/sys/net/ipv4/ip_forward: 1 |
| 02:41 | karsten | 192.168.0.201 0.0.0.0 255.255.255.255 UH 0 0 0 tap0 |
| 02:41 | karsten | 192.168.0.0 0.0.0.0 255.255.255.0 U 0 0 0 tap0 |
| 02:42 | karsten | (route) |
| 02:42 | karsten | cat /proc/sys/net/ipv4/conf/tap0/proxy_arp:1 |
| 02:44 | Newsome | you were pinging 192.168.0.1 or 192.168.0.201? |
| 02:44 | karsten | LAN GW == .1. Host == .33. tun0 == 200. Guest == .201 |
| 02:45 | karsten | I was pinging .1 just now. |
| 02:46 | Newsome | does the tap0 or UML show up in arp? |
| 02:46 | karsten | guest or host? |
| 02:46 | Newsome | guest |
| 02:47 | Newsome | er... |
| 02:47 | karsten | Sec. |
| 02:47 | karsten | Guest can ping its own IP. |
| 02:47 | karsten | # for i in 201 200 33 1; do ping -c3 192.168.0.$i; done |
| 02:47 | karsten | 3 packets transmitted, 3 packets received, 0% packet loss |
| 02:47 | karsten | 3 packets transmitted, 0 packets received, 100% packet loss |
| 02:47 | karsten | 3 packets transmitted, 0 packets received, 100% packet loss |
| 02:47 | karsten | 3 packets transmitted, 0 packets received, 100% packet loss |
| 02:47 | karsten | arping... |
| 02:48 | karsten | arp 192.168.0.33 (incomplete) eth0 |
| 02:48 | karsten | Restart. |
| 02:48 | karsten | uml1:~# arp |
| 02:48 | karsten | Address HWtype HWaddress Flags Mask Iface |
| 02:48 | karsten | 192.168.0.33 (incomplete) eth0 |
| 02:48 | karsten | 192.168.0.200 (incomplete) eth0 |
| 02:48 | karsten | 192.168.0.200 (incomplete) eth0 |
| 02:49 | Newsome | okay, how about from the host side now? |
| 02:49 | karsten | $ /usr/sbin/arp |
| 02:49 | karsten | Address HWtype HWaddress Flags Mask Iface |
| 02:49 | karsten | fritz.introspect (incomplete) eth0 |
| 02:49 | karsten | id.introspect (incomplete) eth0 |
| 02:49 | karsten | navel.introspect ether 00:04:5A:77:A7:83 C eth0 |
| 02:49 | karsten | jung.introspect ether 00:A0:CC:33:28:8F C eth0 |
| 02:49 | karsten | gateway.introspect ether 00:A0:CC:33:28:8F C eth0 |
| 02:49 | karsten | 192.168.0.201 ether FE:FD:C0:A8:00:C9 C tap0 |
| 02:49 | karsten | hplj4-1.introspect (incomplete) eth0 |
| 02:49 | karsten | 192.168.0.201 * <from_interface> MP eth0 |
| 02:49 | karsten | (fritz & id are down) |
| 02:49 | karsten | ...and printer. |
| 02:50 | Newsome | odd that 192.168.0.201 shows up twice |
| 02:51 | karsten | Yes. |
| 02:52 | karsten | Is that a problem? |
| 02:52 | | * karsten groks arp _very_ poorly. |
| 02:53 | karsten | Should I delete and re-add? |
| 02:53 | Newsome | I'm not certain. It may be getting confused since one is on tap0, and the other is on eth0 |
| 02:53 | karsten | Yeah. |
| 02:54 | Newsome | I'd try removing the arp entries for that IP |
| 02:54 | karsten | arp -d 192.168.0.201 |
| 02:54 | karsten | ? |
| 02:55 | Newsome | that looks right |
| 02:55 | karsten | Isn't deleting. |
| 02:55 | Newsome | might have to add "-i tap0" |
| 02:55 | Newsome | and "-i eth0" |
| 02:56 | karsten | Got rid of eth0. tap0 won't die. |
| 02:57 | Newsome | must still be in use. Does the UML IP ping now? |
| 02:57 | karsten | Sec, I'm going to configure tap0 in /etc/network/interfaces |
| 02:57 | karsten | No |
| 03:00 | --- | ---> Johnsie [~john@acs-24-154-32-12.zoominternet.net] has joined #uml |
| 03:01 | karsten | The UML networking docs are confusing. |
| 03:01 | Newsome | yes, they're pretty bad |
| 03:02 | Newsome | Here's a script I've used before to help setup networking - http://uml.tuxrocks.com/tmp/uml_net |
| 03:02 | karsten | For starters, they don't make clear what's guest and what's host. |
| 03:04 | Newsome | I think the networking docs are probably the most difficult of all the pages on that site |
| 03:04 | karsten | Heh. |
| 03:05 | | * karsten should rewrite them and resubmit. |
| 03:05 | karsten | brb |
| 03:05 | karsten | Reading requires food. |
| 03:05 | Newsome | heh |
| 03:08 | Newsome | I think there is some documentation going onto the wiki site, rather than directly onto the main sourceforge page. |
| 03:08 | Newsome | but the networking page could definitely use a revamp |
| 03:08 | karsten | Oi. |
| 03:09 | karsten | I was pretty surprised when I got networking going initially. |
| 03:09 | karsten | The whole devices thing (con/serial) is also confusingly presented. It's pretty straightforward in practice, once you grok it. |
| 03:10 | karsten | Problem seems to be a bunch of unstated assumptions about what's what. |
| 03:10 | Newsome | Yes. I think Jeff Dike probably wrote the docs, but he knows too much about the internals :) |
| 03:10 | karsten | Oi. |
| 03:11 | karsten | The UML kernel also plays a bit in userland, in both host and guest. |
| 03:12 | karsten | Spawning xterms, spawning telnetd. |
| 03:13 | karsten | Um. Should /dev/tap* exist (host) ? |
| 03:14 | Newsome | I don't think so |
| 03:14 | karsten | What does tunctl do? |
| 03:15 | karsten | /dev/net/tun, BTW, and it's there. |
| 03:15 | karsten | "tunctl uid" (in docs) is Just Plain Wrong. |
| 03:16 | Newsome | I'm not entirely sure what tunctl does. |
| 03:16 | karsten | Follows "Create the device with tunctl (available from the UML utilities tarball) " |
| 03:16 | Newsome | I usually do "tunctl -t tap0" or whatever |
| 03:16 | karsten | TUNSETIFF: Device or resource busy |
| 03:17 | karsten | RX bytes:50032530 (47.7 MiB) TX bytes:108853330 (103.8 MiB) |
| 03:17 | karsten | ...btw, before I just downed it. |
| 03:17 | karsten | So yes, it _was_ working. |
| 03:18 | Newsome | so traffic, but routing must have been screwed up |
| 03:18 | Newsome | or something along those lines |
| 03:18 | karsten | Yeah, I'm a tad confused. |
| 03:18 | karsten | Hrm.... |
| 03:18 | karsten | 64 bytes from 192.168.0.33: icmp_seq=1274 ttl=64 time=1.2 ms |
| 03:18 | karsten | Wroking |
| 03:18 | karsten | 1293 packets transmitted, 87 packets received, 93% packet loss |
| 03:19 | karsten | Bidirectional pings working. |
| 03:19 | Newsome | okay |
| 03:19 | karsten | Just started, within 90 seconds, so apparently bouncing tap0 fixed it. |
| 03:20 | karsten | OK, routing's still fscked. |
| 03:20 | --- | <<-- peter [~braam@218.247.142.130] has quit (Read error: Operation timed out) |
| 03:20 | karsten | Looks like I'm not forwarding packets... |
| 03:23 | Newsome | is it still in the arp table? |
| 03:23 | karsten | 192.168.0.201 ether FE:FD:C0:A8:00:C9 C tap0 |
| 03:24 | karsten | Only once now. |
| 03:24 | karsten | Are you at BYU? |
| 03:24 | karsten | http://pel.cs.byu.edu/~sorenson/ |
| 03:24 | karsten | ;-) |
| 03:25 | Newsome | yes, that's me :) |
| 03:25 | karsten | FWIW, I run sco.iwethey.org. |
| 03:25 | Newsome | Yes, I remember :) |
| 03:25 | karsten | ...and co-wrote "Modern SCO Executive" (Pirates of Penguinance). |
| 03:25 | Newsome | only, I've moved all my documents from the BYU site over to sco.tuxrocks.com |
| 03:26 | Newsome | Yes, the Pirates of Pengunance was very fun |
| 03:26 | karsten | Yeah, found that. |
| 03:26 | karsten | PJ rawks. |
| 03:26 | karsten | Heh. |
| 03:26 | karsten | Wrote it out during lunch hour. Tweaked it for the next day or so. |
| 03:26 | karsten | It still cracks me up. |
| 03:27 | karsten | guest: should route point to *LAN* GW, or *HOST* ? |
| 03:27 | karsten | 0.0.0.0 192.168.0.1 0.0.0.0 UG 0 0 0 eth0 |
| 03:27 | karsten | Should default == 192.168.0.33 (host IP) rather than 0.1 ? |
| 03:27 | Newsome | I think the guess should route to the LAN GW (so 0.1) |
| 03:28 | Newsome | er, guest |
| 03:28 | karsten | Testing anyway... |
| 03:29 | karsten | SIOCSIFFLAGS: Device or resource busy |
| 03:29 | karsten | Damnit |
| 03:29 | karsten | uml1:~# ifconfig eth0 192.168.0.201 netmask 255.255.255.0 up |
| 03:29 | karsten | SIOCSIFFLAGS: Device or resource busy |
| 03:29 | karsten | TUNSETIFF failed, errno = 16 |
| 03:29 | karsten | (dmseg) |
| 03:30 | Newsome | hmm... |
| 03:30 | karsten | Rebooting. |
| 03:31 | karsten | (guest) |
| 03:31 | --- | <<-- Nem^^ [~Nem@p54ABC2CB.dip.t-dialin.net] has quit (Remote host closed the connection) |
| 03:31 | --- | <<-- dizzy [~dizzy@ahriman.bucharest.roedu.net] has quit (Server closed connection) |
| 03:32 | --- | ---> dizzy [~dizzy@ahriman.bucharest.roedu.net] has joined #uml |
| 03:34 | --- | <<-- NemLappy^1 [~Nem@p54ABC2CB.dip.t-dialin.net] has quit (Read error: Operation timed out) |
| 03:36 | Newsome | Well, my eyes are starting to swim. I'd better get some sleep. Sorry we couldn't get everything running right tonight. Good luck. |
| 03:36 | karsten | Got it. |
| 03:36 | karsten | Hey, thanks tons for the help, very much appreciated. |
| 03:36 | karsten | I think it's routing or something like that. |
| 03:37 | Newsome | Glad to help. |
| 03:37 | Newsome | take a look at that short network-setup script. It might give you an idea what's wrong |
| 03:38 | karsten | Will do. |
| 03:38 | Newsome | night. |
| 03:38 | --- | <<-- Newsome [~sorenson@byu-gw.customer.csolutions.net] has quit (Quit: Leaving) |
| 03:38 | karsten | toodles. |
| 04:10 | --- | <<-- karsten [~karsten@host-66-81-223-157.rev.o1.com] has quit (Quit: leaving) |
| 04:12 | --- | ---> leighbb [~leigh@cpc1-heck1-4-0-cust245.hudd.cable.ntl.com] has joined #uml |
| 04:25 | --- | ---> braam [~braam@218.247.142.130] has joined #uml |
| 04:32 | --- | ---> NemLappy^ [~Nem@p54ABC2CB.dip.t-dialin.net] has joined #uml |
| 05:21 | --- | ---> schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has joined #uml |
| 05:23 | --- | <<-- DEac- [~deac@xdsl-84-44-151-144.netcologne.de] has quit (Ping timeout: 480 seconds) |
| 05:34 | --- | ---> DEac- [~deac@xdsl-84-44-148-36.netcologne.de] has joined #uml |
| 06:08 | --- | <<-- schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has quit (Remote host closed the connection) |
| 06:35 | --- | <<-- braam [~braam@218.247.142.130] has quit (Ping timeout: 480 seconds) |
| 06:54 | --- | ---> schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has joined #uml |
| 07:04 | --- | <<-- schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has quit (Quit: Client exiting) |
| 08:32 | --- | ---> Newsome [~sorenson@byu-gw.customer.csolutions.net] has joined #uml |
| 08:49 | --- | <<-- nextime [~nextime@213-140-6-96.fastres.net] has quit (Quit: nextime has no reason) |
| 09:02 | --- | ---> NemLappy^1 [~Nem@p54ABE7F4.dip.t-dialin.net] has joined #uml |
| 09:02 | --- | ---> nextime [~nextime@213-140-6-96.fastres.net] has joined #uml |
| 09:09 | --- | <<-- NemLappy^ [~Nem@p54ABC2CB.dip.t-dialin.net] has quit (Ping timeout: 480 seconds) |
| 09:12 | --- | ---> Basic [~Basic@fortress.tanners.org] has joined #uml |
| 09:28 | --- | ---> braam [~braam@221.220.24.132] has joined #uml |
| 09:45 | --- | ---> telekinetic [~roxtar@202.83.43.52] has joined #uml |
| 09:50 | --- | ---> TwoSchubert [~0x6F72746@TwoSchubert.user.oftc.net] has joined #uml |
| 09:53 | --- | <<-- telekinetic [~roxtar@202.83.43.52] has quit (Quit: Leaving) |
| 09:59 | --- | ---> peter [~braam@221.220.24.132] has joined #uml |
| 10:06 | --- | <<-- braam [~braam@221.220.24.132] has quit (Ping timeout: 480 seconds) |
| 10:52 | --- | <<-- jvds [~rghf@office.vaserv.com] has quit (Read error: Connection reset by peer) |
| 10:53 | --- | ---> jvds [~rghf@office.vaserv.com] has joined #uml |
| 11:19 | --- | ---> oOlli [~olli@dsl-084-060-101-187.arcor-ip.net] has joined #uml |
| 11:20 | oOlli | hi @all |
| 11:20 | oOlli | anyone @home? :) |
| 11:26 | --- | ---> tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has joined #uml |
| 11:26 | oOlli | hi tierra |
| 11:26 | --- | ---> olli_ [irc@baltz-online.de] has joined #uml |
| 11:27 | --- | <--- oOlli [~olli@dsl-084-060-101-187.arcor-ip.net] has left #uml (Verlassend) |
| 11:27 | tierra | morning olli_ |
| 11:28 | olli_ | morning? its quite late here in germany |
| 11:28 | tierra | hehe |
| 11:29 | olli_ | did you ever have problems with a freezing UML? |
| 11:29 | tierra | nope |
| 11:30 | olli_ | what kernels do you use? |
| 11:31 | tierra | 2.6.11-rc4 on host, 2.6.10 on guest |
| 11:32 | olli_ | ok. i just tried the 2.4x |
| 11:32 | olli_ | do you use SKAS? |
| 11:33 | tierra | yeah |
| 11:35 | olli_ | i`ll try that |
| 12:46 | --- | <<-- BurtyB [~chris@82-39-41-246.cable.ubr02.benw.blueyonder.co.uk] has quit (Ping timeout: 480 seconds) |
| 13:52 | --- | ---> Electric1lf [~david@usen-221x254x203x118.ap-US01.usen.ad.jp] has joined #uml |
| 13:54 | --- | <<-- ElectricElf [~david@electricelf.chair.oftc.net] has quit (Killed (NickServ command used by Electric1lf)) |
| 13:54 | --- | User: *** Electric1lf is now known as ElectricElf |
| 14:07 | --- | ---> schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has joined #uml |
| 14:24 | --- | <<-- schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has quit (Quit: Client exiting) |
| 14:52 | --- | ---> schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has joined #uml |
| 15:21 | --- | <<-- Johnsie [~john@acs-24-154-32-12.zoominternet.net] has quit (Quit: G'bye!) |
| 15:21 | --- | ---> Johnsie [~john@acs-24-154-32-12.zoominternet.net] has joined #uml |
| 15:52 | olli_ | hi. does anyone know how to devide localhost on host and guest uml? |
| 15:55 | Newsome | divide? |
| 15:57 | olli_ | yes. divide. my UML guests are int he same subnet like the host. if i start an nmap on the localhost of the guest i can see the open ports of the host. |
| 15:58 | Newsome | both UML guest and host should see themselves as localhost |
| 15:59 | olli_ | yes they do. but they seem to have the same localhost :( |
| 16:00 | Newsome | in the guest, does localhost show up with IP 127.0.0.1, or does nmap resolve localhost to another IP? |
| 16:02 | olli_ | 127.0.0.1 |
| 16:03 | olli_ | host "real" ip = 192.168.1.1, tap1 of host for uml1 is 192.168.1.1, too |
| 16:08 | Newsome | odd. I haven't seen that sort of network oddity |
| 16:09 | olli_ | OH sorry. you were right! there was no /etc/hosts on the client. that means: guest asked host for "localhost" |
| 16:12 | Newsome | ah, okay. That makes more sense :) |
| 16:29 | --- | <<-- leighbb [~leigh@cpc1-heck1-4-0-cust245.hudd.cable.ntl.com] has quit (Quit: Download Gaim: http://gaim.sourceforge.net/) |
| 16:48 | loko | tierra did you use someone elses guest kernel or build your own |
| 16:51 | tierra | using these: http://www.linode.com/~caker/uml/kernels/ |
| 16:51 | loko | yea his kernel don't support reboots though =) |
| 16:52 | tierra | I tend to copy the configs though |
| 16:53 | loko | so you just use his config and then recompile the kernel? |
| 16:53 | tierra | yeah |
| 16:53 | loko | yea thats basiclaly what i have been doing as well |
| 16:53 | mikegrb | liar liar pants on fire |
| 16:53 | | * mikegrb runs away |
| 16:54 | | * tierra kicks mikegrb |
| 16:54 | | * loko lites hunter's pants on fire |
| 16:54 | tierra | mikegrb, I've been playing with uml/xen on a home box |
| 16:55 | tierra | not much luck with either, but I'm trying to mimic the same environment as my Linode |
| 17:09 | --- | <<-- tierra [~tierra@dsl093-225-126.slc1.dsl.speakeasy.net] has quit (Quit: bug, n: A son of a glitch.) |
| 17:46 | --- | ---> JViz` [Anomaly@cpe-065-190-045-031.triad.res.rr.com] has joined #uml |
| 17:50 | --- | <<-- JViz [Anomaly@cpe-065-190-045-031.triad.res.rr.com] has quit (Ping timeout: 480 seconds) |
| 17:54 | --- | <<-- TwoSchubert [~0x6F72746@TwoSchubert.user.oftc.net] has quit (Quit: Never underestimate the power of stupid people in large groups!) |
| 18:03 | --- | ---> schlumpf2 [~schlumpf@dsl-084-056-138-247.arcor-ip.net] has joined #uml |
| 18:10 | --- | <<-- schlumpf [~schlumpf@dsl-084-056-128-198.arcor-ip.net] has quit (Ping timeout: 480 seconds) |
| 18:15 | --- | ---> Nem^ [~Nem@p54ABE7F4.dip.t-dialin.net] has joined #uml |
| 18:16 | --- | ---> Nem^^ [~Nem@p54ABDC23.dip.t-dialin.net] has joined #uml |
| 18:17 | --- | ---> NemLappy^ [~Nem@p54ABDC23.dip.t-dialin.net] has joined #uml |
| 18:22 | --- | <<-- Nem^ [~Nem@p54ABE7F4.dip.t-dialin.net] has quit (Read error: Operation timed out) |
| 18:22 | --- | <<-- NemLappy^1 [~Nem@p54ABE7F4.dip.t-dialin.net] has quit (Read error: Operation timed out) |
| 18:23 | --- | <<-- DEac- [~deac@xdsl-84-44-148-36.netcologne.de] has quit (Ping timeout: 480 seconds) |
| 18:35 | --- | ---> DEac- [~deac@xdsl-213-196-202-130.netcologne.de] has joined #uml |
| 19:10 | --- | User: *** silug_ is now known as silug |
| 19:43 | --- | ---> karsten [~karsten@host-66-81-221-62.rev.o1.com] has joined #uml |
| 19:48 | karsten | Newsome: Greets. All sorted. A few restarts and networking mysteriously started working right. |
| 19:49 | karsten | Newsome: ...I'm considering changing the name of UML networking to "newtorking" |
| 19:49 | karsten | Newsome: ...as in "She turned me into a newt, but it got better" |
| 20:01 | --- | <<-- Cowboy [~Cowboy@netblock-66-159-231-38.dslextreme.com] has quit (Remote host closed the connection) |
| 20:19 | --- | ---> lilo_ [debian-tor@lilo.usercloak.oftc.net] has joined #uml |
| 20:20 | --- | <<-- lilo [~lilo@lilo.usercloak.oftc.net] has quit (Ping timeout: 480 seconds) |
| 21:05 | --- | <<-- peter [~braam@221.220.24.132] has quit (Ping timeout: 480 seconds) |
| 21:20 | --- | ---> vvikram [~vvikram@epic26.Stanford.EDU] has joined #uml |
| 21:20 | --- | User: *** vvikram is now known as vulcanoid |
| 21:20 | vulcanoid | hi |
| 21:22 | vulcanoid | first time installing uml; on debian testing machine. i got the Debian image on sourceforge and i am using console and not X. i booted into the debian-uml setup but get "console respawning too fast" messages. |
| 21:22 | vulcanoid | i used con1=tty:/dev/tty6 to boot into console |
| 21:22 | vulcanoid | any idea what i am doing wrong. |
| 21:35 | vulcanoid | ok, /etc/inittab had devfs ttys/1 etc instead of ttyX |
| 21:58 | --- | ---> braam [~braam@218.247.142.130] has joined #uml |
| 22:00 | --- | ---> NemLappy^1 [~Nem@p54ABCF44.dip.t-dialin.net] has joined #uml |
| 22:00 | --- | <<-- orospakr [~orospakr@CPE0004762b7051-CM001225701f0e.cpe.net.cable.rogers.com] has quit (Quit: Leaving) |
| 22:01 | --- | ---> Nem^ [~Nem@p54ABCF44.dip.t-dialin.net] has joined #uml |
| 22:01 | --- | ---> peter [~braam@218.247.142.130] has joined #uml |
| 22:07 | --- | ---> braam_ [~braam@218.247.142.130] has joined #uml |
| 22:07 | --- | <<-- braam [~braam@218.247.142.130] has quit (Read error: Operation timed out) |
| 22:08 | --- | <<-- NemLappy^ [~Nem@p54ABDC23.dip.t-dialin.net] has quit (Ping timeout: 480 seconds) |
| 22:08 | --- | <<-- Nem^^ [~Nem@p54ABDC23.dip.t-dialin.net] has quit (Ping timeout: 480 seconds) |
| 22:10 | --- | <<-- peter [~braam@218.247.142.130] has quit (Ping timeout: 480 seconds) |
| 22:22 | --- | <<-- Newsome [~sorenson@byu-gw.customer.csolutions.net] has quit (Ping timeout: 480 seconds) |
| 22:24 | karsten | vulcanoid: Still here? |
| 22:24 | karsten | vulcanoid: Try 'con=pts'. Sounds like con was trying to run xterm and couldn't |
| 22:24 | karsten | vulcanoid: What's your startup line? |
| 22:25 | karsten | vulcanoid: ...be patient, I'm in a few channels and cooking dinner. |
| 22:49 | --- | <<-- ElectricElf [~david@electricelf.chair.oftc.net] has quit (Read error: Operation timed out) |
| 22:51 | vulcanoid | hi |
| 22:52 | --- | ---> ElectricElf [~david@electricelf.chair.oftc.net] has joined #uml |
| 22:52 | vulcanoid | yes. so i changed /etc/inittab to use tty6 instead of say ttyS/x and it works now |
| 22:53 | vulcanoid | i believe this is related to devfs? i dont know about that. so i just read the docs and made the change. i guess this is why i also had to change fstab to have /dev/ubd0 instead of /dev/ubd/0 ? |
| 22:59 | --- | <--- VS_ChanLog [~stats@ns.theshore.net] has left #uml (Rotating Logs) |
| 22:59 | --- | ---> VS_ChanLog [~stats@ns.theshore.net] has joined #uml |
| 23:04 | --- | <<-- Basic [~Basic@fortress.tanners.org] has quit (Quit: Leaving) |
| 23:13 | --- | <<-- vulcanoid [~vvikram@epic26.Stanford.EDU] has quit (Ping timeout: 480 seconds) |
| --- | Log | closed Mon May 16 00:00:31 2005 |
