[00:34] Doener (~doener@pD9E1277D.dip.t-dialin.net) joined #vserver. [00:34] hi all! [00:44] Simon (~sgarner@apollo.quattro.net.nz) joined #vserver. [00:48] Nick change: Bertl_oO -> Bertl [00:48] hi folks! [00:49] hi bertl [00:52] Anyone here work for the telephone company and have access to who owns what number? [00:52] Such as the ARIN database for IP's. [00:55] hmm, there are CD-Roms containing the # -> person database (at least here in europe ;) [00:56] hmm [00:56] or do you mean # -> phone company [00:56] Every number, or just people in the telephone books? [00:56] # -> person [00:57] phone company changes too easily now here in US. every number is portable accross carriers [00:57] well unlisted numbers will not be listed, wont they? [00:58] That's what I am trying to do at the moment, identify the owners of an unlisted number [00:59] hmm, isn't that illegal? [01:00] it would be against company regulations to give out unlisted information, but I don't think it's illegal [01:02] yes [01:03] hmm, why not 'just' call that number? [01:45] Simon, you around today? [02:20] :e [02:20] re [02:20] hi maja! [02:21] ok, serial cable is plugged in [02:21] now i need to read how to start a getty on my serial port [02:21] nope [02:21] all you do is the following: enable serial console in the kernel config [02:21] ok [02:22] and add console=ttyS0,115200n8 [02:22] to the boot line ... [02:22] on the other end you use minicom ... [02:22] on grub, to the kernel /.... root=... line, am i right? [02:22] configuration 115200 baud 8 bits 1 stop parity=none, hw hanshake yes ... [02:22] yup that line ... [02:25] ok, rebuilding [02:28] while i rebuild the kernel, i've got two questions regarding server rescues [02:28] go ahead ... [02:28] first, if i connect two servers via serial cable to have an emergency console i can use [02:29] do i need 2 cables? [02:29] so if server 1 goes down, i can connect from server2 [02:29] and vice versa [02:29] or is it possible to use 1 cable? [02:30] and has grub the possibility to controll it via serial cable? [02:30] because i do not know what to do if hda failes, where i boot from [02:31] yes, both is possible ... [02:31] ok, i guess i will do a little googling then [02:31] although you might use two cables ro 'cross' connect the servers for the very unlikely case where a power outage boots them at the same time ;) [02:32] for grub, you can simply specify the possible consoles, it will use all of them ;) [02:32] ok [02:40] re [02:40] my pc rebooted :) [02:40] (workstation) [02:41] its not good to switch a monitor that often, so you press your vga card out of the slot... [02:41] hmm true 8-) [02:41] (i forgot to add a screw to it) [02:45] you said i should use minicon? [02:45] is that one for windows? [02:47] minicom, no it's for unix/linux ... [02:47] who uses windows?! [02:48] me ;) [02:48] i like those nifty games [02:48] well then find your nifty vt100 terminal emulation ;) [02:48] ill try to look, if putty supports that :D [02:50] ok, no putty for that [02:50] doesn't the windows telnet client have vt100 emu? [02:51] yes, but i don't know how to access the serial port ;) [02:51] Use hyperterminal [02:51] That's in windows itself. [02:51] hyperterm :) [02:51] Or use SecureCRT. [02:52] mhm, i guess i press reset and look what it does ;) [02:52] should i be able to press some keys? [02:54] cause i receive an empty screen [02:54] hmm, well or 'just' boot into linux and use minicom ;) [02:54] the serial ports are always enabled? [02:55] because i usually disable everything i do not need :) [02:55] well, you can disable them in the bios ... [02:55] so my linux machines come with no usb/parport/... support [02:55] is it possible to use the comport from inside a vserver? [02:56] yes sure, but why do you want dto do that? [02:56] because i've got a vserver at home, which i trash with different programms and packages [02:57] so the host system stays clean [02:58] hmm, how many machines do you have? [02:58] i guess, that the comport which is closest to the ps2 ports is com1? [02:59] usually they are labeled A/B or 1/2 [02:59] A=1= ttyS0 [02:59] currently, i got 1 windows machine and 3 linux servers [02:59] so why use the windows machine for the serial link?! [03:00] my old router, my new router (which hast to be fully configured yet) and the server [03:00] i allready switched to the linux server [03:00] ah okay, so this is solved, right? [03:00] yes [03:01] i've got no /dev/vttyS0 on my vserver :) [03:01] well, using the host now [03:01] devfs or hand made /dev? [03:02] hand made [03:02] ill install devfs afterwards [03:02] /dev/ttyS0 should be there ... [03:02] as well as cua/cub [03:02] well, it isnt ;) [03:03] i only got [03:03] full/initctl/log/null/ptmx/pts/random/reboot/shm/ttyp/urandom/xconsole/zero [03:03] on my vserver [03:03] anyways [03:03] what do i need to pass to minicom? [03:04] jsut start it with -so [03:04] then select serial port setup ... [03:04] A - Serial Device : /dev/ttyS0 [03:04] B - Lockfile Location : /var/lock [03:04] C - Callin Program : [03:04] D - Callout Program : [03:04] E - Bps/Par/Bits : 115200 8N1 [03:04] F - Hardware Flow Control : Yes [03:05] G - Software Flow Control : No [03:05] ok [03:05] and in the modem/dialing setup [03:06] remove the strings A-H [03:06] done [03:06] then either select safe setup as default or as name ... [03:06] and then exit ... [03:06] make sure your terminal has at least 80x25 ... [03:07] most terminal open at 80x24 ;) [03:07] (in which case you would lose the status line ;) [03:08] i use putty, lots of space there :> [03:09] on exit, i get a "cannot open /dev/ttyS0" [03:09] CosmicRay (~jgoerzen@glockenspiel.complete.org) left irc: Ping timeout: 480 seconds [03:09] mhm [03:09] crw-rw---- 1 root dialout 4, 64 Mar 14 2002 /dev/ttyS0 [03:10] okay, so you probably really managed to configure the serial port away, ether by bios or kernel options ... [03:10] maybe there is a module to load? [03:10] i got monolithic kernels [03:10] hmm, well, then I guess not ... [03:10] is the problem with the server or the client [03:10] or could it be either one [03:11] only the machine the minicom runs on atm ... [03:11] could be a permission problem? [03:12] ok [03:12] i guess it was a perm problem [03:12] Welcome to minicom 2.1 [03:12] OPTIONS: History Buffer, F-key Macros, Search History Buffer, I18n [03:12] Compiled on Nov 12 2003, 19:21:57. [03:12] Press CTRL-A Z for help on special keys [03:12] CTRL-A Z for help |115200 8N1 | NOR | Minicom 2.1 | VT102 | Offline [03:13] okay, that looks better ... [03:13] is the cable connected? [03:13] yes [03:14] hmm, you should see an online there ... but maybe the other side doesn't connect yet ... [03:14] let's test this ... [03:15] s0:23:respawn:/sbin/mgetty -s 115200 -r -p "@: " -i /etc/issue.line /dev/tts/0 [03:15] add this line on the server (other end) [03:15] in the /etc/inittab [03:16] done [03:16] and activate the changes with telinit Q [03:16] hmm you might actually touch the /etc/issue.line before ;) [03:17] and make sure that mgetty is installed ... [03:17] done [03:17] root 445 0.2 0.0 1744 656 ? S 03:18 0:00 /sbin/mgetty -s 115200 -r -p @: -i /etc/issue.line /dev/tts/0 [03:17] CTRL-A Z for help |115200 8N1 | NOR | Minicom 2.1 | VT102 | Online 00:00 [03:17] hey it's working ;) [03:18] wheee \o/ [03:18] press enter ... [03:18] woot [03:18] hehe [03:18] nice [03:18] red.ipax.tk login: [03:18] okay, that's the normal console ... this doesn't interfere with the debug/boot stuff, so you can leave it in ... [03:19] you can always use that to reconfigure the network for example ;) [03:19] thoughts like these came to my mind [03:19] or enter the rescue partition ... [03:19] pretty nice actually [03:19] i think ill play around with that [03:19] after the work is done ;) [03:20] i just noticed, that my hyperterm received stuff while i was installing minicom [03:20] md1: sync done and stuff [03:20] Doener (~doener@pD9E1277D.dip.t-dialin.net) left irc: Quit: Leaving [03:20] anyways [03:20] okay, you added the boot line to the grub config? [03:21] yes [03:22] kernel /vmlinuz-2.4.24-vs1.23-2.95 root=/dev/md0 console=ttyS0,115200n8 [03:25] looks good ... [03:25] now modify the grub config file ... [03:26] serial --unit=0 --speed=115200 [03:26] terminal --timeout=5 serial [03:26] add that at the top ... [03:27] k [03:28] then reboot the machine ... [03:28] during the actual reboot, your minicom should show Offline, and then Online again ... [03:29] then grub asks for a key/selection . [03:29] yes [03:29] hehe [03:29] somehow, i'm really fascinated by that [03:29] i can't tell [03:29] really [03:29] :) [03:31] okay, now have you booted the kernel? [03:32] yes [03:32] try CTRL-a-z-f [03:32] and then h [03:32] SysRq : HELP : loglevel0-8 reBoot tErm kIll saK showMem showPc unRaw Sync showTasks Unmount [03:33] hmm actually CTRL-z should do, but usually I use a screen ;) [03:33] but that is the magic sysreq, you should already know ;) [03:33] hehe [03:34] i use screen too [03:34] all the time [03:34] but i guess i will go to the toilet ... [03:34] before the hard work start [03:34] s [03:35] okay, machine is ready for testing ... if it 'hangs' you should be able to use sysreq-P and sysreq-T to investigate ... [03:35] well, i do not know what to test ;) [03:35] brb [03:36] IIRC, you where able to crash the vs1.23 kernel (compiled with 2.95/96?) more than once, that would be quite interesting [03:48] re [03:48] so [03:48] shall i login and crash the server? [03:48] yeah, but do not login via the console for that purpose ... [03:49] ok [03:49] keep the console for your investigations ;) [03:50] crashed [03:51] okay use the sysreq-p and then sysreq-t [03:51] so [03:51] ctrl-z-f-p? [03:51] you might need to adjust the scrollback buffer first ... [03:52] SysRq : [03:52] screen/keyboard history ... [03:53] ok, i told putty to keep 10000 lines [03:54] hmm, that is possible? [03:54] though somehow i now lost everything inside the windows with minicom :) [03:54] IIRC there are only 4 digits ;) [03:55] lines of scrooback: 10000 [03:55] whatever [03:55] i do not know how i get something out of that window [03:55] /server [03:55] CTRL-a-z-f-h [03:56] nothing [03:56] or better CTRL-a z f h [03:56] yeah [03:56] after f, it tells "sending blank" [03:56] and then i press h [03:56] nothing anymoure [03:56] break I hope ... [03:56] ok [03:56] yes [03:56] hehe [03:57] it disappears too fast to read it properly [03:57] well, that means that your system doesn't handle the magic sysrq ... [03:57] which further points to a hardware issue?! [03:57] mhm [03:57] i had the sysrq stuff displayed before [03:57] maybe putty/.. messed something up? [03:58] could it be that your cpus overheat very fast? [03:58] no [03:58] not that i think so [03:58] i stressed the server from yesterday around 3 until midnight [03:58] 4 paralell kernel compiles [03:58] load avg 3.4-3.9 [03:58] 63 and 61 degrees celcius as i quit [03:59] hmm, strange ... doesn't look like a typical kernel issue to me ... [03:59] btw: fear my proper english: "not that i think so" [03:59] well, ill reboot [03:59] and do that -p again [04:00] minicom was still online [04:00] btw [04:00] as i press reset, it's offline [04:01] hmm, maybe you should try with a non-vserver kernel, and start a fork bomb (or even killer) just to see how that looks like? [04:02] i could try that [04:02] missed the grub screen [04:02] second reset [04:03] change the 5 seconds to 25 on the next boot ... [04:03] ok: ctrl-a-f-break-h -> [04:03] SysRq : HELP : loglevel0-8 reBoot tErm kIll saK showMem showPc unRaw Sync showTasks Unmount [04:03] ./killer [04:04] press 8 [04:04] it hangs [04:04] now? [04:04] without anything? [04:04] no aon the help [04:04] this is now with vanilla or vs1.23 again? [04:04] vs1.23 [04:04] mhm [04:05] Last message repeated 1 time(s). [04:05] ctrl-a-f-break-h is not working anymore?! [04:05] i shall enter: [04:05] ctrl-a-f-break-h [04:05] and then 8? [04:05] or 8 on the minicom help screen [04:05] like after pressing ctrl-a [04:05] well, actually you might add a '8' after ctrl-a-f-break-h [04:06] which sets the loglevel to 8 [04:06] well, i get no response... [04:06] that is the interesting part ;) [04:06] nothing gets sent back on pressing ctrl-a-z-f-h [04:07] okay, lets try with a vanilla 2.4.24 ... [04:07] i have to compile one [04:07] shouldn't take too long ... right? [04:07] but i've got a 2.4.24-ck1, which i will use afterwards [04:07] is that ok too? [04:07] or should it be vanilla [04:07] ah, i have to enable sysreq [04:08] yeah would be better ck1 changes too much ... [04:08] i go for a vanilla one [04:08] this gets a little boring because nothing happens ;) [04:15] well, you can pipe the kernel compile output into xmms, this gives great music ;) [04:17] rebooting [04:19] ok [04:19] ill run the fork bomb [04:19] ok? [04:19] okay ... [04:19] SysRq : HELP : loglevel0-8 reBoot tErm kIll saK showMem showPc unRaw Sync showTasks Unmount [04:19] running [04:19] ./forkbomb.sh: fork: Resource temporarily unavailable [04:19] :) [04:20] what next [04:20] try the killer script ... ignore the fact that it can't change with this kernel ... [04:20] during the forkbomb? [04:20] mhm [04:21] lets see [04:21] :) [04:21] ok [04:21] it "should" be running [04:22] sometomes the "./forkbomb.sh: fork: Resource temporarily unavailable [04:22] is broken by something [04:22] but im not able to see it [04:22] ctx: -1 [04:22] it says [04:23] hmm seems okay .. and it seems stable ... strange ... [04:23] try to activate the sysreq now [04:24] ctrl-z-a-f-h [04:24] SysRq : HELP : loglevel0-8 reBoot tErm kIll saK showMem showPc unRaw Sync showTasks Unmount [04:24] SysRq : Show Regs [04:24] Pid: 1, comm: init [04:24] EIP: 0010:[] CPU: 0 EFLAGS: 00000287 Not tainted [04:24] EAX: 000000ff EBX: f6ac6000 ECX: 00000000 EDX: c031e8c0 [04:24] ESI: cf864000 EDI: 00000000 EBP: c1c3ff18 DS: 0018 ES: 0018 [04:24] so it works here ... [04:24] CR0: 8005003b CR2: 4018646c CR3: 37da1000 CR4: 000006d0 [04:24] Call Trace: [] [] [] [] [] [04:25] yes [04:25] wow [04:25] -t shows _a lot_ of content [04:25] lots of forkbomb.sh [04:25] :D [04:25] hmm ... okay, lets try to start the killer on vs1.23 (gcc 2.95/96 right?) [04:26] via the serial console ... [04:26] ok [04:26] ad your gcc 2.95 question [04:27] any possibility to be real sure that it is 2.95? [04:27] really [04:27] i changed the two lines in the Makefile [04:27] sure, use strings on the kernel and grep for linux [04:27] ok [04:28] strings /boot/vmlinuz-2.4.24-vs1.23-2.95 |grep linux [04:28] linux: fatal error: A20 gate not responding! [04:28] /home/raoul/src/linux-2.4.24/include/asm/spinlock.h [04:28] /home/raoul/src/linux-2.4.24/include/linux/highmem.h [04:28] strings kernel | grep 2.4.24 [04:28] seems better ... [04:29] 2.4.24-vs1.23-2.95 (root@red.ipax.tk) #3 SMP Thu Jan 15 02:27:30 CET 2004 [04:29] /home/raoul/src/linux-2.4.24/include/asm/spinlock.h [04:29] /home/raoul/src/linux-2.4.24/include/linux/highmem.h [04:29] but i named my kernel that way [04:29] hmm, right ... okay [04:29] (2.4.24-vs1.23-2.95 [04:29] the info is there somewhere ... let me check ... [04:31] ah okay, boot it, and use cat /proc/version ... [04:31] ok [04:31] red:/home/raoul# cat /proc/version [04:31] Linux version 2.4.24-vs1.23-2.95 (root@red.ipax.tk) (gcc version 2.95.4 20011002 (Debian prerelease)) #3 SMP Thu Jan 15 02:27:30 CET 2004 [04:31] good to know [04:31] ok [04:31] i'll start the killer from the serial line [04:32] ctrl-a-z-f-p works (nothing started yet) [04:33] i compiled the killer with gcc-3.3 [04:33] but i guess that has nothing to do with that [04:33] or am i wrong? [04:34] still running [04:34] load average: 50.67, 17.04, 6.12 [04:35] starting another killer [04:35] wtf? [04:35] why did it crash before? [04:35] ok [04:35] crash [04:35] hehe [04:35] no respond to ctrl-a z f p [04:36] first killer.c: [04:36] ctx: 60196 [04:36] ctx: 60198 [04:36] ctx: 60199 [04:36] ctx: 60201 [04:36] ctx: 60202 [04:36] ctx: 60204 [04:36] ctx: 60205 [04:36] ctx: 60207 [04:36] ctx: 60208 [04:36] ctx: 60210 [04:36] ct [04:36] --- [04:36] second: [04:36] ctx: 61070 [04:36] ctx: 61071 [04:36]