[08:02] there's a new kernel in hardy-updates, so would be nice to have a d-i image with it === greenwhich_ is now known as greenwhich [09:52] partman-auto-lvm: cjwatson * r203 partman-auto-lvm/debian/changelog: releasing version 27ubuntu1 [10:03] partman-base: cjwatson * r93 ubuntu/ (67 files in 5 dirs): merge from Debian 120 [10:05] partman-base: cjwatson * r94 ubuntu/debian/changelog: releasing version 120ubuntu1 [10:22] hmm, the d-i seems to be using UTC but the chroot sometimes uses the local time (at least user/groupadd log entries do). is there a way to make d-i use the correct time? I think we are having some problems because of that (installer waiting 3h for some reason) [10:23] EEST happens to be UTC+3 [10:30] hm, I guess the problem could be that if the installed machine used to have windows... [10:30] need to investigate further [10:33] it's a rather long-standing bug [10:34] not trivial :-/ [10:34] we haven't had this issue before, so it's probably something to do with the machines in question [12:55] oem-config: cjwatson * r464 oem-config/ (45 files in 10 dirs): Internationalise .desktop files (LP: #233949). === davmor2 is now known as davmor2_dinner === cjwatson_ is now known as cjwatson === davmor2_dinner is now known as davmor2 [14:06] hey all, I'd like to know if anyone else has ssh'd into a d-i recently. [14:08] when I get the three options: start basic install, start expert install, drop to shell; when i select drop to shell it quits immeatiatly [14:08] when selecting expert install and then drop to shell from that menu i am able to execut a shell [14:09] I suspect it's a conflict with upkg refusing to run two instances like dpkg. one during the already running install and one that gets executed when you select 'drop to shell'. [14:09] it's a while since I tried network-console; let me take a look [14:10] I'm not aware of any locking to speak of in udpkg, though [14:10] the drop to shell option executes as upkg and with a force configuration flag on the shell [14:10] ah [14:10] anything interesting in /var/log/syslog? [14:11] i lost that log, there was something in it but I can't recall at the moment [14:12] cjwatson, can you tell me where to find the source (template?) for the 'expert install' menu so I can compare how the shell is being executed? [14:12] apt-get source network-console [14:12] I'm grabbing the hardy mini.iso now and I'll have a look [14:12] Awesome, thanks for searching with me [14:13] oh, you mean the regular menu? [14:13] Yes [14:13] that's in main-menu (and debian-installer-utils for this menu item) though it may be difficult to follow through that if you aren't already familiar with the architecture [14:20] hmm, my local mirror is not so happy [14:22] cjwatson, I'd like to run a test changing this line in network-console "exec udpkg --configure --force-configure di-utils-shell" by removing the udpkg --configure --force-configure [14:22] that certainly won't work [14:22] can you give me a point into the right direction on how to do that. [14:22] no [14:22] di-utils-shell is a package name, not a command [14:22] I see [14:22] but if you want to edit something, just use nano and edit it on the fly [14:23] usually the easiest way [14:23] that is a shell script, nice [14:25] excuse me, there'll be a slight delay here because anna is giving me grief [14:25] that's alright, I appreciate you're advice so far [14:26] In the past I thought that udpkg worked similar to dpkg regarding locking [14:27] I'm going to rebuild a machine now and set it up with network-console so I can grep some logs [14:33] blink, what the heck is going on here [14:33] oh, I'm running into bug 234486 [14:33] yow, best fix that for 8.04.1 [14:34] looks like it happens when you tell net-retriever to look at hardy, hardy-security, and hardy-updates - it doesn't de-duplicate them and ends up trying to install the same package multiple times [14:35] though I guess I can work around it temporarily to figure out your (separate) problem [14:36] :-) [15:04] tbielawa: hmm. Having got to the point where I can actually test this, executing a shell from the network-console menu works just fine for me [15:05] tbielawa: this is in Ubuntu 8.04; are you using the same version? [15:13] *back* [15:13] cjwatson, we are using 8.04 here as well [15:14] ok, in that case I need the syslog from a failure [15:14] cjwatson, just so we're on the same page, you were presented with an ncurses menu with three options, and you selected the third option (execute a shell)? [15:14] correct [15:15] cjwatson, ok cool. I am getting my preseed configured now (lost our settings) [15:15] I will get the logs from syslog as soon as I can for you [15:23] 234486 is stopping me [15:23] I'm getting that error in d-i now [15:24] and you weren't before? I wonder why not [15:24] that is correct [15:24] I'm getting the error by using: d-i preseed/early_command string anna-install network-console [15:25] which obviously calls in the libcrypto for ssh [15:25] anyway, my workaround was as follows: as soon as the installer starts, press alt-f2, press enter, 'nano /usr/lib/debian-installer/retriever/net-retriever', search for codenames, and after the if ... fi block add a line reading: codenames="$codename" [15:25] then alt-f1 and continue through the installation [15:25] I'm afraid this won't work with automatic installation though [15:26] your work around is for libcrypto, or so you could get a shell to execute? [15:26] also be sure to not ssh in over an untrusted network, since the openssl security updates won't be applied that way [15:26] for the libcrypto thing [15:34] cjwatson, i tried applying your workaround and I am still getting the same message [15:35] think you must have made a mistake ... [15:38] entirely possible. Our setup here is very in-house for creating bootable media [15:39] I see my mistake, typo [15:52] cjwatson, i've replicated my error. ssh'ing in as installer and selecting 'Start Shell' the tail of my syslog is at http://pastebin.ubuntu.com/15090/ [15:58] if I touch /var/log/lastlog i get a different error message, but the ssh connection still fails [15:58] http://pastebin.ubuntu.com/15095/ [15:59] OK, thanks; I'll look at that shortly [16:00] and finally (sorry for the information overload) having touched /var/log/lastlog and selecting 'execute a shell' from the 'expert menu' i am able to execute a shell, pastebin: http://pastebin.ubuntu.com/15096/ [16:00] I don't want to bug you any more but I'll stick around in the channel. If there's bug testing I can help you with please let me know. Thanks for taking the time to talk to me this morning [16:12] tbielawa: the lastlog failure is cosmetic; sshd warns about that, but it won't cause a failure [16:15] tbielawa: could you use ssh -vvv to attempt to ssh in, and post the client-side debugging output for me? [16:15] cjwatson, sure, one sec and I'll get that for you === cjwatson_ is now known as cjwatson [16:17] tbielawa: actually, you know what, I'm on crack - ssh is working perfectly well so there's no need for debugging output from it [16:18] tbielawa: are you installing in English, or in some other language? [16:20] We are using English [16:21] to be exact, in our preseed we specify: d-i debian-installer/locale string en_US [16:21] If you'd like to see the preseed I'm using I could give you a link to it [16:22] ok, could you try another test for me? edit /bin/network-console-menu and change the 'set -e' to 'set -ex'; then I need /var/log/syslog again [16:22] preseed probably wouldn't hurt [16:27] syslog coming up [16:35] http://pastebin.ubuntu.com/15101/ [16:36] drat, nothing new there [16:37] an interesting thing happened. having selected expert install, then selecting start a shell, then concurrently opening another shell connecting I was able to select 'start shell' from the main 3 option menu without it failing [16:37] same text in the syslog, however [16:40] could you confirm which image you're using to start with? I'm testing with the netboot mini.iso [16:42] That will be difficult to answer. I will investigate and report back when I figure that out [16:45] just the URL you downloaded it from will be fine [16:46] do you remember a fellow, jeckersb? [16:46] we're using a tool we developed in house in the wvu cs department and I'm not 100% on it's internals yet [16:46] I'm confiring with it's creator now [16:46] I don't think so, and nor do my logs [16:47] though they only date back so far [16:47] oh, are you using a home-built initrd image? that could be relevant [16:47] I believe so [16:48] this software is in need of an overhaul I believe [16:50] it could be that you simply don't have di-utils-shell in the initrd :-) [16:51] unfortunately, the errors sort of go to /dev/null [16:51] hmm, actually, they *ought* to land in syslog [16:51] but maybe don't for some horrible reason [16:52] we're using a netboot iso. it's available here: http://mirror.lcsee.wvu.edu/ubuntu/dists/hardy/main/installer-i386/current/images/netboot/ubuntu-installer/i386/ [16:52] actual iso (i.e. http://mirror.lcsee.wvu.edu/ubuntu/dists/hardy/main/installer-i386/current/images/netboot/mini.iso) or the broken-down initrd and kernel components? [16:53] initrd.gz actually [16:53] have we discovered something? [16:53] no, just trying to clarify [16:53] Alright [16:57] I'll return in a short while === tbielawa is now known as tbielawa_lunch [17:07] well, the md5sum of the netboot/mini.iso on that site is identical to the one I have ... [17:50] hm [17:50] curious === tbielawa_lunch is now known as tbielawa [18:34] back [18:45] cjwatson, did you happen to see my question on figuring out which file fails the 'is the cd good' check? [22:13] xivulon: Hi! I'm curious what your use case for being in ubuntu-bugcontrol is. [22:14] hi bdmurray, from time to time I had to set bug priority for bugs I was playing with outside of wubi (but related to) [22:15] also am planning on expanding my interests [22:15] given time constraints of course [22:16] Okay, that makes sense to me then. [23:47] greearb: yeah, I answered [23:47] 22:09 greearb: mount the CD loopback on (say) /cdrom (or boot the live CD), and run 'cd /cdrom && md5sum -c md5sum.txt'