[00:44] That moment where, after several hours of troubleshooting, you realize that there's no problem with your config, your WiFi adapter is just bad. [00:45] ow [00:45] that's rough [00:50] What's really dumb is if I'd figured it out right off, I could've run down to Fry's, grabbed an Ethernet cable, and set up my old WiFi router as a bridge to the one Comcast has us using that my roommate refuses to replace. [01:25] replace your room mate :P [05:49] I have installed mailutils, postfix, dovecot-imapd, & dovecot-pop3d, followed a some tutorials as to set them up so my server will function as a send-only SMTP server, I can send email from telnet to localhost successfully, yet from the php form hosted on my apache2 server I get no response. Here is the output from netstat: http://dpaste.com/0MGVRVB [05:51] here in /etc/postfix/main.cf : http://dpaste.com/2TD1X44 [05:53] ufw status: http://dpaste.com/0MEG50R [05:55] and /etc/dovecot/dovecot.conf : http://dpaste.com/0592VD8 [07:02] ahasenack: I hit the trace you linked with ustriage rather often, but never with git ubuntu [07:02] ahasenack: afaik I tohught it would be related to my kde & keystore setup [07:03] ahasenack: I was sure it is either that or part of the language settings (lang: en; numbers/formats: german) that make this to hit just me [07:03] ahasenack: there even was a bug about it in the context of ustriage, but we didn't get far IIRC [07:10] Good morning [09:07] hello, I'm testing canonical-livepatch and it says that there is nothing to apply but I have a reboot required with the kernel, I'm using the HWE kernel, is this why? or am I misunderstanding how it works? output of the console: https://apaste.info/FEfI [09:08] from https://www.ubuntuupdates.org/package/core/xenial/main/proposed/linux-image-4.10.0-40-generic, shouldn't the canonical-livepatch service work for CVE-2017-12188 ? [09:19] I cannot find out if canonical-livepatch works for the HWE kernel... any idea? [09:21] Slashman, no (https://assets.ubuntu.com/v1/ac3aa269-DS_Canonical_Livepatch_Service_screen-AW_08.17.pdf <- see system requirements, only 4.4 kernels) [09:22] smb: thanks, that's too bad and make it useless for me... === _ruben_ is now known as _ruben [09:37] So I'm setting up an email server for the first time, running into some expected trouble. I'm on the Ubuntu fourms trying to resolve this, but I figured I would try here also. executing 'echo "test" | sendmail me@gmail.com' returns no errors, but I'm not getting any mail in my gmail account. Here is all the relevant information I could think to gather, http://dpaste.com/16JXRHQ [09:37] Any help would be greatly appreciated, as I am currently stuck. === jelly-home is now known as jelly [10:14] any suggestions on the nicest/cleanest way to add gretap tunnels and bridge them to a bridge on ubuntu, i.e. ifupdown /etc/network/interfaces [10:18] persistently, obviously [12:14] ahasenack: is Launchpad struggling? I'm getting lag. [12:14] Could that be the reason? [12:15] it's fine from here [12:30] rbasak: ok, so this line: [12:30] 11/21/2017 19:47:34 - INFO:Importing patches-unapplied 0.12~pre2.dfsg0-1 to debian/sid [12:30] 11/21/2017 19:48:04 - ERROR:Command exited 1: gbp import-orig --no-merge --upstream-branch do-not-push --pristine-tar --no-interactive --no-symlink-orig '--upstream-tag=importer/upstream/debian/%(version)s.gz' /home/ubuntu/deb/sysstat/gitrepo/.git/git-ubuntu-cache/elinks_0.12~pre2.dfsg0.orig.tar.gz [12:31] "ERROR:stderr: gbp:error: The orig tarball contains .git metadata - giving up." [12:31] rbasak: in the end, 0.12~pre2.dfsg0-1 should exist in the repo I'm creating, right [12:31] besides that error [12:31] I do get that in my run. [12:31] Yes, importer/import/0.12_pre2.dfsg0-1 should exist. [12:31] But the pristine-tar will be missing, which is expected. [12:31] maybe we should add some verbosity to the escaping function. If -v was passed, it could log that it's escaping some .+git directories [12:32] ok, let me run again [12:32] You can see that it escaped by checking out one of the errant import tags [12:32] of course I removed the previous repo [12:33] For example [12:33] git ls-tree importer/import/0.12_pre2.dfsg0-1 [12:33] Contains 040000 tree 4b825dc642cb6eb9a060e54bf8d69288fbee4904 ..git [12:33] However, you don't see it on checkout because the original .git directory was empty. [12:33] rbasak: when I check that out, I'll also have ..git, right [12:33] ah [12:33] So that leads to the git empty dir case. [12:33] But you can see that it's correctly in the tree object. [12:34] rbasak: when we rename .git to ..git, we change the timestamp of that directory? [12:34] (unrelated question) [12:34] git doesn't store timestamps [12:34] * ahasenack checks stat [12:35] https://git.wiki.kernel.org/index.php/Git_FAQ#Why_isn.27t_Git_preserving_modification_time_on_files.3F [12:36] ok [12:37] what was the other package that also had a .git in its tarball? [12:41] It's listed in bug https://code.launchpad.net/bugs/1730778 but I can't reach it right now :-/ [12:41] lockfile-progs [12:43] hm, that one crashed [12:44] http://pastebin.ubuntu.com/26019672/ [12:44] That's interesting. [12:44] I think it's a separate bug though. [12:44] nacc pointed out that our orphan tag handling is wrong. [12:44] But we'd never hit a broken case. [12:44] Now we have :) [12:45] I'll file it, worst case we close it as invalid [12:45] or duplicate [12:45] I'll also check with the snap [12:47] the snap fails before because of the .git issue [12:50] filed [13:07] rbasak: have you seen this crash before? http://pastebin.ubuntu.com/26013467/ [13:08] ahasenack: didn't you have that yesterday? :) [13:08] yes, I didn't remember if I showed it to you back then :) [13:09] I'm going over my todo list for today and that item was there: "file bug somewhere about that crash" [13:09] Feel free to file a bug on it. But I want to focus on the importer right now. [13:09] I'm not sure it's in git-ubuntu, feels like it's not [13:10] rbasak: you can probably merge https://code.launchpad.net/~racb/usd-importer/+git/usd-importer/+merge/334033, unless you want another review [13:10] Thanks! [13:38] rbasak: I finished the local import of elinks, and I'm trying build-source on it now [13:39] rbasak: it's failing like this: http://pastebin.ubuntu.com/26020011/ [13:39] is this because of the pristine-tar failure during import? [13:39] it's a debian sync package, btw, no ubuntu changes [13:40] Seems likely. [13:40] But really these are all separate bugs. [13:41] ahasenack: oh [13:41] I like "oh" :) [13:42] It's probably because you haven't branched the correct pristine-tar names [13:42] Basically you need to rename importer/* -> *. [13:42] Inside refs/heads and refs/tags IIRC. [13:42] is that another step after import finishes? [13:43] I mean, does our cron job in that bastion host do that as a separate step? [13:43] or is that done when the repo is pushed, which I skipped? [13:43] Importing locally and not pushing is an edge case that we don't really support for users - just developers for testing. [13:44] So basically all the refs are wrong because they've been prefixed. [13:44] and the "push" step "fixes" that? [13:44] We did that for safety - to avoid local changes accidentally making it to the published repos, the importer operates entirely inside importer/* [13:44] Yeah. The push un-importer/*-izes, IYSWIM. [13:45] ok, that's good enough [13:45] In the long term we could make this better for the developer testing use case. [13:45] But for now it's a poor UX I admit. [13:48] I'll push to my namespace [13:48] You could just push to another local repo. [13:48] If you want to test. [13:48] Or, as a hack, check out something like a new orphan empty branch [13:48] Then rename from inside .git/refs [13:48] Move everything in .git/refs/importer/ one level up [13:49] If you push somewhere, you have to set the refspec manually. The importer deliberately doesn't use the default refspec. [13:50] ok, so just setting --lp-owner isn't enough? [13:51] and importing again without --no-push? [13:51] What exactly are you trying to do? [13:54] rbasak: I want to verify that after all these manipulations we are doing, when I build a source package out of the git branch I will get the .git directory inside the orig tarball that we generate [13:54] the original .git problematic directory [13:54] I see. [13:54] That won't work at the moment, since I haven't implemented the unescape end in git ubuntu build. [13:55] I intend to open a bug for that and then leave it, as it's unimportant. [13:55] Right now, you would end up with a ..git. [13:55] To test that, I suggest you try my hack above to unprefix all the refspecs. [13:58] is unescaping missing somewhere else? [13:59] that we know of, of course [14:01] I can only think of the build for unescaping and import for escaping. [14:02] ok [14:05] rbasak: let's land it [14:05] just gave a +1 [14:05] Thanks! [15:33] rbasak: did the imports resume? [15:34] after the git-escaping fix [15:35] hi, is there a way to create a new dhcp entry in dhcpd.conf using shell commands? something like webmin but in cli [15:38] ahasenack: I've not updated the bastion yet. Working on the applied branch fix. [15:41] Guest74793: I think there is a dhcp cli [15:42] Guest74793: in the case of isc's dhcp server, it's omshell [15:53] how long does it take for the welcome screen when one does ssh to not show the number of packages that can be updated (meaning, I recently updated packages on another ssh session) [15:54] DammitJim: if you log off then log back in it will be up to date [15:54] oh, how weird [15:55] I guess since it doesn't keep state, one has to kick it [15:58] DammitJim: it's a know bug [16:04] finally found it: LP: #1368864 [16:04] Launchpad bug 1368864 in shadow (Ubuntu) "old motd is displayed on login" [Medium,Fix released] https://launchpad.net/bugs/1368864 [16:33] anybody has openvpn running in a container in bridge mode by any chance? [16:46] powersj: so, you've switched to using priority instaed of debconf/priority, correct? [16:46] yes and I got ISOs published the last two days [16:47] last night things broke again, but due to a perl depend: https://paste.ubuntu.com/26021153/ [16:47] ack [16:47] so, you might want to change netcfg/get_hostname to just 'hostname' as well [16:47] the issue is due to changes in busybox [16:48] are these changes that are official and not bugs? [16:48] an official change to fix some bug, that breaks assumptions make in the installer [16:48] ok thanks for following up [16:50] I'm looking at how to fix this now === jelly-home is now known as jelly [18:04] omshell, oh, thanks i go to see [20:24] Do you have to do anything special to boot off of mdadm with linux-image-generic-hwe-16.04? [20:24] vs linux-image-4.4.0-101-generic? [20:40] kyle__: not that I know of, but you want to make sure that the initramfs was updated with your raid config, altho I'm not even sure that's needed anymore, it should autodetect everything [20:40] what problem are you having? [20:41] drab: I installed with hwe, and it installe dhte 4.4.0 and the 4.10.0(hwe) kernel. If I boot with the 4.4.0 it comes all the way up, if I boot with the 4.10.0, it doesn't. [20:42] I can pull up the error on kvm, but there's no output, no nothing, just dies not being able to locate root on md0 [20:42] do you get dropped into an initramfs / busybox shell? [20:42] oh [20:42] do you need to install an -extra package or similar? [20:42] Yeah. It's curious. [20:43] the generic should pull in extra (as opposed installing virtual which installs minimal) [20:43] root@kh25-65:~# dpkg --get-selections|grep 4.10 [20:43] linux-image-4.10.0-40-generic install [20:43] linux-image-extra-4.10.0-40-generic install [20:43] That extra? [20:43] but yeah, extra should be there, minimal has no raid [20:43] yep, that extra [20:43] update-initramfs -u should run as part of the kernel install process, but maybe worth another try [20:43] ? [20:43] * kyle__ is really hoping for a, "You forgot to install XYZ you idiot" type answer and it will magically fix it. [20:43] I did try that earlier, but I'm willing to try again. OK [20:44] if you did that already that's fine, just running through some ideas [20:44] the only weird thing is the no busybox shell... if it can't find the root it should drop you into one [20:45] I don't get that part, don't think I've seen it before [20:45] You know, I never tried remove it first, update-initramfs -k all -u is muscle memory by now. [20:45] Trying -k all -c first, then the -u [20:46] * drab puts up his "trust in sarnold and everything will be ok" sign [20:46] it seems to help with weird problems [20:47] heh, i'm usually useless at boot / install things, I only ever do that once or twice a decade [20:47] * kyle__ sighs. [20:47] Supermicro boxes take SOOOo long to cycle [20:47] but forgetting to install the -extra package bit me in the ass once, so I remembered that one. [20:48] once I had this woman who wanted to hang a banner with my name next to her screen because she kept calling me for a problem but when I went there it was always fine [20:48] drab: lol [20:48] and she insisted the computer knew I was there and behaved [20:50] drab: you must have a good EMI karma ;) [20:50] lol [20:51] I know all the secret pressure points, like the reset button or the switch on the PSU [20:54] Here's what it says after the fully done update-initramfs [20:54] https://imagebin.ca/v/3iD0zFYx9WDc [20:56] lol, I just tried to pgup the picture :P [20:57] how did you assemble the raid? did you use /dev/sdc or by id or something? [20:57] I'm wondering if for some reason once you use hwe kernel modules the order in which it detects drives changes or somethings [20:57] and throws mdadm off [20:59] /dev/sda /dev/sdb, in debian preseed. [21:00] Well... you give debian preseed the drives you want to use, and it partitions and does all the stuff for you, and you sit there and hope it doesn't decide to do anything horribly weird, because it's debian-preseed. [21:00] if you can wait, which I'm sure you can't, at some point today I can try on my spare SM [21:01] I have a raid1 on two SSDs and disks in the front trays, it's an X9 tho [21:01] they boot fine on generic, can try to install hwe and see what happens [21:01] drab: Might not help, just because this works on the two older generations of the same SM hardware. [21:01] :/ [21:01] ah, ok, so it's an X10 thing, mmmh [21:01] * kyle__ nods [21:01] we're too poor for x10s [21:02] ok, well, the hard way would be to take that initrd.img and unpack it, can't think of anything else to do [21:02] and see that it has the modules and the mdadm stuff [21:02] I don't think it's poor as much as, oh my god we're strapped for space! You there! Fit two full size servers into each U, go do... whatever it is you IT people do [21:02] * kyle__ thinks that's how the conversation to buy these went. [21:04] here's what I mean when I say poor... I built the rack myself with 2x4s :P [21:04] You get extra macguyver points for that though... [21:04] srly tho, X9s on ebay are in a very good spot now, X10s not quite [21:05] Really? OK I need to check that out... [21:05] lol, now if only homedepot let you redeem macguyver points... :D [21:05] that would be a kickass marketing strategy [21:06] afk for a while [21:06] Hrumm..... sounds like a marketing opprotunity. [21:06] but yeah, unpacking initrd.img and checking for modules + mdadm stuff would be my next step [21:06] ttyl, wish u luck [21:08] Thank man. Goodnight! [21:25] How to write shell scripts in linux which control the other applications? I also want to speed up in linux. [21:30] If you're talking about scripting running applications, you need an application that exposes something it can control. [21:30] kyle__: What? [21:30] Generally shell scripts launch other applications with options, not interact with ones already running. The language expect lets you launch and automate text based applications if that's what you want. [21:32] kyle__: So, I can control text based application with shell scripting? [21:32] using expect, yes. But it's automating an application it launches. It is not automating an application that is already running before your script. [21:33] kyle__: OK [21:33] Scripting running applications is highly dependent on the application, and what that application exposes. [21:33] kyle__: Is 'expect' any program [21:34] expectinteracts with stdio only [21:34] standard in standard out. [21:34] I think stderr as well. [21:35] kyle__: Are their any examples online that can help me that how can I control my applications with shell scripting [21:36] Plenty, just search for what you're trying to do in vague terms. [21:36] Or lookup the bash shell scripting tutorial on ldp.org [21:36] wait, wrong site [21:36] www.tldp.org [21:36] That one [21:36] kyle__: OK [21:37] I never had luck with expect except for two cases, (a) ppp chat script (b) using autoexpect to learn the interaction [21:38] kyle__: Are their more or any guidance or tools that you would suggest to use on linux for faster usage of system and also as responxe to user by the system. [21:38] sarnold: I am trying to understand its man page. [21:39] sarnold: I didn't write it, but there's an expect script for start-scumming nethack. It's brilliant [21:40] kyle__: What is start-scumming nethack? [21:41] pankaj_: nethack is an ancient game, one of the first computer games.. https://nethackwiki.com/wiki/Start_scumming [21:41] sarnold: OK [21:43] sarnold: Is their any guide or anything you will suggest (for mode of learning) for anyone learning about ubuntu servers? [21:46] sarnold: OK. BYE. THANKS [21:46] pankaj_: when I was starting out, I read something like a "student guide to unix" which had a nice overview of things.. and then all the manpages on the system, and all the documentation in /usr/share/doc/ , all the howtos on www.tldp.org .. [21:46] heh [21:47] The ubuntu server manual can even sometimes be useful [23:43] Well this is annoying, for one of the servers at work ASUS seems to have only released Windows executables for updating the Intel management engine, I have no obvious avenue for updating the firmware on this motherboard other than briefly running Windows I guess? [23:44] (I suppose that's better than the one for which the manufacturer still hasn't released an update...) [23:49] powersj: still around? [23:50] cyphermox: yep [23:50] powersj: dpb1 : what is your opinion on making the server iso splash have a timeout, and carry on to "Install Ubuntu Server" automatically? [23:51] I propose a timeout of 5 seconds, that matches up with what is already in place for desktop images. [23:51] that automatic install would still then stop at asking the user for a language in the installer [23:52] I feel like the only issue is if someone would want one of the other install options, is 5 sec enough. [23:52] it would just be jumping right into d-i (or subiquity) rather than waiting for input at the bootloader [23:52] Then again I'm not sure the other install options are used very much, given some were broken for a while [23:52] yep [23:53] I'm +1, but I think worth at least asking ubuntu-server list [23:53] well, 5 sec is enough if you consider any keypress cancels the timeout [23:53] if you have a good network connection for those doing remote installs [23:54] as necessary we can put an exception to increase the timeout, but I'd rather burn that bridge when we get there [23:54] ok still worth an email to the server list [23:54] :/ [23:54] heh [23:55] I killed the i386 iso with zero negative responses [23:55] keithzg: btw, will the intel microcode firmware package be the -entire- update when it is published? [23:55] powersj: timeouts can quickly because bikeshed [23:56] well, how about we change it then and see if anyone even notices or cares [23:56] bah, I'll send an email to the list [23:57] might be agreeably surprised