[01:38] <nacc> ahasenack: just fyi, the import is still going, it's done with debian (and patches-applied of debian), doing ubuntu patches-applied now (up to lucid), so it'll probably finish sometime this evening (my time)
[03:51] <jak2000> why cant do an apt update?   https://pastebin.com/uBNFwv7w
[07:42] <lordievader> Good morning
[09:42] <jsheeren> hi
[09:42] <jsheeren> can i force a process to take over a port still in use by a defunct zombie process?
[11:46] <mdeslaur> rbasak: is someone working on mysql-5.7 5.7.19 in artful?
[11:47] <rbasak> mdeslaur: blocked on me at the moment :-/
[11:47] <mdeslaur> rbasak.clone()
[11:48] <rbasak> I've still not reviewed Skuggen (not here)'s merge of 5.7.18 :-/
[11:48] <rbasak> It's on my list to spend some time catching up on MySQL properly.
[11:49] <mdeslaur> ok, thanks
[11:49] <rbasak> mdeslaur: go ahead and bump the upstream to 5.7.19 in artful I guess?
[11:49] <mdeslaur> rbasak: sure, I can do that
[11:49] <rbasak> mdeslaur: you might find https://anonscm.debian.org/cgit/pkg-mysql/mysql.git/log/?h=mysql-5.7/lars/ubuntu helpful.
[11:49] <rbasak> That was for 5.7.18.
[11:50] <mdeslaur> I was trying to get someone else to go through autopkgfail hell :)
[11:50] <rbasak> Oh, but that's already there.
[11:50] <rbasak> https://anonscm.debian.org/cgit/pkg-mysql/mysql.git/log/ is on 5.7.19 if that helps.
[11:50] <rbasak> It should be almost identical, apart from the mecab component delta.
[11:51] <mdeslaur> I'll just do what I did for the zesty security update, and it can get replaced with a proper sync afterwards
[11:51] <rbasak> Sure. Thanks!
[12:10] <christos_> What is the difference between "crontab -e" and "/etc/crontab"?
[12:11] <christos_> I am trying to setup some jobs to run and not sure what is the best place to put what
[12:14] <nav-> christos_: nothing
[12:14] <nav-> 'crontab -e' opens file /etc/crontab for editing using your default text editor
[12:15] <nav-> '/etc/crontab' is a file that stores your crontab entries
[12:20] <rbasak> No.
[12:20] <rbasak> "crontab -e" edits your user's crontab entry.
[12:20] <rbasak> /etc/crontab is the system's crontab entry and is a different syntax (it has an additional user field).
[12:20] <christos_> nav-: is there any reason why I can't just vim /etc/crontab? Is that bad to do?
[12:21] <rbasak> christos_: unless you want them to run as a specific ordinary user, editing /etc/crontab is fine.
[12:21] <christos_> Also, I want to get an email after the cronjob finishes, how do I do that? I tryed adding MAILTO my /etc/crontab and then putting a script in cron.hourly to test but it ran once and then never again?
[12:21] <rbasak> christos_: if you want them to run as your own user, then you can still use /etc/crontab, though I'd use "crontab -e" as I feel that's cleaner.
[12:22] <nav-> christos_: no, nothing is stopping you from just editing /etc/crontab in vim
[12:22] <christos_> rbasak: When I make an entry in crontab -e and look in /etc/crontab, it does not show what I added?
[12:22] <rbasak> christos_: correct. Because they are different sets of entries.
[12:22] <rbasak> One is for your user. The other is for the system.
[12:23] <christos_> is there a location for what I am editing with crontab -e
[12:23] <christos_> Like i can go to that file somewhere?
[12:23] <rbasak> /var/spool/cron/crontabs, but you aren't supposed to edit that directly.
[12:24] <rbasak> I expect you'll have sync issues if you do.
[12:24] <rbasak> You may edit /etc/crontab directly, but for user crontabs, you should go through the crontab command.
[12:25] <christos_> rbasak: I understand that now thanks
[12:26] <nav-> christos_: why not just use 'crontab -e' and tell it to use 'vim' to edit it? and then use 'crontab -l' to list your entries to verify it worked
[12:26] <christos_> and "cat" the /var/spool/cron/crontab for the users to see what is what.
[12:27] <rbasak> christos_: as nav- says, use "crontab -l"
[12:27] <christos_> nav-: That is what I was doing before, I was just curious about the difference
[12:27] <nav-> gotcha
[12:28] <christos_> Thanks guys! Is there a way to get emails when a job runs?
[12:29] <nav-> sure. have a script that sends one and schedule it to run at the end of your cronjob
[12:29] <christos_> there is a site called cronguru where they help with the schedule spacifics and you can add "&& curl -sm 30 k.wdt.io/<email-address>/<cronjob-name>?c=5_4_*_*_*" to your job and it will email you but I don't want to use their service
[12:30] <christos_> nav-: that sounds easy enough. thanks!
[12:30] <nav-> np
[12:30] <christos_> I collect comics and it is easy to get my computer to do all my collection leg work than me.
[12:32] <nav-> once you get past the frustrating part of setting everything up, it certainly will be
[12:35] <rbasak> cron automatically emails the output of every job to whatever MAILTO is set to. So for a basic email you just have to have the job output something, and you should get it. This is subject to the system mail daemon being correctly configured. If it's not, that's worth fixing anyway, as there are other things on the system that also expect to be able to email the administrator.
[12:37] <coreycb> jamespage: senlin switched to using the u-c version of python-docker (2.4.2) but i don't think we want to sync a new python-docker without a new docker-compose. i'm going to add a minor patch to senlin's docker driver to use python-docker 1.9.0 which is what we currently have in the archive.
[12:38] <jamespage> coreycb: ok
[12:38] <christos_> ok, last question for the day. I am sending an email and I know how to add an attachment to my email, but is there a way to cat the contents of the file to my email?
[12:39] <christos_> Oh. I got it guys
[12:39] <christos_> "< filename"
[12:39] <christos_> Thanks for everythign nav- and rbasak
[12:39] <nav-> or you can pipe the output of 'cat file' as the input to the email program
[12:40] <nav-> but i dont want to make it more complicated than it needs to be, no problem
[12:57] <ahasenack> smb: hey, around?
[12:59] <smb> ahasenack, yes, but give me a sec
[13:00] <ahasenack> ok
[13:01] <smb> ahasenack, ok now
[13:01] <ahasenack> smb: ok, it's about this bug: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1707400/comments/12 I'm linking to my comment where I believe i sorted it out (confirmed by the reporter later on)
[13:02] <ahasenack> smb: tl;dr if you look at the attached dmesg.txt file, you'll see apparmor denied errors, preventing libvirt-bin from connecting to a socket
[13:02] <ahasenack> smb: I debugged it, and I have two possible patches:
[13:02] <ahasenack> smb: a) http://pastebin.ubuntu.com/25219930/
[13:02] <ahasenack> smb: b) http://pastebin.ubuntu.com/25219940/
[13:02] <ahasenack> ubuntu > xenial is using (a)
[13:03] <ahasenack> but I found it odd that the dh_apparmor lines were there in d/rules but commented, maybe there was a good reason back then. git log didn't help me find any
[13:04] <smb> ahasenack, historic reasons that up to a certain point we were not using apparmor profiles in examples/libvirt but some copies under debian/apparmor (or so)
[13:04] <ahasenack> smb: I tested (a) in my broken system, and libvirt installed just fine then
[13:04] <ahasenack> I wonder if the code in (b) should even still exist
[13:04] <ahasenack> it still exists in artful
[13:05] <ahasenack> where postinst calls apparmor_parser twice then: once like in (b), and later on again via the snippet introduced by dh_apparmor
[13:05] <smb> Christian and I were working on moving back to what debian does. which is to use dh_apparmor helper and that generates a lot of the maintscripts
[13:05] <ahasenack> yeah, in terms of a minimum change for an SRU, dh_apparmor sounds more intrusive
[13:06] <ahasenack> in > xenial, where we use dh_apparmor, the service changed names even
[13:06] <ahasenack> it's now a proper systemd job iirc
[13:06] <smb> I thought Christian dropped our mainscripts patches in artful in favour of using dh_aa
[13:06] <ahasenack> libvirt-system-daemon or something
[13:06] <ahasenack> yeah man, talk about patches
[13:06] <ahasenack> $ find debian/patches/ -type f | wc -l
[13:06] <ahasenack> 83
[13:07] <smb> Its a long and slow process to get rid of them again
[13:07] <ahasenack> smb: anyway, I think we need -T and -W in the apparmor_parser call, otherwise certain upgrades just fail
[13:08] <ahasenack> I wasn't aware of that apparmor cache before debugging this bug
[13:10] <smb> ahasenack, I would tend to do the postinst change in Xenial and Zesty if needed. Because there we modify postinst and either need to remove things as well or have duplicate steps when enabling dh_apparmor
[13:11] <ahasenack> smb: >=yakkety are fine wrt the bug
[13:11] <ahasenack> apparmor_parser is called correctly there
[13:11] <ahasenack> via dh_apparmor
[13:11] <smb> ah ok
[13:11] <ahasenack> not that existing snippet in postinst, that one still uses just -r
[13:12] <smb> Rigth just done unnecessary twice. and with artful the snippet should be gone
[13:12] <ahasenack> correct, done twice: incomplete the first time, fully correct the second time because of dh_apparmor
[13:12] <ahasenack> same in artful
[13:13] <ahasenack> it's still there, we could probably remove it
[13:13] <smb> ahasenack, just to be sure, you know that libvirt is kept in a lp git tree
[13:13] <ahasenack> smb: I had a hint yesterday
[13:13] <ahasenack> smb: I'm indeed just checking the package in artful, and its imported git tree into that usd-importer-team
[13:14] <smb> ahasenack, ok, latest artful might be a spearate branch too
[13:15] <smb> ahasenack, Christan has a habit of doing merge branches so latest is artful-3.5
[13:15] <ahasenack> oh, and I think you are right, the artful package just uses dh_apparmor
[13:15] <ahasenack> I just fetched it with pull-lp-source
[13:16] <ahasenack> no apparmor_parser call in debian/blabla.postinst
[13:17] <smb> Ah ok, so more what I expected. Could be that the importer needs hinting with multiple branches.
[13:17] <ahasenack> cool
[13:17] <ahasenack> so nothing to do in artful
[13:17] <ahasenack> I asked for the importer to run yesterday after I noticed the xenial branch was out of date
[13:17] <ahasenack> I didn't check artful again, let me refresh my tree, the new import run should be done now
[13:19] <smb> Xenial has only one branch but zesty and artful each 2. One without version and one with when merging a new upstream version.
[13:20] <smb> personally I would get rid of the versioned branches or at least reset the unversioned ones but not consider this my personal space ;_)
[13:20] <ahasenack> ok, artful git branch is good
[13:20] <ahasenack> so it's just xenial
[13:21] <ahasenack> maybe trusty
[13:21] <ahasenack> and zesty is calling it twice, no harm
[13:22] <ahasenack> I can attach these two patch suggestions to the bug
[13:22] <ahasenack> christian is back next week?
[13:22] <smb> ok, yeah, so for x and possibly t I would do the direct postinst snippet change as I am not sure how well debhelper plays together with what we got there
[13:22] <smb> I believe so
[13:24] <smb> ahasenack, Oh actually for T I would be careful and verify that aa has a cache back then already
[13:24] <ahasenack> yep, saw something about these -T -W options being "new"
[13:26] <ahasenack> it probably does, otherwise this bug wouldn't have happened
[13:37] <christos_> how can I add `date '+%m-%d-%Y' to the subject of my email that doesn't show up as '08-01/2017' and the actual string
[13:38] <christos_> "8-01=2017 Report" instead of "date '+%m-%d-%Y Report"
[14:34] <ahasenack> smb: a default trusty install: http://pastebin.ubuntu.com/25220361/
[14:34] <ahasenack> -T -W is used already
[14:35] <smb> ahasenack, hm, so we lost this over time. bah! but at least one less release to worry about
[14:36] <smb> oh not for libvirt maybe
[14:36] <ahasenack> right, not libvirt
[14:36] <ahasenack> that's just the name of my container
[14:36] <ahasenack> I tend to create a lot, one per bug I'm working on, or worked on
[14:36] <ahasenack> $ lxc list|grep PERSISTENT|wc -l
[14:36] <ahasenack> 58
[14:36] <ahasenack> :)
[14:37] <ahasenack> smb: can you accept nominations? https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1707400
[14:37] <smb> I was not looking directly at the pastbin but the comment here sounded like it was already done for libvirt.
[14:37] <smb> ahasenack, I think I can for that package
[14:37] <smb> ahasenack, done
[14:37] <ahasenack> smb: thx
[15:19] <nacc> ahasenack: fyi, libvirt appears to be fully imported now
[15:40] <ahasenack> nacc: it is, thx
[15:40] <ahasenack> got a whole bunch of new stuff
[15:42] <nacc> ahasenack: :)
[15:42] <sdeziel> I am wondering if someone is working on the security fixes for PHP to address CVE-2017-922{4, 6, 7, 8, 9}?
[15:42] <ahasenack> sdeziel: try the #ubuntu-hardened channel, that's where the security team lingers
[15:43] <sdeziel> ahasenack: thank you
[15:43] <ahasenack> np
[15:46] <ahasenack> nacc: I got this during git ubuntu build-source:
[15:46] <ahasenack> $ git ubuntu build-source --sign
[15:46] <ahasenack> 08/01/2017 12:14:50 - INFO:New upstream version detected (1.2.2) which is after the last published upstream version (1.2.16). Falling through to just calling dpkg-buildpackage.
[15:46] <ahasenack> d/changelog has 1.2.2
[15:47] <ahasenack> I don't know where it got 1.2.16 from
[16:02] <nacc> ahasenack: this was to build an SRU?
[16:03] <ahasenack> yes, I was on a xenial branch
[16:03] <ahasenack> I wonder if it did ascii comparison/sorting
[16:06] <nacc> ahasenack: file a bug please, i'll try and get it fixed
[16:06] <ahasenack> k
[16:06] <nacc> ahasenack: xenial branch of libvirt?
[16:07] <ahasenack> nacc: yes
[16:07] <nacc> hrm, shouldn't that be 1.3.1 someting?
[16:08] <ahasenack> sorry, trusty
[16:08] <ahasenack> xenial worked
[16:08] <ahasenack> I was on trusty now, and that's where I saw that output
[16:08] <nacc> ahasenack: ah ok, yeah, i'm not sure
[16:49] <nacc> rbasak: oh i meant to ask in a comment in the blog draft -- do you have anchors in your post?
[16:50] <nacc> rbasak: so, e.g., i can refer to a particular section
[16:51] <rbasak> nacc: no but I expect I can add them.
[16:51] <nacc> rbasak: ok, that'd be handy for my referencing back to your post
[16:51] <nacc> (and e.g., to point directly to the index section)
[16:59] <rbasak> nacc: http://www.justgohome.co.uk/blog/2017/07/developing-ubuntu-using-git.html#multi-post-index
[17:00] <nacc> rbasak: thanks! and one for the experimental status, too, if you can :)
[17:00] <rbasak> ack
[17:02] <rbasak> nacc: done: http://www.justgohome.co.uk/blog/2017/07/developing-ubuntu-using-git.html#experimental-status
[17:03] <nacc> rbasak: excellent, tyvm
[17:04] <oerheks> nacc, copy that to #ubuntu-news too ;-)
[17:24] <teward> anyone got any experience with postfix and mail relaying? I've got a case where some messages need to just go straight out via SMTP in some cases and the rest need to be relayed to another system before being sent out over the Internet, a little confused to set it up
[17:47] <sdeziel> teward: you can define "routing" rules using transport: http://www.postfix.org/transport.5.html
[17:55] <nacc> ahasenack: yep, it's a bug, i've got the fix staged
[17:55] <nacc> ahasenack: let me know when/if you file it
[17:55] <ahasenack> nacc: ah, cool, do you still need the bug? I just came back from lunch
[17:55] <ahasenack> will do
[17:56] <nacc> ahasenack: thanks
[17:57] <nacc> ahasenack: once rbasak approves of the sru-versioning-fixes MP for the linter, i can push this fix in on top (it's easiest to do it that way, as i'm using a bit of shared code that is only shared after the changes in that branch). Alternatively, I can cherry-pick that across and just fix the bug; rbasak any opinion there?
[17:58] <ahasenack> nacc: https://bugs.launchpad.net/usd-importer/+bug/1707983
[18:00] <teward> sdeziel: thanks
[18:00] <sdeziel> np
[18:41] <teward> sdeziel: one more question: can a 'routing' rule be set to just send to the actual destination, while all other mail is relayed to a different server?  That is to say, a straight SMTP outbound connection rather than relaying in certain circumstances (as transport destinations for regexp header matches for example)
[18:43] <sdeziel> teward: in that case, I'd use a global relayhost and would put just one entry in your transport_maps with the domain you want to be sent directly
[18:46] <teward> hmm, OK, i'll have to think this through, we are trying to special-case handle out of office automatic replies on Exchange at postfix, so that we can have it go right out instead of pissing around with mcafee web gateway eating those.
[18:46] <teward> so it'd be pretty much "Send to expected destination" for any matching subject, relay otherwise.
[18:46] <teward> not *as* simple a config :p
[18:48] <sdeziel> right, a bit more involved then ;)
[18:48] <teward> yup.  any tips are helpful.  :)
[18:50] <sdeziel> have you tried this: https://serverfault.com/questions/679975/postfix-selecting-relay-host-based-on-from-mail-header-rather-than-envelope-se
[18:51] <teward> sdeziel: yeah that's what i'm working on now, gonna run a test shortly to see what happens and if anything explodes in my face xD
[18:52] <sdeziel> teward: a header_checks can return "FILTER transport:destination" to have it pick a specific transport
[18:53] <sdeziel> this would remove the need from using a transport_maps
[21:22] <hashwagon> Anyone know how I'd add a countdown timer on the grub menu of a bootable ubuntu server install USB? I can't seem to find documentation on it, I just want it to kick off an install automatically.
[21:27] <oerheks> edit  /etc/default/grub change GRUB_TIMEOUT=<seconds> sudo update-grub
[21:27] <oerheks> or -1 for indefinitly ?
[21:29] <hashwagon> On the root of my USB I don't have /etc/default I have /boot/grub/grub.cfg am I missing something?
[21:30] <hashwagon> I have /isolinux as well
[21:30] <oerheks> err, the server installer is not a live iso
[21:30] <oerheks> this applies to an install only
[21:32] <JeevesMoss>  has anyone hit this bug yet when trying to install mongoDB?  "insserv: warning: script 'K01mongodb' missing LSB tags and overrides"
[21:32] <nacc> JeevesMoss: it's just a warning and can be ignored, iirc
[21:33] <hashwagon> Ahhh, I was hoping there was something I could put at the end of the grub.cfg line: linux /install/vmlinuz [...] timeout=0 quiet ---
[21:34] <JeevesMoss> nacc, I can't get the dumb server to start.  :-(
[21:35] <nacc> JeevesMoss: what are you trying? exact command and output (in a pastebin)
[21:37] <drab> hashwagon: you should have a isolinux.cfg
[21:37] <JeevesMoss> nacc, https://pastebin.com/f77WnL0n
[21:37] <drab> there are two options in there, prompt and timeout
[21:37] <drab> hashwagon: those should do what you want
[21:37] <nacc> JeevesMoss: that's not an ubuntu pacakge, you should contact the repository owner
[21:37] <drab> since it's a bootable stick the menu etc should be the isolinux one, not the grub one
[21:37] <nacc> JeevesMoss: what version of ubuntu are you on?
[21:37] <hashwagon> drab, okay I do have isolinux.cfg, I'll try that now
[21:37] <JeevesMoss> nacc, 16.04
[21:37] <drab> hey nacc .o/
[21:38] <nacc> JeevesMoss: you definitely don't want some pacakge with the word 'upstart' in it then :)
[21:38] <drab> hashwagon: it should already have in it prompt 0 , timeout 0
[21:38] <nacc> *some repo
[21:38] <nacc> drab: heya
[21:38] <drab> hashwagon: or something like that
[21:38] <drab> hashwagon: http://www.syslinux.org/wiki/index.php?title=SYSLINUX#TIMEOUT_timeout
[21:39] <drab> hashwagon: http://www.syslinux.org/wiki/index.php?title=SYSLINUX#PROMPT_flag_val
[21:39] <hashwagon> drab, I do have both at zero. My isolinux.cfg also includes: include menu.cfg default vesamenu.c32
[21:39] <drab> so change prompt to 1 and timeout to 10 or something
[21:39] <hashwagon> I'll give it a shot thanks
[21:39] <drab> hashwagon: that's normal, and those files can override those variables, but iirc (and you can check grepping), none do
[21:39] <JeevesMoss> nacc, time to go find the correct package
[21:39] <nacc> JeevesMoss: iirc, there is a mongodb-server package in ubuntu
[21:40] <drab> hashwagon: actually my prompt.cfg does override prompt to 0 again
[21:40] <drab> so you may want to change that too just in case
[21:40] <drab> if it gets included
[21:40] <drab> but it shouldn't
[21:41] <JeevesMoss> nacc, thanks.  purging that old junk.   I'm trying to install GenieACS, and the maintainers script is a MESS
[21:41] <nacc> JeevesMoss: good luck!
[21:42] <oerheks> JeevesMoss, lots of new commands to handle mongodb, maybe this page is any help https://www.digitalocean.com/community/tutorials/how-to-install-and-secure-mongodb-on-ubuntu-16-04
[21:42] <oerheks> * services
[21:48] <JeevesMoss> oerheks, looks like I got it.  time to reboot and see if it comes back up!
[22:03] <CodeMouse92__> Ooooookay, not sure how syslog got to be nearly 2G, but I think it needs cleaning
[22:12] <tomreyn> it probably says why in this very file ;)
[22:18] <hashwagon> drab, I appreciate your help. I'm using preseed and for some reason it's booting using /boot/grub/grub.cfg.. I don't remember why since it's been months since I've set it up. Anyway, adding set timeout and set default resolved the issue.
[23:21] <drab> hashwagon: ah, good to know, thanks for circling back