[04:48] <fishcooker> how to show changelog on every ubuntu package update; ie: http://vpaste.net/YnNXo ?
[05:37] <cpaelzer> good morning
[07:07] <sobukus> https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1587142 … I wonder if suddenly something happens on that one.
[07:07] <sobukus> If I should count on this being fixed in 16.04 some time or if I should make a local hack.
[07:07] <sobukus> Servers not being able to reboot is not nice.
[08:39] <Pjusur> Good Morning ppl, is there a way to make journald not pick syslog messages from a specific services? just syslog to rsyslog
[08:45] <necrophcodr> If I do `SetEnv SOMETHING "else"` and `PassEnv SOMETHING` in a virtual host, then that variable still isn't accessible in, say, a .htaccess file with RewriteCond %{ENV:SOMETHING}
[08:45] <necrophcodr> How can I make that happen?
[08:45] <necrophcodr> It seems that SetEnv and PassEnv work great for CGI or FCGId based stuff, but has no affect on .htaccess files.
[09:25] <trippeh> hm. unbound needs some work in artful. it is set up as type notify in systemd, but the apparmor profile doesnt allow it. if you allow it, reloads are still broken, probably a sd_notify missing somewhere.
[09:33] <sobukus> OK, I'm down to 'only' a RAID resync on each boot, from endless hang on any reboot/shutdown attempt.
[09:34] <sobukus> Please, Ubuntu folks, implement the proper strategy of RAID array build-up/tear-down as indicated in the linked fedora issues for https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1587142 .
[09:34] <sobukus> Or bug mdadm folks to push mdmon into the kernel to avoid the trigger for this whole mess.
[10:18] <RoyK> necrophcodr: it's no difference between running apache on a vm compared to running it on metal - I guess something is messed up in the apache config
[10:21] <necrophcodr> RoyK, i'm not refering to virtualization
[10:21] <necrophcodr> I'm refering to <VirtualHost>
[10:22] <necrophcodr> I guess I should've specified it was regarding Apache2 setup :)
[10:22] <RoyK> ah - I see
[10:23] <RoyK> well, that was pretty obvious, but I read it as you had a problem on a "virtual machine"
[10:23] <RoyK> necrophcodr: not sure if there are much apache gurus here, though
[10:24] <necrophcodr> I'm not sure where else to ask about apache stuff though
[10:24] <RoyK> you may want to try #httpd
[12:43] <hhee> good time. mod in apache2 in ubuntu (16.04) shown in packages list as libapache2? or libapache2-mod? i need to find out numbers ands names of installed apache2 modules
[13:06] <ahasenack> hhee: libapache2-mod-* seems to be the norm
[13:11] <hhee> ahasenack, got it
[13:41] <sobukus> Ahw, maaaaan. The latest kernel update broke Infiniband RDMA here. Works with 4.4.0-83-generic, broken with 4.4.0-93-generic. What changed?!
[13:42] <sobukus> I thought that is why we use LTS!
[13:43] <sobukus> apt-get changelog is not helpful.
[13:44] <sdeziel> sobukus: take a look at /usr/share/doc/linux-image-4.4.0-93-generic/changelog.Debian.gz
[13:48] <sobukus> Ah, thanks. I realise that there is a massive amount of changes between these two. I guess I have to test some versions inbetween.
[13:48] <sobukus> I don't see an overt reference to infiniband.
[13:52] <sobukus> Hm, I need to hunt down 4.4.0-84
[13:53] <sdeziel> sobukus: the Xenial -proposed kernel (-94) has some fixes related to RDMA so you might want to take a look there as well
[13:56] <sobukus> That might be a better direction.
[14:04] <sobukus> sdeziel: Ah, magic! With -94, rdma works again.
[14:04] <sdeziel> glad to hear that
[14:05] <sobukus> Now I need to figure out if I can reboot the servers without upsetting clients …
[14:20] <madLyfe> so does 17 become the LTS? or is it a different #? after 16.
[14:20] <ahasenack> madLyfe: 18.04 will be the next LTS
[14:21] <madLyfe> which is just everything from 17, just in stable form?
[14:22] <sobukus> madLyfe: That's the idea;-)
[14:22] <sobukus> LTS every two years … in between the big rockin' changes.
[14:22] <genii> madLyfe: LTS releases are the first release in even numbered years. So 14.04 16.04 18.04
[14:23] <madLyfe> that makes it easy
[14:23] <madLyfe> so in the 16 installer you can select HWE installer. is this default in 17?
[14:26] <M3mphiZ> After enabling PAM authentication using private-key and google-authenticator using this tutorial (https://www.digitalocean.com/community/tutorials/how-to-set-up-multi-factor-authentication-for-ssh-on-ubuntu-16-04), my gitlab instance doesn't accept any pull/push anymore. how can i tell it to not use pam with the git user?
[14:26] <sdeziel> madLyfe: HWE kernels are only available to LTS versions
[14:28] <madLyfe> oh
[14:28] <madLyfe> HWE for server is the future though, correct?
[14:33] <sdeziel> not only servers, no. https://wiki.ubuntu.com/Kernel/LTSEnablementStack
[14:38] <madLyfe> its the future stack for all releases is what i mean
[14:41] <sdeziel> it's a backport of a more recent kernel (and/or X stack) for a LTS release
[14:42] <sdeziel> in otherwords, you get the kernel of a non-LTS readily installable on your LTS
[14:43] <madLyfe> but im saying, in like the next LTS(18) will HWE be the default in installer and not how GA kernel is right now?
[14:46] <sdeziel> "Server installations will default to the GA kernel and provide the enablement kernel as optional." from the link above
[14:46] <madLyfe> ya but thats for 16. what about 18?
[14:47] <madLyfe> im asking if the plan is to move everything to HWE?
[14:50] <sdeziel> HWE used to be the default for ${LTS}.2
[14:50] <sdeziel> now it's the default only for desktop installs
[14:50] <sdeziel> for servers, the default is the GA kernel
[14:55] <madLyfe> are they planning on changing the default for server to HWE?
[15:08] <nacc> madLyfe: 18.04 hasn't started yet (17.10 isn't out yet). I don't believe there are plans to change the above
[15:11] <madLyfe> so they will just keep offering HWE during install?
[15:11] <nacc> madLyfe: i assume so
[15:11] <madLyfe> thats confusing
[15:11] <nacc> madLyfe: why?
[15:12] <madLyfe> i mean.. is there a reason not to run HWE over GA?
[15:12] <nacc> madLyfe: yes
[15:12] <nacc> madLyfe: if you don't need it
[15:12] <nacc> it's more change to audit
[15:12] <nacc> kernels have bugs too
[15:57] <nacc> powersj: are artifacts of jenkins jobs saved at all (e.g., the snap)
[15:57] <powersj> nacc: they can be, but only a certain number
[15:57] <nacc> powersj: ok, np
[15:58] <nacc> powersj: just wondering
[19:55] <madLyfe> do you have to install grub?
[20:19] <madLyfe> i cant figure out why this happens after i install server to a microSD card and then reboot after install to boot for the first time: https://usercontent.irccloud-cdn.com/file/QakJoRAR/irccloudcapture683196535.jpg everything works fine when i install to a USB stick.
[21:40] <nacc> madLyfe: is it actually booting from the microsd card?
[21:42] <madLyfe> i dont know. when i restart from the install, i pull the installer usb stick and its just the mSD card and thats the msg i get right when it would show the ubuntu boot process or grub
[21:43] <madLyfe> if i pull the SD card it just says cant find boot drive or whatever. thats normal though because there is nothing there to boot to.
[21:44] <nacc> madLyfe: does `ls` from the grub rescue prompt work? does it display anything?
[21:47] <madLyfe> nacc: https://usercontent.irccloud-cdn.com/file/EUlEjkxt/irccloudcapture365147137.jpg
[21:48] <nacc> madLyfe: try `ls (hd0,msdos1)/`
[21:48] <nacc> madLyfe: tab complete *should* work, as well
[21:48] <nacc> madLyfe: my guess is it was unable to read or find a config file
[21:52] <madLyfe> not sure what you mean by tab complete https://usercontent.irccloud-cdn.com/file/mHopvUFC/irccloudcapture1198195483.jpg
[21:54] <nacc> madLyfe: you can hit <tab>
[21:54] <nacc> madLyfe: so that's your root fs
[21:56] <nacc> madLyfe: can you do something like `root (hd0,msdos1)`
[21:56] <nacc> madLyfe: and then `configfile /boot/grub/grub.cfg`
[21:58] <madLyfe> sorry had a phone call
[21:58] <madLyfe> 'root' unknown command
[21:59] <nacc> madLyfe: ok, just hit tab and see what commands are available
[21:59] <madLyfe> tab does nothing
[21:59] <nacc> madLyfe: and then google around for the commands
[21:59] <nacc> madLyfe: ok
[22:02] <madLyfe> hmmm help is an unknown command but says use that to get a complete command list
[22:02] <madLyfe> https://help.ubuntu.com/community/Grub2/Troubleshooting
[22:04] <nacc> madLyfe: probably because it's in the recovery mode
[22:04] <nacc> madLyfe: rescue mode, i mean
[22:04] <nacc> madLyfe: it seems like it read past the end of the disk while trying to laod itself
[22:06] <madLyfe> ive read where some ppl have this issue with partitions on large disks but this is only a 16gb sd card
[22:06] <nacc> madLyfe: sd cards may not be well supported as boot media on this machien?
[22:06] <nacc> madLyfe: not sure
[22:07] <nacc> powersj: do you have a log handy of the failed `quilt` ?
[22:07] <nacc> *quilt push
[22:07] <madLyfe> actually i talked to HPE and they said the only thing that will boot on this blade is the SD card and not USB. so far its the opposite.
[22:07] <nacc> powersj: at what version it happened at, i mean
[22:07] <nacc> madLyfe: fun :)
[22:08] <madLyfe> but the sd card would ride the usb chip?
[22:08] <madLyfe> just thinking if i turned off something
[22:08] <madLyfe> cuz i def turned off the network stuffs and probably the sata controller
[22:08] <madLyfe> scsi stuff
[22:08] <nacc> madLyfe: did you need to turn them off?
[22:09] <madLyfe> well for a mining server, saving power is key.
[22:09] <madLyfe> turn off everything that isnt needed.
[22:09] <powersj> nacc: whatever master was last Friday
[22:09] <powersj> https://jenkins.ubuntu.com/server/job/git-ubuntu-ci/55/console
[22:10] <nacc> powersj: ok, that's fixed then
[22:10] <nacc> i'll let the job finish and merge it
[22:10] <nacc> madLyfe: functionality seems more key first
[22:10] <nacc> madLyfe: :)
[22:10] <madLyfe> i wasnt ever planning to use sd card though so not using it isnt a huge deal but would save a few bucks per server if i could get it to work.
[22:11] <madLyfe> i wouldnt need a usb hub then.
[22:11] <madLyfe> $10 per server. adds up.
[22:12] <madLyfe> but at the top is says 'attempting boot from usb drivekey' and then shows the dirs so it must be using the usb chip
[22:16] <madLyfe> The proper paths must be set via the set prefix command.
[22:16] <madLyfe> Many GRUB 2 commands will not work until the correct path is set.
[22:16] <madLyfe> If the path to the grub folder (normally /boot/grub) is not correct, an unknown command or file not found message is likely.
[22:17] <madLyfe> i have to jet for a bit
[22:17] <madLyfe> thanks
[22:20] <elico> *: Maybe anyone knows why the bug fix is not being backported? https://bugs.launchpad.net/ubuntu/+source/iptstate/+bug/1562333
[22:21] <sarnold> elico: because you haven't prepared and tested a debdiff to fix it yet :)
[22:22] <sarnold> elico: packages in universe are community-maintained
[22:22] <elico> sarnold: I took the package from zetsy and installed on xenial and it works..
[22:23] <elico> sarnold: is there a guide about doing a debdiff??
[22:23] <sarnold> elico: nothing easy.. here's a starting point for the process to add a patch to a package https://wiki.ubuntu.com/StableReleaseUpdates
[22:25] <elico> sarnold: but if the zetsy package works fine, why not use it??
[22:25] <nacc> elico: becuase that's not how things work.
[22:25] <nacc> elico: and mixing and matching release is a bad idea (tm)
[22:26] <elico> nacc: OK but nobody does anything..
[22:26] <sarnold> elico: depending upon how much changed that might even be a solution, but it takes a -lot- of convincing to make that happen. applying a patch to a package is usually the far faster route
[22:26] <sarnold> elico: aye. tragedy of the commons lived out in software :/
[22:26] <nacc> elico: well, that's rather rude of you to think. people do *plenty* of stuff
[22:26] <nacc> elico: maybe not in this particular bug
[22:26] <elico> nacc: nobody responded in the bug(talking about this one..)
[22:27] <sarnold> well, that's not exactly a good analogy.. loads of people do loads of stuff, and it's not a fixed pool of resources under competition. but still, it's soooo easy to Let Someone Else Handle That..
[22:27] <elico> I created a package for OpenSUSE but I am not so good with debian packages..
[22:28] <sarnold> heh rpms sure feel easy in comparison to debs
[22:28] <elico> for me... maybe someone that worked a lot with deb's feels better with it..
[22:33] <elico> sarnold: where do I see the packages which are in community compared to others and if others then in which it belongs?
[22:34] <sarnold> elico: if you run apt-cache show iptstate one of the lines output is Filename: pool/universe/i/iptstate/iptstate_2.2.5-1build1_amd64.deb
[22:34] <sarnold> elico: the 'universe' there means that package is community supported
[22:34] <elico> sarnold: OK thanks!
[22:34] <elico> sarnold: are you from the server team?
[22:34] <sarnold> elico: no, security team
[22:36] <elico> If I will show you how I am building my DEB's you will find it "un-orthodox"
[22:37] <sarnold> probably, no two people build their debs the same way :)
[22:39] <elico> sarnold: You haven't seen mine... it's one of the most amazing ways to create a deb.
[22:39] <elico> http://gogs.ngtech.co.il/elicro/squid-debian-deb-build
[22:40] <elico> ..The Makefile..
[22:43] <sarnold> elico: hahah :) I can confirm I've never seen a Makefile to build debs quite like this one before :)
[22:47] <elico> sarnold: I had a talk with Amos the squid-cache project manager and he created a deb file for debian and ubunutu.
[22:47] <elico> He has a really nice one. Now I wanted to provide a package but after reading the DEB guide and book more then three times I got tired and tried this recipe and it appears that my server got hit hard because of it so I had to start QOS the traffic.
[23:39] <nacc> ahasenack: was your backtrace with g-u review of cpaelzer's MP: ValueError: Not sure how to decompose 0ubuntu13.1.22: series 1 not in expected format
[23:40] <nacc> ?
[23:40] <nacc> ahasenack: if so, can you file a bug for it? it means our version deduction alogrithm needs to be made more flexible
[23:53] <keithzg> Are there any prefab LAN search engines out there? I basically just want a local network version of Google/DuckDuckGo to help users with our sprawling set of different web services at the office (wiki, bugtracker, client database, doxygen, etc etc). I would have thought someone would have written something, but my google-fu appears to be failing...
[23:54] <sarnold> keithzg: tada https://enterprise.google.com/search/products/gsa.html
[23:56] <keithzg> sarnold: I guess to be picky I should stipulate that I was hoping for something FOSS-licensed, but uhh also apparently it's being discontinued?
[23:56] <sarnold> ah
[23:56] <keithzg> http://fortune.com/2016/02/04/google-ends-search-appliance/
[23:57] <keithzg> So I couldn't even get one purchased for the office if I wanted to!
[23:58] <keithzg> So far the "best" option seems to be hacking together something myself with Apache Nutch + Solr or ElasticSearch