[00:00] <Patrickdk> !webmin
[00:00] <Patrickdk> !virtualmin
[00:00] <Patrickdk> !ebox
[00:03] <WACOMalt> virtualmin/webmin supports 14.04 though?
[00:04] <WACOMalt> Patrickdk, does zentray offer plenty of options? Aka can I host say, 5 separate domains on it, each using a different IP etc etc
[00:05] <WACOMalt> mail server, FTP, mysql, postgre
[00:05] <Patrickdk> how should I know?
[00:05] <WACOMalt> oh... you recommended it so I figured maybe you used it
[00:06] <Patrickdk> I didn't recommend anything
[00:06] <Patrickdk> I just listed *options*
[00:06] <WACOMalt> ok, sorry then
[00:06] <Patrickdk> there might be more options
[00:06] <Patrickdk> but those are the ones that I know of
[00:09] <Sachiru> Also can anyone recommend a shared web host or cloud email provider? The shared web hosting will be used strictly for IMAP/email purposes (we need custom branding for email for 15 users, but we create and delete users a lot due to the nature of the nonprofit, so if we'll go by created users a month it will be more like 30 users)
[00:18] <WACOMalt> !ehcp
[00:37] <WACOMalt> I need to manually add a public key for a repo I think.. but I have no idea how to do that. This is the repo: http://archive.zentyal.org/zentyal/dists/3.5/
[00:38] <WACOMalt> there is a Release.gpg file there. I assume this is the key I need to add
[00:58] <jak2000> how to restart a ububtu server? sudo shutdown -g0 -r     ? or?
[00:58] <Patrickdk> reboot? :)
[00:58] <jak2000> yes reboot
[01:05] <sarnold> WACOMalt: you need something like: sudo apt-key adv --recv-key E23F4777
[01:05] <sarnold> WACOMalt: (I found the keyid by downloading the Release and Release.gpg file, and running "gpg Release.gpg"
[01:05] <WACOMalt> I wound up figureing it out from this https://wiki.zentyal.org/index.php/Daily_Builds#Installing_from_APT_repository
[01:05] <WACOMalt> ah
[02:02] <jak2000> how to restart/reboot a ububtu server? sudo shutdown -g0 -r     ? or?
[02:03] <sarnold> jak2000: what does -g0 do? I don't see it in my shutdown(8) manpage
[02:03] <jak2000> sarnold :) bad practice other linux
[02:03] <jak2000> the: sudo shutdown -r   its ok?
[02:04] <sarnold> jak2000: I always use sudo shutdown -r now
[02:07] <rena_> sudo reboot
[02:08] <sarnold> I don't recommend reboot since some unix systems instantly reset the CPU on 'reboot': no dirty buffer flushing, no filesystem umounts, no process killing..
[02:11] <TJ-> sarnold: reboot calls shutdown unless runlevel is {0,6}
[02:25] <jak2000> sarnold wich VPS reccomend me? cheaper, good speed, good processors, and much hard drive? :)
[03:20] <fraq> if I want to download all the dependencies for a package (which may already have been downloaded)
[03:21] <fraq> what is the recommended way to go about that?
[05:44] <mechamjoseph> hey all
[05:45] <mechamjoseph> my server is running quite slow. when i run “top” i see quite a large number of “sendmail-mta” processes. is this normal?
[06:05] <RoyK> mechamjoseph: check the mail logs - also - why on earth are you running sendmail?
[06:05] <mechamjoseph> thank you Royk
[06:05] <mechamjoseph> i didn’t set this server up, i’m just helping out with a company, and i’m not really familiar with running a server
[06:06] <mechamjoseph> what should i do instead of sendmail?
[06:06] <RoyK> postfix or exim - I prefer postfix - but anyway - if it works, don't fix it
[06:06] <mechamjoseph> haha it doesn't
[06:06] <mechamjoseph> it’s totally busted
[06:06] <RoyK> but then - I can't recall ever seeing a process called sendmail-mta
[06:07] <RoyK> it should be just sendmail
[06:07] <RoyK> pastebin "ps axfwww" please
[06:08] <RoyK> !pastebin | mechamjoseph
[06:09] <histo> RoyK: is that like super wide output? www?
[06:09] <RoyK> histo: mhm
[06:09] <RoyK> histo: ps tends to cut lines
[06:28] <histo> RoyK: one w would be sufficient, is what I was trying to say. www does nothing vs. w
[06:28] <sarnold> histo: have you tried?
[06:29] <histo> sarnold: tried what?
[06:29] <sarnold> histo: ps auxw vs ps auxwww ?
[06:29] <histo> sarnold: yes
[06:30] <RoyK> histo: no, one w alone doesn't do what two does. three was perhaps a bit overkill
[06:30] <RoyK>        w               Wide output. Use this option twice for unlimited width.
[06:30] <sarnold> histo: note one w:
[06:30] <sarnold> sarnold  24919  0.0  0.0  11564   736 ?        Ss   Jul18   0:00 cscope -dl -f /home/sarnold/ubuntu/security/audits/librevenge/utopi
[06:31] <sarnold> histo: two ww:
[06:31] <sarnold> sarnold  24919  0.0  0.0  11564   736 ?        Ss   Jul18   0:00 cscope -dl -f /home/sarnold/ubuntu/security/audits/librevenge/utopic/librevenge-0.0.1/cscope.out
[06:31] <histo> ahh
[06:34] <sarnold> hunh. where'd that terminal go anyway?
[07:00] <lordievader> Good morning.
[07:07] <RoyK> morning
[07:11] <lordievader> Hey RoyK, how are you?
[07:11] <RoyK> lordievader: fine, thanks
[07:37] <histo> Now wait a sec, sarnold when I tested diff <(ps auxw) <(ps auxwww) no difference here. I assume that depends heavily on output so I'll take your word for it.
[07:41] <histo> s/output/running pids/
[07:41] <user123321> http://pastebin.com/EghdcWfA <--- this in the interfaces file. I tried pinging 8.8.8.8 but the responses are way too slow. I'm getting like only 3 responses within 15-20 seconds. The problem came after adding CARP setting. Any ideas?
[07:48] <lordievader> user123321: Default route timeout?
[07:51] <user123321> lordievader, Each response came with time around 70 ms. Would installing traceroute be helpful?
[07:52] <lordievader> user123321: Check your routes: ip route/route
[07:53] <user123321> lordievader, ip route gives, default via 192.168.1.1 dev eth0
[07:53] <user123321> 192.168.1.0/24 dev eth0  proto kernel  scope link  src 192.168.1.3
[07:55] <lordievader> user123321: Judging from your interfaces file I'd say that is correct, however I have no experience with a carp device and I do not know the influence of such a thing.
[07:59] <user123321> Yep, problem comes only after CARP, lol.
[08:00] <user123321> If I remove the CARP setting, then it will work. hmm.
[08:02] <user123321> What's a good HA and LB method for Ubuntu?
[08:03] <user123321> At least for, 2 servers. Main server + Backup server
[08:03] <user123321> They would have, for example, an Apache.
[08:20] <user123321> Is HAProxy good?
[08:27] <user123321> Hello, I would appreciate some advice regarding the possibility to install this in VM's --> http://support.severalnines.com/entries/23612682-Install-HAProxy-and-Keepalived-Virtual-IP-
[08:28] <user123321> For example, without getting actual physical machines, is it alright to install some of the parts in VM's, any recommendations?
[10:03] <mephx> can a xfs filesystem be mounted during ubuntu setup in an alternative console? mount doesn't support -t xfs in the current 14.04 setup.
[10:06] <user123321> Any comments for this network design? Load Balancer 1 and Web server 1, both in one physical machine (LB1 in a VM), Load Balancer 2 and Web Server 2 in another physical machine (LB2 in a VM). The web servers and load balancers are identical. Both the load balancers will have one virtual IP.
[10:06] <user123321> So basically, there's a reduction from 4 machines to 2 machines. Any ideas?
[11:25] <weeb1e_> Apparently, someone managed to remove the kernel entirely from an ubuntu server
[11:26] <weeb1e_> And grub is not showing any previous kernel to load
[11:26] <weeb1e_> What are the steps to recover the system and reinstall a kernel? I need to provide them via email as I have no direct access to the machine
[12:01] <peetaur2> weeb1e_: :D
[12:01] <peetaur2> weeb1e_: boot rescue/live cd, chroot into root, make sure boot is mounted, apt-get install linux-image-...
[12:02] <peetaur2> weeb1e_: and then:    optional, not required unless you also lost /boot:  grub-install /dev/sdX          then maybe optional:  update-initramfs -u          and important:  update-grub
[12:03] <peetaur2> I have also written this which has the same chroot and grub install steps:  http://pastebin.com/4NLFkdzQ
[12:05] <peetaur2> and I wrote this answer for even if you rm -rf /boot: http://askubuntu.com/questions/117027/cannot-recover-grub-due-to-missing-vmlinuz-and-intrd-img/134278#comment181504_134278
[12:05] <weeb1e_> peetaur2: Yeah I found instructions online and passed them on
[12:05] <weeb1e_> I saw the box respond to pings while live booted, but it has not come back up since rebooting
[12:06] <weeb1e_> So I assume it did not go well
[13:44] <sdfer> For some reason my laptop just *wont* boot grub after having installed ubuntu. It just keep booting Windows
[13:44] <sdfer> I've tried with and without secure boot enabled
[13:53] <pds_corp> hi guys, is there a tutorial regarding on how to renew a certificate for an apache server
[13:56] <peetaur2> pds_corp: as long as you know where the files are, should be easy...
[13:56] <peetaur2> pds_corp: the trouble is just finding the files. eg. someone in ##proxmox changed their cert and found that they had to also open a .jar file to replace the one for the VNC client too.
[13:56] <pds_corp> peetaur2: nope guess staring with .csr file
[13:56] <peetaur2> to find the old certs, just find the config that points there
[13:57] <peetaur2> and you also change the key, not just cert
[13:57] <peetaur2> client needs cert, but server needs key
[14:06] <jamespage> o/
[14:07] <gnuoy> o/
[14:08] <gnuoy> jamespage, I'm looking at a package which has a change log entry that says "Rebuild to drop files installed into /usr/share/pyshared". But deb diff shows nothing relating to /usr/share/pyshared.
[14:10] <rharper> \o
[14:11] <rharper> jamespage: last time I sent you some diffs for etc_keeper,  you had some comments on them, but I wasn't what needed done next.
[14:11] <gnuoy> jamespage, would that be a change in the code tarball then ?
[14:12] <lutostag> o/
[14:12] <lutostag> anyone want a autofs debdiff?
[14:16] <rbasak2> o/
[14:16] <rbasak2> Sorry I'm late
[14:19] <coreycb_> o/
[14:19] <rbasak2> So I have access to not very much, but I can do web, Launchpad and a shell on an Ubuntu system.
[14:20] <rbasak2> So I can look at stuff as needed I hope, but I have no overlayfs, schroot or sbuild (awkward armhf system running on a vendor non-Ubuntu kernel)
[14:20] <rbasak2> I will help as much as I can!
[14:20] <rbasak2> Does anyone have the pad link from last time?
[14:20] <rbasak2> (no IRC logs here either :(
[14:21] <jamespage> gnuoy, is it a python package?
[14:21] <rharper> rbasak2: what was the domain ?
[14:21] <rbasak2> pad.ubuntu.com
[14:21] <rbasak2> Like ubuntu-server or something
[14:21] <lutostag> http://pad.ubuntu.com/server-team-merges
[14:21] <rbasak2> That's it. Thanks!
[14:21] <gnuoy> jamespage, no. I have it for pacemaker and moin
[14:22] <rharper> emelectrictx.com
[14:22] <rharper> huh
[14:22] <rharper> lutostag: has it
[14:22] <jamespage> gnuoy, pacemaker is python
[14:22] <jamespage> gnuoy, basically prior to droppnig py 2.6 python got installed to /usr/share/pyshared and then symlinked to 2.6 and 2.7 install locations
[14:23] <jamespage> when 2.6 got dropped things just got installed direct to 2.7
[14:23] <lutostag> do sponsors here prefer debdiffs or lp branches?
[14:24] <coreycb_> I wasn't here last fri, do we just pick a pkg at random?
[14:24] <rbasak2> I prefer debdiffs, but am happy to work with a LP branch
[14:24] <rbasak2> Though right now I don't have my gpg key to sign uploads :-/
[14:24] <lutostag> thats ok, doesnt need to go thru right now; but now I'll create a bug
[14:25] <jamespage> lutostag, I prefer a branch personally
[14:25] <rbasak2> coreycb_: yes, please pick a package at random. We're managing a list of what people are working on in http://pad.ubuntu.com/server-team-merges to avoid duplicate work.
[14:25] <jamespage> rharper, can't remember - that was last week!
[14:25] <rharper> hehe
[14:25] <rharper> lemme find the pastes
[14:26] <coreycb_> rbasak2, will diddly do.  if you know of an easy one for a first timer, let me know.
[14:26] <rharper> Jul 25 11:22:38 <rharper>	jamespage: debian_1.12_to_1.12ubuntu1.debdiff -> http://paste.ubuntu.com/7857095/  ; ubuntu_1.11ubuntu1_to_ubuntu_1.12ubuntu1.debdiff -> http://paste.ubuntu.com/7857097/
[14:28] <rbasak2> coreycb_: I only just did exim4, and it's been updated in Debian already! I know it's fairly trivial. Probably mostly educational though - we don't need it merged again as a priority, but it might be a good one to start with.
[14:30] <coreycb_> rbasak2, great thanks, I'll take a look at that
[14:48] <lutostag> ah so you guys want both debian(new)->ubuntu(new) and ubuntu(old)->ubuntu(new) for debdiffs?
[14:50] <rbasak2> lutostag: yes, that's what sponsors generally tend to want.
[14:51] <rbasak2> rbasak2: I never understood that though. I want to see debian(old) -> ubuntu(old) and compare that to debian(new) -> ubuntu(new) personally. But I can generate what I need.
[14:51] <lutostag> rbasak2: ok, I'll get the first half of it into bug 1351337 then; thanks
[14:52] <rbasak2> It is pretty awkward when you need sponsoring. I remember it well. Different sponsors tend to want different things. I think sponsors understand this difficulty though, and are generally quite flexible, and I try to be the same.
[14:53] <lutostag> yeah I figure for next merge I'll do a branch with 2 commits first is debian updates, second is ubuntu fixes
[14:55] <jamespage> gnuoy, OK _ pacemaker uploaded - I tweaked the changelog a bit
[14:55] <jamespage> take a look
[14:56] <gnuoy> thanks
[14:59] <arrrghhh> Nivex, still nothing official that I can really find on "why" has 14.04.1 not been made available...
[15:00] <jamespage> rharper, looking now
[15:00] <rharper> jamespage: sure
[15:01] <gnuoy> jamespage, where can I see it ?
[15:01] <jamespage> gnuoy, you should have an email
[15:02] <rbasak2> arrrghhh: it is available. It's just not automatically being recommended if anything.
[15:02] <arrrghhh> rbasak2, semantics, but it's not here http://changelogs.ubuntu.com/meta-release-lts
[15:02] <arrrghhh> I realize I can get a image from the release site... that's not what I'm after :)
[15:03] <Nivex> https://bugs.launchpad.net/ubuntu/+source/update-notifier/+bug/1344762
[15:03] <Nivex> rbasak2: ^^
[15:04] <jamespage> rharper, http://paste.ubuntu.com/7924702/
[15:04] <jamespage> I'd expect a copy-paste of the previous merge comments
[15:04] <jamespage> and a validation that they are needed
[15:04] <jamespage> rharper, also note that if you are the sole change author, no need for [ Ryan Harper ]
[15:05] <gnuoy> jamespage, I see no email (well I see lots of email but I'm guessing you don't mean the one about the Muppet movie being available on Netflix)
[15:06] <gnuoy> just got it
[15:06] <rharper> jamespage: not following the first one -- the only patch applied is the conversion of bzr to git and the tutorial -- I don't really have a justification for that;  that's just what was already done...
[15:06] <jamespage> rharper, that debdiff I pasted is the full diff, which includes all of the changes in the changelog
[15:07] <jamespage> rharper, merging is normally just about doing what we did last time and resolving any conflicts or drops
[15:11] <rharper> jamespage: ok -- so I posted the wrong thing ?  I mean for etc keeper, the merge change just looked to me like we kept the same "use bzr by default" patch between versions
[15:11] <rharper> jamespage: not sure how to express that in debdiffs
[15:11] <jamespage> rharper, the debdiff was just fine - but your changelog entry needed updating - that's all I changed
[15:12] <rharper> oh -- the cut-n-paste of the previous ubuntu patch changelog
[15:13] <jamespage> rharper, indeed
[15:13] <jamespage> rharper, that's where I normally start
[15:13] <jamespage> copy-and-paste and then check each comment
[15:13] <rharper> gotcha
[15:13] <rharper> jamespage: so, let me update it and send you new debdiffs?
[15:14] <rharper> that way I know I got it right this time?
[15:17] <jamespage> rharper, sure
[15:17] <jamespage> rharper, redo the process and check against my debdiff
[15:18] <gnuoy> jamespage, I've thrown Bug #1351331 up as well
[15:19] <rharper> jamespage: thanks!
[15:31] <coreycb_> rbasak2, so for exim4 the only conflict is that ubuntu has additional patches in debian/patches/series
[15:32] <coreycb_> rbasak2, which is kind of odd actually b/c I don't see the patches in debian/patches
[15:33] <gnuoy> jamespage, rbasak2, I think I'll just unassign Bug #1351331 from myself and subscribe ~ubuntu-sponsor . I don't think it's controversial
[15:49] <rharper> jamespage: ok, if my debdiff is the same as yours on the etckeeper (with the email updated as well) -- what's next?  the merge bug and attach the two debdiffs ?
[15:50] <jamespage> rharper, nah - I'll just upload it
[15:53] <rharper> jamespage: ok
[15:55] <qman> I'm having an issue with zfs on ubuntu server, where after a reboot, my pool has devices missing or corrupted. I'm pretty sure that zfs is just loading too quickly and my drives aren't ready, because if I export and import again, it's fine. How can I introduce some boot delay here to try and fix it?
[15:56] <hallyn> gnuoy: I didn't test-build or anything, but moin debdiff lgtm
[16:13] <jkyle> anyone have a working example of a working kickstart file for trusty/14.04 with lvm partitioning?
[16:14] <jkyle> the one I have keeps failing to create the volume groups
[16:16] <jkyle> or a working preseed expert_string, though the kickstart is way cleaner if I can get it to work
[17:01] <hallyn> jkyle: I don't, but pls post if you get one :)  we should really have a wiki page for various preseed segments
[17:01] <hallyn> per-iso, probably
[17:01] <hallyn> (well, per-iso-style - i.e. netbookt, server iso, desktop iso)
[17:15] <med_> is zul still in transit?
[17:15] <zul> no hes trying to recover from being in transit
[17:33] <jkyle> hallyn: I'm working on one. currently, I've given up on doing anything remotely not vanilla. just trying to get a standard atomic partition going.
[17:35] <jkyle> it'd be nice if they just ditched current installer and ported over anaconda
[17:39] <sarnold> histo: ah. perhaps ps is determining what to do whether or not the output is a terminal or not! that'd explain both of our experiences very quickly :)
[17:46] <LarsN> is there a script/utility that generates preseed configurations based on an existing system?
[17:47] <hallyn> jkyle: in that case, I'
[17:47] <hallyn> I'm using http://people.canonical.com/~serge/uquick
[17:47] <hallyn> gives me standard ubuntu user/password and simple disk layout
[17:54] <jkyle> LarsN: you can get it going with debconf-get-selections
[17:55] <jkyle> LarsN: but they rarely work without tweaking and usually include a whole lot of extra, unneeded (or unwanted) options
[17:56] <jkyle> hallyn: let me try that one, I just want a running system at this point.
[17:57] <hallyn> jkyle: it's based on kirkland's that he used to ship at uquick.ly or somesuch.  should we create a wiki page wiki.ubuntu.com/preseeds?
[18:00] <jkyle> hallyn: I think they have some in ubuntu community documentaiton. just need to update with a working. this canned preseed will probably work, I was just hoping ot get it going with a kickstart
[18:03] <jkyle> hallyn: that preseed stops at language selection
[18:03] <jkyle> for me
[18:04] <hallyn> jkyle: the very last time i tried it just ddnt work at all, but not too long ago it was working 100% for me with a server iso
[18:04] <hallyn> (my machien is hammered right now, can't try atm)
[18:05] <LarsN> jkyle, hallyn Thanks to both of you
[18:06] <jkyle> there's a lot to like about debian, but it's installer isn't one. what's frustrating is this has been a trivial process for ages in most distros
[18:07] <hallyn> jkyle: in principle it should be nice to use;  i don't know why it's so hit-or-miss, and yes it's horribly frustrating :(
[18:20] <inv3nt0r> I'm attempting to upgrade my Ubuntu 12 server to Ubuntu 14 using the do-release-upgrade method, and getting 404 errors for the 'Upgrade tool signature' and the 'Upgrade tool'
[18:20] <inv3nt0r> Any thoughts?
[18:24] <mov_agn_ntrl> i just got here myself a min ago - looks like none here
[18:26] <mov_agn_ntrl> i just found this inv3nt0r  http://askubuntu.com/questions/65911/how-can-i-fix-a-404-error-when-updating-packages
[18:26] <mov_agn_ntrl> and this inv3nt0r   http://askubuntu.com/questions/462326/upgrade-ubuntu-server-11-04-signature-error
[18:27] <inv3nt0r> Thanks
[18:27] <patdk-wk> why post something about 11.04? that has been unsupported for years
[18:28] <mov_agn_ntrl> yeah inv3nt0r it's what i thought - you could write a list of your manual upgrades you have done - and then i think you could also make a boot disk with al you r current settings just in case you need to go back to 12 for some reason, but looks lik einstalling a new version of 14 might be your best bet
[18:29] <mov_agn_ntrl> does this help inv3nt0r https://answers.launchpad.net/ubuntu/+source/update-manager/+question/251853
[18:29] <inv3nt0r> Crappity.
[18:29] <mov_agn_ntrl> I found this one - it looks prmiosing, that's all i got inv3nt0r  best of luck  http://www.cyberciti.biz/faq/howto-upgrade-to-ubuntu-14-04-from-ubuntu-13-10-or-12-04/
[18:30] <mov_agn_ntrl> *pro...
[18:32] <inv3nt0r> Hmm
[18:32] <mov_agn_ntrl> anything ring a bell or help you see whats wrong inv3nt0r ?
[18:32] <inv3nt0r> Well
[18:32] <inv3nt0r> I'm jhaving mixed results
[18:33] <mov_agn_ntrl> well best of luck man :)
[18:33] <mov_agn_ntrl> or woman - idk what kind of inv3nt0r you are :P
[18:34]  * inv3nt0r = man
[18:34] <mov_agn_ntrl> same here dude
[18:34] <inv3nt0r> Haha
[18:34] <inv3nt0r> Well
[18:34] <inv3nt0r> Maybe I don't need to upgrade
[18:34] <inv3nt0r> This server won't be my problem after 5 more work days
[18:35] <inv3nt0r> I hate to leave something broken behind though...
[18:35] <inv3nt0r> Right now I cant even apt-get install zip
[18:35] <inv3nt0r> So
[18:35] <inv3nt0r> I guess I can do a deb install
[18:35] <mov_agn_ntrl> you could always do the complete download and install on a partitioned area of disk - so if soehting goes wrong, you can go back? but isnt 12 almost out of service?
[18:37] <inv3nt0r> It appears that way, thus why I cant even install zip support :)
[18:41] <mov_agn_ntrl> what is zip support?
[18:42] <mov_agn_ntrl> the only reason i ever use zip files anymore is to compress a folder and send it by email, but most people that receive it still dont know what to do
[18:43] <patdk-wk> oh? you can send zip via email without it being rejected? :)
[18:47] <user123321> What kind of settings might I need to check for HAProxy and Keepalived for reducing the delay in case of an IP fail over?
[18:47] <andol> user123321: If nothing else the healthchecker interval matters.
[18:48]  * inv3nt0r gives himself a pat on the back
[18:48] <user123321> andol, cool.
[18:48] <inv3nt0r> Looks like zip_3.0-6 is not in the repository, but luckily 3.0-4 was so I installed it via dpkg
[18:50] <andol> user123321: Well, except that I might have answered the wrong question :) The healthchecker is really about detecting broken backends, no for failing over the actual HA frontend.
[18:51] <user123321> Aha
[18:52] <patdk-wk> heh?
[18:52] <patdk-wk> keepalived can failover the frontend easily
[18:52] <andol> patdk-wk: Yes, but were refering to the keepalived healthchecker specifially, which might not have been relevant gives user123321's question.
[18:53] <patdk-wk> well, if he is using haproxy, keepalived shouldn't be checking backends
[18:53] <patdk-wk> but only the haproxy frontend
[18:53] <patdk-wk> and failingover based on that
[18:54] <andol> Another point in favor of my answering the wrong quetsion :)
[18:54] <user123321> Right now, I have Keepalived1+Haproxy+web server in one VM, and Keepalived2+Haproxy2+web server in another VM. If I fail one VM, my browser waits some seconds.
[18:55] <patdk-wk> it takes time
[18:55] <patdk-wk> you have many issues there
[18:55] <patdk-wk> you have how long it takes for haproxy to detect a bad backend
[18:55] <patdk-wk> you have the mount of time it takes keepalive to detect a bad haproxy
[18:55] <patdk-wk> and you have the time it takes for move the ip to the good one
[18:56] <patdk-wk> and it doesn't sound like you have the time for it takes to enact stonith/fencing
[18:56] <user123321> Any ideas to reduce?
[18:56] <patdk-wk> that would require me to learn keepalived, not something I have ever liked
[18:56] <patdk-wk> I'll stick with corosync
[18:56] <user123321> Aha
[18:57] <patdk-wk> but still, no matter what, it will be a few seconds :)
[18:57] <patdk-wk> at best, likely 2seconds
[18:57] <inv3nt0r> mov_agn_ntrl: I needed to create a .zip file for a Windows user who would never in a million years figured out how to open a .tar.gz
[18:58] <patdk-wk> use 7zip :)
[18:58] <mov_agn_ntrl> oh. .tar.gz? i never heard of it but i would probably get a non-malware program to open in in a few mionutes form a download site like download.com etc
[18:58] <mov_agn_ntrl> idk though, i just dont zip files anymore
[18:58] <user123321> patdk-wk, is it really impossible to make it like, 0.1 seconds? I'm thinking like, "just ping and if no response, redirect to the other or something" :p
[18:59] <inv3nt0r> Believe me, I'd have them use something more advanced but they would eff up, and install scumware from CNET
[18:59] <patdk-wk> heh? how is that going work?
[18:59] <patdk-wk> how can you redirect, if it's dead
[19:00] <patdk-wk> you have to bring the ip up online
[19:00] <user123321> redirect to the live server.
[19:00] <patdk-wk> and you have to update the mac caches on everything upstream
[19:00] <patdk-wk> you can't redirect
[19:00] <patdk-wk> this is ip, not http
[19:00] <patdk-wk> there is no redirect
[19:00] <patdk-wk> how would a dead server redirect?
[19:00] <user123321> umm I mean, something like redirect :d
[19:00] <patdk-wk> nothing like a redirect!
[19:01] <user123321> dead server doesn't do anything, right? So, if the live server could see that the other server is dead, can't it become active quickly?
[19:02] <patdk-wk> the issue is, detecting if it is dead in .1 second isn't normally trivial
[19:02] <patdk-wk> and if you ACTIVATE it
[19:02] <patdk-wk> and the other server is not dead, you have issues
[19:03] <user123321> if the live server pings the dead server, if the response isn't received within 0.05 seconds?
[19:04] <user123321> I'm actually thinking about whatever the settings that could do it quickly ;)
[19:04] <patdk-wk> you really think that *dead server* is never going be overloaded
[19:04] <patdk-wk> due to like *php* load
[19:04] <patdk-wk> or a disk going bad
[19:05] <user123321> I see, so, during the transition period, let's assume, 2 seconds. If, during this time, if 1000 new connections are requested by clients, what would happen?
[19:06] <user123321> will they get no connection error in any case
[19:06] <user123321> or just getting response in 2 seconds?
[19:07] <user123321> let's make it maximum, 1 million etc.
[19:07] <patdk-wk> it's tcp
[19:07] <patdk-wk> and it will depend on many things
[19:07] <user123321> what if both http and tcp
[19:07] <patdk-wk> likely, it will just wait 2seconds before it connects
[19:08] <user123321> I see
[19:08] <patdk-wk> if it was keepalive, they will get a connection reset, and maybe it tries again
[19:08] <patdk-wk> what od you mean both http and tcp
[19:08] <patdk-wk> there is only TCP/ip
[19:08] <patdk-wk> http is not a transit layer protocol
[19:08] <user123321> Oh I meant, apache etc.
[19:09] <patdk-wk> user123321, before attempting ANY of this
[19:09] <patdk-wk> learn about the iso model, and about ip/tcp
[19:09] <patdk-wk> what does apache have to do with it? that is even LESS than http does
[19:09] <patdk-wk> and http has NOTHING to do with it
[19:09] <user123321> umm, there's layer 7 LB.
[19:09] <patdk-wk> haproxy is a layter 7 lb
[19:09] <patdk-wk> but we aren't talking about that
[19:10] <patdk-wk> we are talkg about BEFORE haproxy
[19:10] <patdk-wk> last I knew, keepalive was not a layer 7 lb
[19:10] <patdk-wk> or even ANY lb at all
[19:11] <user123321> that i'm not sure, I thought keepalived just made the 2 LB's highly available.
[19:12] <patdk-wk> yes
[19:12] <patdk-wk> and if we are attempting to make the lb high available, how can we do that using a layer 7 lb?
[19:12] <user123321> regarding apache, during the fail over, my browser says "connecting" and eventually connected.
[19:12] <patdk-wk> you can't protect something using itself
[19:12] <patdk-wk> why does apache matter?
[19:13] <patdk-wk> you -> internet -> isp -> keepalived -> haproxy -> apache
[19:13] <patdk-wk> if it's broken at the keepalived layer (server died, ip broken, ...)
[19:13] <patdk-wk> how does apache matter at all?
[19:14] <user123321> Well I didn't get a connection reset.
[19:14] <patdk-wk> you sure?
[19:14] <patdk-wk> that would happen i nthe background
[19:14] <patdk-wk> and depend on EXACT TIMING
[19:14] <user123321> yeah, tried many times.
[19:14] <patdk-wk> if you are already connected to the first server
[19:14] <patdk-wk> and it dies
[19:14] <patdk-wk> you failover ip to second one
[19:15] <patdk-wk> second one will see packet but won't have session state for it, and respond with a reset/denied
[19:15] <user123321> Oh, the failover while downloading data. I suspect that would reset.
[19:16] <user123321> btw, any suggestions for haproxy settings that would reduce delay?
[19:16] <patdk-wk> I don't use haproxy
[19:16] <patdk-wk> I don't see a need to use to double up like that
[19:16] <user123321> Ah
[19:17] <user123321> what do you use?
[19:17] <patdk-wk> ipvs
[19:17] <user123321> I see
[19:18] <user123321> What's the failover delay in your case?
[19:18] <patdk-wk> around 15seconds
[19:18] <user123321> ok
[19:18] <patdk-wk> if the balancer where to die
[19:19] <patdk-wk> it takes a long time to bring up that many ip addresses
[19:19] <patdk-wk> if a backend dies? <1second
[19:19] <user123321> Cool
[19:19] <patdk-wk> I have only had the balancer die once in 4 years
[19:20] <user123321> how many balancers you've got?
[19:21] <patdk-wk> 2
[19:21] <patdk-wk> one active, one passive
[19:22] <user123321> Aha
[19:27] <user123321> patdk-wk, Is it sensible to install servers in the load balancing machine? Is it usually the practice to dedicate a separate computer only for a load balancer?
[19:29] <patdk-wk> that depends on scale
[19:30] <patdk-wk> using the same machine, makes it slower to failover
[19:30] <patdk-wk> but using dedicated onces takes more resources
[19:31] <user123321> Cool
[20:08] <user123321> patdk-wk, what's your suggestion for load balancing and making highly available between, let's say, only 2 servers? Should I get 2 separate LB's or..?
[20:09] <user123321> What would be my options?
[20:56] <flubby_nanabo> So, If I wanna run a script on startup, but it's supposed to run only once, and exit, should I put it on rc.local, or make it a service on /etc/init (or init.d)?
[21:17] <sarnold> flubby_nanabo: I'd do an upstart service, might as well make it integrate easily with the other tools. cron's @reboot might be handy if it is a user-script instead of system-script
[21:23] <Patrickdk> sarnold, please :)
[21:23] <Patrickdk> at this point, I wouldn't recommend upstart anymore, but systemd
[21:23] <Patrickdk> considering upstart is being removed
[21:24] <sarnold> Patrickdk: except he isn't running systemd today, and he isn't running systemd tomorrow.
[21:24]  * Patrickdk still needs to redo his
[21:24] <Patrickdk> well then :)
[21:24] <Patrickdk> if he has 14.04 he could be running systemd today
[21:24] <sarnold> Patrickdk: maybe if he upgrades to 14.10 and if pitti goes bonkers, -maybe-....
[21:24] <rberg_> this is how I do a first boot scrip in rc.local http://paste.ubuntu.com/7927450/
[21:26] <flubby_nanabo> I do have 14.04, but I am using the ubuntu-upstart docker image, so, I guess I'm sort of stuck with upstart for now
[21:27] <Patrickdk> :)
[21:27] <flubby_nanabo> Patrickdk, upstart is being deprecated? I thought it was the new shinny thing ...
[21:27] <Patrickdk> I just hate to write something, that I will have to rewrite
[21:27] <Patrickdk> itwas, back in 12.04?
[21:27] <Patrickdk> but debian went systemd, and ubuntu opted to not fight it
[21:28] <flubby_nanabo> Patrickdk, I need a docker image with a init system that the puppetlabs-mysql puppet module understands, and the ubuntu-upstart was the only one I could find, so, I don't see myself rewriting it any time soon
[21:29] <Patrickdk> :)
[21:29] <Patrickdk> odd
[21:29] <Patrickdk> actually, you should be able to in a few months
[21:29] <Patrickdk> with rhel7 out, using systemd
[21:29] <Patrickdk> and debian too
[21:29] <Patrickdk> that should be around soon
[21:29] <Patrickdk> no worries though
[21:30] <flubby_nanabo> Patrickdk, I'm not planning on ditching ubuntu, with works pretty well, and I don't have experience with rpm distros.
[21:30] <flubby_nanabo> *it owrks
[21:30] <flubby_nanabo> *it works
[21:30] <flubby_nanabo> damn fingers!
[21:30] <Patrickdk> I mean, the pupplet-mysql-for-systemd should be out soon
[21:30] <flubby_nanabo> Patrickdk, Oh, that.
[21:31] <flubby_nanabo> I should probably take a look at the code, see if there's any changes in some new release
[21:31] <flubby_nanabo> IIRC, last time I checked it just either used upstart, or just failed when starting the services
[21:39] <bilde2910> Hello! I have a laptop with Ubuntu 14.04 server, as the main server for my website. Earlier today, the server got disconnected, seemingly without reason. I tried restarting the machine, but the network does not go up. Running ifconfig only shows loopback interface lo, `sudo lshw -C network' lists the interface em1 (the ethernet interface) as "*-network DISABLED". Card is an RTL8101E/RTL8102E PCI Express Fast Ethernet
[21:39] <bilde2910> Controller". Any help resolving this? I need to get the machine online as fast as possible because thousands of users of my web services are unable to connect to the website
[21:41] <bilde2910> I furthermore tried to sudo dhclient em1, which brought the interface up with DHCP. It gives the wrong IP, therefore breaking port forwarding rules. sudo /etc/init.d/networking restart doesn't output anything, `sudo ifdown em0 && sudo ifup em1' gives errors: "ifdown: interface em1 not configured", "RTNETLINK answers: File exists" and "Failed to bring up em1".
[21:45] <markthomas> Does anyone know if it's possible to configure network bonding via preseed?
[22:51] <zartoosh> hi I am using ubuntu 14.04. I keep observing this in my /var/log/messages init: upstart-file-bridge main process ended, respawning
[23:05] <jaawerth> zartoosh: does it coincide with wifi connectivity problems by any chance?
[23:08] <jaawerth> zartoosh: if so, this may be relevant http://askubuntu.com/questions/259530/why-does-my-wifi-internet-intermittently-disappear