[00:14] <Aison> what's wrong here? nfs-idmapd.service: Control process exited, code=exited status=1
[10:48] <curmet> I've developed and tested web app (using a framework) in windows machine successfully, but after I move the app and framework folder , from c:\xampp\htdocs\ to /var/www/html to Ubuntu machine, it encounters errors, "cannot find the requested view" . What changes should I make to Ubuntu machine?
[10:48] <curmet> THanks in advance :)
[11:35] <Repox> Hello. I just installed Ubuntu 16.04 and I want to setup Jenkins. Now, I have some options regarding installing OpenJDK, and I can see that most information refers to openjdk-7-jre, but is there any reason why I should want to avoid openjdk-8-jre or openjdk-9-jre?
[13:23] <Dulcin> Hi, is there a way to see how much data my server is using?
[13:24] <ReScO> ifconfig ?
[13:25] <patdk-wk> what is *data*?
[13:25] <Dulcin> Well the problem is, my host tells me my server (which has been there for years) all of a sudden started using 1TB data per day since about 3 weeks ago
[13:25] <Dulcin> and they charge 50 cents per gig
[13:26] <Dulcin> Is there a way to verify this?
[13:26] <patdk-wk> you are running mrtg/munin/cacti/... on it?
[13:26] <patdk-wk> my bet is your infected and sending spam email
[13:26] <patdk-wk> and attacking other websites
[13:27] <Dulcin> I haven't installed those myself. It's a default ubuntu server and I only set up a web server on it
[13:27] <patdk-wk> well, that is safe
[13:27] <Dulcin> so what should I start with to find out?
[13:27] <patdk-wk> atleast till you install php/cgi'/...
[13:28] <patdk-wk> well, see what is running on the server
[13:28] <patdk-wk> log all open connections
[13:29] <poohbear82> my website is running on ip that starts with 10. I understand it is on the local network. How do I make it public?
[13:30] <Dulcin> ok thank you I'll start with that
[13:30] <patdk-wk> poohbear82, ask your network admin
[13:31] <poohbear82> well conceptually... what has to happen?
[13:32] <poohbear82> I'm trying to understand
[13:32] <hateball> poohbear82: NAT has to happen
[13:35] <patdk-wk> maybe
[13:35] <patdk-wk> we don't know what has to happen, as we don't manage his network
[13:35] <patdk-wk> nat, or public ip assigned, or vlan move, or ....
[13:35] <patdk-wk> so many options, nothing we can really help with
[13:36] <patdk-wk> or just forget about this ipv4 stuff, and use ipv6
[13:36] <poohbear82> Ok, thank you.
[14:00] <Dulcin> patdk-wk, I'm a bit at a loss here, can you point me in the right direction?
[14:00] <Dulcin> I checked the auth.log and I see a bunch of ssh attemps on root user, but that's locked
[14:00] <Dulcin> when I check 'who' I'm the only one logged on
[14:01] <Dulcin> I checked netstat -tupn
[14:01] <Dulcin> Most seems to be webserver connections
[14:07] <patdk-wk> I never said to check auth.log
[14:07] <patdk-wk> that is normally useless
[14:07] <Dulcin> Sorry I'm a web developer with limited knowledge on these things
[14:08] <patdk-wk> if your compromised, and it shows in auth.log, you have issues
[14:08] <patdk-wk> it's almost always the web that is broken
[14:13] <Dulcin> patdk-wk: but these attemps seem to fail
[14:13] <Dulcin> What do I do to check if I'm compromised?
[14:13] <patdk-wk> yes, and failed attempts are not a compromise
[14:13] <patdk-wk> they are just normal random internet noise
[14:14] <patdk-wk> everything :)
[14:14] <patdk-wk> you don't know how you are compromised, or if you are, so the only way to know is to check everything
[14:14] <patdk-wk> but there are only two reasons for increased internet traffic
[14:14] <patdk-wk> ligit, increased usage of your server
[14:14] <patdk-wk> or non-ligit, compromised server
[14:15] <patdk-wk> I'm betting it's compromised, and likely via the webserver
[14:15] <patdk-wk> so looking in auth.log is likely pointless
[14:15] <patdk-wk> unless you have really weak passwords
[14:16] <Dulcin> I dont think I do, they all have lower-upper case and special characters in them
[14:16] <patdk-wk> it's possible someone compromised your password, but normally that is very rare
[14:17] <patdk-wk> compared to a website getting compromised
[14:17] <Dulcin> so, then I should check the web server logs?
[14:17] <patdk-wk> sure, but that won't likely tell you much
[14:17] <Dulcin> how do I proceed?
[14:17] <patdk-wk> normally the most help those are, is after you find the comprise, to locate when it was first used, and put in place
[14:18] <patdk-wk> like I said above
[14:18] <patdk-wk> look for anything running that sholdn't be
[14:18] <patdk-wk> look for any new files on your server
[14:18] <patdk-wk> anything that doesn't belong
[14:18] <patdk-wk> or if you just don't care, scrape the server and reinstall it
[14:19] <patdk-wk> hopefully without whatever was that let them in
[14:20] <Dulcin> patdk-wk: just got a call from my host, they finished their investigation...
[14:20] <Dulcin> patdk-wk: they made an error in their calculation...
[14:20] <Dulcin> I did not use 13TB in 12 days :|
[14:21] <patdk-wk> :)
[14:22] <patdk-wk> I would recommend you install mrtg/munin/cacti/... to monitor your usage
[14:22] <Gazby> Hi everyone. Is it OK to install the ubuntu-server package (http://packages.ubuntu.com/xenial/ubuntu-server) post-install? The basic images you often get are missing several packages from a server install so I like to fix that, but open-iscsi fails to start when installing ubuntu-server so the install fails.
[14:22] <patdk-wk> if nothing else, you will be able to see and predict when your usage changes and account for it
[14:23] <Dulcin> patdk-wk: any preference to those?
[14:23] <patdk-wk> personally I perfer munin
[14:23] <patdk-wk> though I use cacti also
[14:23] <patdk-wk> just cacti is kind of overbearing
[14:24] <patdk-wk> lots of poeple seem to like graphite
[14:24] <patdk-wk> all have different levels of setup complexity
[14:24] <Gazby> +1 for munin
[14:25] <Gazby> graphite is amazing, but a lot more work than munin, and probably for a more narrow use case
[14:28] <Pici> I'm using graphite + grafana here for some random stuff, I like it a lot.
[14:29] <Gazby> oh that's very sexy
[14:29] <RoyK> I've started setting up zabbix for testing.
[14:29] <Gazby> we used dashing http://dashingdemo.herokuapp.com/sample
[14:30] <RoyK> Gazby: commercial thing?
[14:30] <Gazby> no
[14:30] <Gazby> built by shopify, floss
[14:32]  * patdk-wk should floss
[14:32] <Pici> grafana+graphite: http://i.imgur.com/uIyQWSS.png
[14:33]  * patdk-wk is now just a stat on pici's graph
[14:33] <Gazby> that's gorgeous
[14:41] <genii> Pici: Are those three spikes netsplits?
[14:43] <Pici> genii: yes. I put an annotation thing in, but it doesn't work that well yet: http://i.imgur.com/A6OS1Bk.png
[14:45] <Pici> I'll open it up to ops once I figure out how to separate out my own personal dashboards out.  I have some of my own servers in there, plus I'm tracking unread emails and I'm working on another small project to feed some more esoteric data in.
[14:47]  * genii makes sure the coffee keeps coming
[14:53] <Gazby> have posted my question here: http://askubuntu.com/questions/766920/how-to-install-packages-from-basic-with-a-minimal-install
[14:54]  * patdk-wk never uses anything > ubuntu-minimal
[14:55] <patdk-wk> but why would open-iscsi fail?
[14:55] <patdk-wk> are they using some odd kernel? openvz thing?
[15:00] <Gazby> it happens on linode, standard kvm, and vultr
[15:01] <Gazby> my understanding is that it always fails and the ubuntu installer just ignores it
[15:01] <Gazby> you'll find it failing in the dmesg of any basic install
[15:03] <TJ-> Gazby: looks like bug 1576341
[15:05] <patdk-wk> that wouldn't affect kvm
[15:05] <Gazby> ya
[15:05] <Gazby> i had seen that, the log output differs from mine also
[15:09] <Pici> Gazby: hrm.  I remember needing to a bit of steps on my linode when I setup my linode, but that was a while ago and I don't recall seeing that error.
[15:11] <Gazby> i could probably cobble something together is `tasksel --list-packages` was outputting a package list. i've no idea what the output of "^server" means
[15:11] <mike-zal> I don't have apache on my server and yet during a boot system tries to open it. I get: The apache2 instance did not start within 20 seconds... etc. it slows down my boot
[15:12] <mike-zal> how can I get rid of this automatic apache start?
[15:12] <Pici> Gazby: you may want to ask #linode on oftc.net as well.
[15:13] <Pici> which reminds me, I should rejoin that too.
[15:13] <mike-zal> it showed afer I tried and installed webmin and then uninstalled it (with a script, just like instructed on their site)
[15:13]  * patdk-wk is testing it
[15:13] <patdk-wk> mike-zal, well, then you screwed it up
[15:13] <mike-zal> I know :P
[15:13] <patdk-wk> and you will have to ask their website what the script did
[15:13] <patdk-wk> and undo it
[15:13] <patdk-wk> not sure we have any idea what happened
[15:13] <mike-zal> shit, sorry
[15:14] <patdk-wk> Gazby, it worked fine for me
[15:14] <mike-zal> ok, so new install. it's testing anyway so far
[15:14] <patdk-wk> I installed a ubuntu-minimal
[15:14] <patdk-wk> then did an apt-get install ubuntu-standard
[15:14] <patdk-wk> it installed and worked just fine
[15:14] <patdk-wk> no issues
[15:14] <mike-zal> next time I won't install webmin. it's overkill for me
[15:15] <mike-zal> ajenti works just fine and should be sufficient
[15:15] <patdk-wk> so whatever they did to make their *custom* ubuntu 16.04 image, broke it
[15:15] <patdk-wk> assuming it really is kvm, full vritualization and not openvz
[15:15] <patdk-wk> but then, everything is assumed to be broken on openvz/lxc/...
[15:16] <mike-zal> is it safe to install server 16.04?
[15:16] <patdk-wk> what is safe?
[15:16] <Gazby> patdk-wk: i'm trying to instal ubuntu-server not ubuntu-standard
[15:16] <mike-zal> desktop version has cleary issues and I wonder if that's not transferred into server, although it shouldn't since there is no gui
[15:16] <patdk-wk> hmm
[15:16] <mike-zal> just asking what is the general opinion on this
[15:17] <mike-zal> they screwed desktop and that lowever trust overall in 16-04
[15:17] <patdk-wk> depends on what you call *safe*
[15:17] <mike-zal> hence my wondering wheter I should insyall it or 14.04
[15:17]  * patdk-wk wonders what is screwed up in desktop
[15:17] <mike-zal> mostly the new software center
[15:18] <patdk-wk> people actually use that?
[15:18] <mike-zal> it doesn't find packages that are available through apt-get
[15:18] <nacc> that issue is supposed to be fixed soon, though, iiuc?
[15:18] <patdk-wk> my only issue is the gnome toolbar
[15:18] <patdk-wk> that the gnome method to disable them doesn't work
[15:18] <mike-zal> also this other program, what was it? for manging packages. syn something?
[15:18] <mike-zal> synapitic
[15:18] <nacc> synaptic
[15:19] <mike-zal> also doesn't see those packages. but they are possible to install via terminal
[15:19] <mike-zal> that's a major screw up and people are talkig about it
[15:19] <mike-zal> anyway, that's gui problem so probably not related to servers
[15:20] <nacc> also "screwed" is a bit of an overstatement (IMO), and that's also why LTS -> LTS upgrades aren't offered yet :)
[15:20] <nacc> (again IMO)
[15:20]  * patdk-wk did a raw install, no upgrades
[15:23] <patdk-wk> oh, ubuntu-server is a new package, didn't used to exist
[15:23] <patdk-wk> no wonder I was confused
[15:23] <patdk-wk> looks like everything I don't use, no wonder I'm not too interested in it
[15:24] <mike-zal> I red release log, there are mainly changes in packages version
[15:24] <patdk-wk> only items in it, ethtool, patch, vlan
[15:25] <Gazby> it exists in 14.04, that's why i'm trying to install it on 16.04, been doing it on trusty forever
[15:25] <mike-zal> I'll give 16.04 try. there is no point of using old system and packages for the next few years
[15:25] <patdk-wk> I looked in 14.04, it tells me no package
[15:25] <patdk-wk> apt-cache show ubuntu-server
[15:25] <patdk-wk> N: Unable to locate package ubuntu-server
[15:25] <patdk-wk> E: No packages found
[15:25] <mike-zal> and release upgrae is a bit risky
[15:25] <nacc> rmadison says only xenial + yakkety, fwiw
[15:27] <patdk-wk> http://packages.ubuntu.com/search?suite=all&searchon=names&keywords=ubuntu-server
[15:27] <mike-zal> nacc: did you install 16.04?
[15:27] <nacc> mike-zal: a few months ago, yeah
[15:27] <patdk-wk> I have around 20 16.04 installs so far
[15:27] <mike-zal> and no problems?
[15:27] <patdk-wk> only the one on my laptop that uses the gnome scrollbars
[15:28] <patdk-wk> it's damned impossible to use those scrollbars on a 4k screen that is 15"
[15:28] <mike-zal> ok, got the iso. will launch startup
[15:28] <nacc> mike-zal: none, but this is on a desktop (and i'm running ubuntu gnome, technically, i guess)
[15:28] <mike-zal> I'm installing in VB for educational purposes
[15:28] <mike-zal> learning some basics
[15:29] <mike-zal> I'm planning to buy cloud VPS and move my shop there
[15:30] <mike-zal> ovh will have 16-04 probably ready in summer, at least so they told me over the phone a month ago.. :P hopefully quicker
[15:30] <patdk-wk> yuk
[15:30] <patdk-wk> ok, ubuntu-server installed manually on ubuntu-minimal
[15:30] <patdk-wk> install worked, reboot worked
[15:44] <rbasak> patdk-wk: ubuntu-server is a metapackage generated from the server seed, new in Xenial. It's there so that if a user removes packages depended on by ubuntu-server, they know that they may be breaking something by no longer having what we call "Ubuntu Server", and on future release upgrades we can add to the list so users' systems can follow along.
[15:45] <daftykins> hey folks, i have an OpenVPN setup running on a 14.04.4 server (so vivid HWE) that is proving to be wholly unreliable, might the 4.2 kernel have issues with openVPN?
[15:45] <rbasak> ubuntu-desktop etc. already did something similar.
[15:46] <daftykins> reliability woes manifest as webpages sporadically loading through the tunnel... network resources failing one moment that don't the next... etc.
[15:46] <mike-zal> out of curiosity, if you decided for ubuntu server then why? why ubuntu and not centos or debian?
[15:46] <daftykins> mike-zal: just checking, was that query to someone else?
[15:46] <patdk-wk> rbasak, you oviously where not following along
[15:47] <patdk-wk> rbasak, read Gazby
[15:47] <mike-zal> that was for anyone who is willing to share
[15:47] <patdk-wk> !poll
[15:47] <patdk-wk> !best
[15:47] <patdk-wk> damn bot
[15:47] <daftykins> mike-zal: run what you like, it's not apt to ask for opinions in distro specific channels
[15:48] <mike-zal> ok, then let me rephrase. why do you like ubuntu server? just curious. I have my reasons to choose it but I'd like to know why others use it
[15:49] <mike-zal> I'm not looking for other server if that is what you are implying daftykins
[15:49] <daftykins> mike-zal: sorry i don't follow
[15:49] <patdk-wk> yes, but those reasons are specific to that person, and rarely relates to others
[15:50] <mike-zal> mine are simple. I don't know much about servers so I
[15:50] <mike-zal> so I'd better stick to ubuntu since I know it a little
[15:50] <mike-zal> while debian or red hat systems are totally strange for me
[15:50] <daftykins> debian'd be the same experience pretty much
[15:51] <mike-zal> plus, there is lot of stuff on the net about ubuntu server so in case of questions, it's easy to look for answers
[15:51] <mike-zal> also, ubuntu seems to be more newbie friendly
[15:52] <patdk-wk> ubuntu is *debian based*
[15:52] <patdk-wk> so for debian to be strange, would be really strange :)
[15:52] <mike-zal> I know but there must be sime differences. never tried pure debian
[15:52] <Pici> Ubuntu's releases are easier to understand.
[15:53] <daftykins> in the past i found ubuntu had sane configuration defaults for packages, whereas you may install the same package on a debian machine and it just wouldn't even run
[15:54] <mike-zal> so it seems that ubuntu is more user friendly after all
[15:56] <WOW> Hello
[15:56] <WOW> I meet a problem and can't figure it out for a couple of days...
[15:56] <Pici> Whats up?
[15:56] <WOW> need help...
[15:56] <WOW> We have multiple trusted domains.
[15:57] <daftykins> WOW: type it all on one line if you can, volunteers don't want to read a page of scrollback
[15:57] <WOW> I built one ubuntu server to join domain A..Works perfect..
[15:57] <WOW> I can use command id, su -...
[15:57] <WOW> ok
[15:58]  * patdk-wk wonders what a trusted domains is
[15:59] <daftykins> windows server term probably
[15:59] <daftykins> i.e. active directory domains
[16:00] <WOW> I built another ubunbu server to join domain B, same steps, only different domain name..after I join the ubuntu server to domain B. I can see this server listed in domain B. I can login to the ubuntu server by build in administrator account only..Can't use any other domian accounts to login to the server or run id or su - command..If I run id command, it gives me " no such users"
[16:01] <patdk-wk> ya, active directory defines it, domain doesn't could be dns, could be network, could be hundreds of things
[16:03] <WOW> I have compared the settings of these two boxes...the only different is domain name...DNS works will
[16:03] <WOW> well
[16:03] <WOW> any special logs I can check...
[16:05] <daftykins> yes authentication related ones
[16:11] <WOW> I am using sssd+ realmd
[16:11] <WOW> by default all the domain use should be able accesss the box once it is joined to domain...
[16:12] <WOW> no issue on box one which was joined to domain A
[16:12] <WOW> Doesn't work on the second box which was joined to domain B...
[16:12] <WOW> runnning out of idea at all
[16:13] <WOW> I can see the second box in the AD...
[16:14] <daftykins> yeah so read the auth logs when trying as a domain user
[16:16] <tgm4883> Has anyone used altermime to add a multiline disclaimer to all postfix emails? I've got it adding it, but it's making my disclaimer a single line
[16:19] <WOW> Apr 29 15:16:18 ubuntutest login[2551]: pam_unix(login:auth): check pass; user unknown Apr 29 15:16:18 ubuntutest login[2551]: pam_unix(login:auth): authentication failure; logname=administrator@MYDOMAIN.COM uid=0 euid=0 tty=/dev/tty1 ruser= rhost= Apr 29 15:16:18 ubuntutest login[2551]: pam_sss(login:auth): authentication failure; logname=administrator@MYDOMAIN.COM uid=0 euid=0 tty=/dev/tty1 ruser= rhost= user=MYDOMAIN.COM Apr 2
[16:19] <WOW> here is the logs
[16:19] <WOW> I don't quite understand...
[16:20] <ReScO> LDAP is kicking my ass :(
[16:21] <ReScO> can't think of a proper structure
[16:27] <WOW> You know what, I am going to try samba + winbind
[16:27] <WOW> and see if different result...
[16:29] <mike-zal> why sudo restart ssh says that there is no such command?
[16:30] <genii> mike-zal: Because you want something more like sudo service ssh restart
[16:31] <mike-zal> genii: thanks :)
[16:31] <ReScO> Anyone in here a bit more experienced with setting up LDAP in a corporate environment?
[16:35] <mike-zal> shit, can't connect to ssh from host computer. it's not doing anything :(
[16:35] <mike-zal> does ssh has to be opened in ufw first?
[16:37] <tarpman> ReScO: #ldap is a good channel for discussions of LDAP in general
[16:37] <genii> mike-zal: You're syaing that: ssh localhost  ..or: ssh 127.0.0.1     or: ssh <exact IP of the host computer>  while on the host computer does not work?
[16:37] <mike-zal> yeap
[16:38] <mike-zal> I do: ssh root@localip and nothing. cursor moves to the next line
[16:39] <mike-zal> I got actually only netowork IP 192.168.1.16 and I use that
[16:39] <genii> Try a non-root account
[16:39] <genii> work, AFK
[16:40] <mike-zal> the same
[16:40] <mike-zal> maybe I have something wrong on server
[16:43] <genii> mike-zal: Do you have openssh-server package installed?
[16:43] <mike-zal> ok, found it. ufw was blocking it. added rule for ssh
[16:43] <mike-zal> genii: thanks
[16:45] <genii> np
[16:45]  * genii wanders back to work 
[17:00] <mike-zal> how to install mysql database?
[17:00] <teward> mike-zal: sudo apt-get install mysql-server
[17:00] <mike-zal> sudo mysql_install_db seems to be deprecated
[17:00] <mike-zal> I already got it
[17:00] <teward> mike-zal: then I fail to understand your question
[17:00] <teward> the 'automatic configuration' is done during the apt-get install phase
[17:00] <mike-zal> it says "Please consider switching to mysqld --initialize"
[17:00] <teward> and sets up the 'mysql' server
[17:01] <mike-zal> ok, then maybe that is why I get "[ERROR] --initialize specified but the data directory has files in it. Aborting."
[17:02] <mike-zal> I'm looking on 14.04 instalation guide so some things are outdated obviously
[17:02] <mike-zal> in 14.04 it had to be done manually
[17:07] <mike-zal> is php5-fpm ok for 16-04?
[17:10] <mike-zal> I guess not, because I got that it doesn't have accessible version
[17:16] <teward> mike-zal: there's no php5-fpm in the repositories
[17:17] <teward> mike-zal: I wrote a blog post about how to get *a* version of php5.6 for the system, but it's not an official Ubuntu-supported method because PPA
[17:17] <teward> !ppa
[17:17] <teward> mike-zal: http://dark-net.net/?p=128
[17:22] <teward> mike-zal: you can try php7.0-fpm first though
[17:22] <teward> mike-zal: though you may have some headaches getting it to the same point your php5-fpm was at
[17:24] <mike-zal> yeah, I found and installed php7
[17:25] <mike-zal> was just looking on outdated info
[17:25] <teward> mike-zal: MOST guides are not updated for 16.04 :P
[17:25] <mike-zal> on digital ocean is
[17:26] <mike-zal> at first I thought they didn't so I followed old tutorial but then I found newer version
[17:26] <mike-zal> so far they have the best guides
[17:26] <daftykins> bear in mind DO is a bit different from stock ubuntu, such as being email'd root passwords instead of a user which can sudo - although you can easily implement this setup then disable root SSH login (which is sane security practice)
[17:27] <teward> ^
[17:27] <teward> daftykins: though, most VPS providers fall into that category
[17:27] <teward> (RamNode let me use an ISO to reinstall my VPS though, so I was able to do the full from-scratch setup on the KVM VPS heh)
[17:28] <mike-zal> ok, thanks. will read about securing server later. in guide there are some basic notes how to disable the most insecurities but I'm sure that's not all. besides this is testing server so I don't want to set all security measures just yet
[17:29] <mike-zal> I'm complete newbie so I have to go with some very simple to understand guides
[17:31] <daftykins> no you should do the user management bits from the beginning
[18:01] <axisys> how do install a older version of package? I want to install offlineimap from trusty into xenial
[18:02] <axisys> sudo apt-get -t=trusty install offlineimap did not work
[18:02] <axisys> The value 'trusty' is invalid for APT::Default-Release as such a release is not available in the sources
[18:03] <daftykins> you can't pick a version based on distro name
[18:04] <daftykins> it's also *highly* inadvisable to try to take one from an older release
[18:04] <hallyn> smb: sarnold: arges: *completely* unpredictably, i'm running into trouble trying to use both groups libvirt and libvirtd :)
[18:05] <axisys> daftykins: had been trying to current version working for few days.. so using docker to run older version and running offlineimap from it with shared folder for Maildir/ .. but I like to run it from outside docker
[18:06] <daftykins> axisys: so the point is you're running... 16.04? and this ones is broken?
[18:07] <axisys> daftykins: actually correctly speaking.. this one fixed the certificate issue and I cannot make the imap ssl work anymore.. older version silently ignored it and was working fine.. :-)
[18:08] <daftykins> heh
[18:08] <axisys> daftykins: may be I should take this discussion back to #offlineimap again :-)
[18:08] <daftykins> or run 14.04
[18:08] <axisys> last time it was quite...
[18:08] <axisys> already upgraded to 16.04
[18:09] <axisys> so downgrading to 14.04 would be lot of work.. so vagrant or docker are my only solutions to use older distro
[18:09] <daftykins> well you can't downgrade, you'd simply spin up a VM for these times
[18:09] <daftykins> back to their channel it is then by the sounds :)
[18:09] <axisys> I do not want to start a whole new VM with vagrant.. so picked docker..
[18:09] <axisys> daftykins: :-)
[18:10] <axisys> daftykins: docker works really well.. but i need to work a little bit more for a persistent cache.. or it is kind of slow..
[18:11] <hallyn> bah, bc it's trying to be too smart:
[18:11] <hallyn> sudo addgroup --system --gid 117 libvirt
[18:11] <hallyn> addgroup: The GID `117' is already in use.
[18:12] <hallyn> -o i guess
[18:22] <daftykins> docker != VM as far as i follow it
[18:27] <sarnold> hallyn: *shocked* I'm shocked I say!
[18:27] <hallyn> sarnold: i'm gonna give it one more go and if that fails i'll just punt on the switching to libvirt gorup for now :)
[18:29] <ppetraki> hallyn, apw, so we just find the coolest g++-5.0 compiler bug  - https://bugs.launchpad.net/ubuntu/+source/gcc-5/+bug/1577891
[18:30] <hallyn> ppetraki: "c++ lolz"  :)
[18:30]  * hallyn hasn't used a c++ class since 2001
[18:31] <ppetraki> hallyn, it's how real work get's done ;)
[18:31] <hallyn> you know this actually rings a bell
[18:31] <hallyn> wasn't there some libc bug or kernel bug causing a memset to always write 0s instea dof the requested bytes, or something?
[18:31] <ppetraki> hallyn, me neither, until I started working here. No reason to use anything else.
[18:31] <sarnold> ppetraki: the heck does this mean? :) "Signature *sp = new(p) Signature();"
[18:31] <ppetraki> sarnold, "write your stuff here"
[18:32] <ppetraki> aka placement new
[18:32] <sarnold> in. sane.
[18:32] <ppetraki> you get used to it, rarely used to be honest, but when you need it you need it
[18:33] <ppetraki> what's insane is optimizing out functions that I call
[18:33] <sarnold> I can imagine that if you were going to have afew million objects of a given type that you might want to use this thing with slabs that grow and shrink etc..
[18:33] <ppetraki> we do
[18:33] <ppetraki> represents a SHA1
[18:33] <hallyn> ppetraki: well you wanted optimization, give us a prize for giving you what you want
[18:33] <hallyn> what's faster than "don't do it"
[18:34] <ppetraki> well, since it's a sha, and sha names the data, we'll never find it again. so that's kinda bad for a storage array
[18:34] <hallyn> just blame it on btrfs
[18:35] <ppetraki> hallyn, you can blame everything on that. It's not even fun anymore
[18:35] <daftykins> can't spell butterfingers without btrfs!
[18:35]  * daftykins ducks
[18:35] <hallyn> heh ppetraki wants a challenge
[18:35] <sarnold> ppetraki: hopefully useful: http://www.daemonology.net/blog/2014-09-04-how-to-zero-a-buffer.html http://www.daemonology.net/blog/2014-09-06-zeroing-buffers-is-insufficient.html
[18:35]  * hallyn looks on amazon for  BOFH calendar
[18:36] <hallyn> sarnold: maybe but not until the compiler is fixed i'm guessing :)
[18:36] <ppetraki> sarnold, interesting...
[18:36]  * ppetraki tries
[18:37] <sarnold> hallyn: that may still be the end result, yes. but memset_s() or secure_memzero() may be ways to tell the compiler that they aren't noops and can't be optimized away
[18:37] <hallyn> oh, i see.
[18:39] <axisys> daftykins: right docker is a container.. VM is OS over Host+Software as hypervisor
[18:40] <de-facto> does xenial server support using i/o schedulers as kvm guest?
[18:40] <ppetraki> sarnold, is it in 16.04? can't build using this example, http://en.cppreference.com/w/c/string/byte/memset
[18:40] <bekks> de-facto: that question doesnt make much sense.
[18:40] <patdk-wk> de-facto, what does that mean?
[18:41] <bekks> de-facto: an I/O scheduler is not some kind of vm.
[18:41] <de-facto> bekks im asking because some distros deactivate schedulers when used as kvm guests (e.g. you cannot set schedulers on /dev/vda)
[18:42] <patdk-wk> you can do that, but it's rather pointless
[18:42] <patdk-wk> cause any scheduling you do, will be *redone*, so your just wasting cpu/memory
[18:43] <sarnold> de-facto: I understand only cfq scheduler supports the io priorities anyway, and afaict everyone hates that scheduler
[18:43] <sarnold> ppetraki: it may require a different -std= ..
[18:44] <patdk-wk> cloud-at-cost has a 10second delay programmed into their scheduler :)
[18:44] <sarnold> patdk-wk: holy
[18:44] <de-facto> yeah but what if i want to do it anyhow? is it possible on xenial or blocked?
[18:44] <sarnold> patdk-wk: that's one way to discourage disk IO :)
[18:44] <sarnold> de-facto: I've never seen an error message from ionice :) I assume you'll continue executing just without the classes you expected
[18:46] <de-facto> e.g. will         echo "cfq" > /sys/block/vda/queue/scheduler ; cat /sys/block/vda/queue/scheduler         yield something else than "none" on xenial as kvm guest (virtio /dev/vda)?
[18:49]  * patdk-wk still wonders what your attempting to gain
[18:51] <sarnold> de-facto: hah, looks like that fails.
[18:52] <ppetraki> sarnold,  nm -D /lib/x86_64-linux-gnu/libc.so.6 | grep memset_s
[18:52] <ppetraki> sarnold, no results, memset is there, I don't think it made it https://sourceware.org/ml/glibc-bugs/2015-01/msg00193.html
[18:52] <sarnold> ppetraki: :(
[18:53] <ppetraki> sarnold, but I can look for alternatives, now that I know that this exists. Thank you, that was helpful.
[18:53] <patdk-wk> does the vda driver even support schedulers?
[18:53] <patdk-wk> still don't get why you would want to
[18:54] <patdk-wk> any ordering you do, will be redone with the hosts scheduler
[18:56] <de-facto> sarnold so its not possible to change scheduler in xenial on such a scenario either then?
[18:57] <patdk-wk> "For better performance of I/O-intensive applications, a new I/O path was introduced for the virtio-blk interface in kernel version 3.7. This bio-based block device driver skips the I/O scheduler, and thus shortens the I/O path in guest and has lower latency. It is especially useful for high-speed storage devices, such as SSD disks. "
[18:57] <sarnold> de-facto: at least my xenial guests on xenial host doesn't. i'm not sure why you'd want to, the guest hasn't got a clue what storage looks like..
[18:57] <patdk-wk> it is not possible on ANY linux kernel to do it
[18:59] <de-facto> sarnold i just want to try different schedulers to compare performance. on an old wheezy i got "cfq" and its fast, on jessie i got "none" and its slow. so i want to change it for comparison
[18:59] <de-facto> hence my question if xenial would allow me to do that
[18:59] <sarnold> sure
[18:59] <sarnold> just not in VMs
[18:59] <de-facto> :-(
[19:08] <patdk-wk> de-facto, you have a lot more differences than just the scheduler
[19:08] <patdk-wk> cause the only reason the scheduler doesn't work, is if your using bio mode of that driver
[19:08] <patdk-wk> so on wheezy, it doesn't support bio mode
[19:10] <de-facto> how can i disable bio mode then?
[19:13] <de-facto> patdk-wk also where can i see if bio mode of what driver is used?
[19:16] <de-facto> i guess you are referring to /lib/modules/3.16.0-4-amd64/kernel/drivers/block/virtio_blk.ko or similar then
[20:54] <de-facto> so did everyone pull the openssl update in already?
[20:55] <karstensrage> yup
[20:55] <karstensrage> restarted apache, stunnel?
[20:56] <karstensrage> whats else?
[20:58] <de-facto> karstensrage maybe something like "apt-cache rdepends openssl" can reveal?
[20:58] <de-facto> is openssl the correct dependency in that case? or lib...?
[20:59] <trippeh> libssl1.0.0
[20:59] <rattking> I did a "lsof | grep DEL | grep ssl" to find them
[20:59] <sdeziel> karstensrage: I use this script to find processes needing a restart: https://github.com/simondeziel/puppet-unix-tools/blob/master/files/sbin/check-deleted-libs
[21:00] <trippeh> I kind of want to match processes to cgroups which could be matched to systemd units
[21:00] <trippeh> want to make something to, that is
[21:03] <de-facto> sdeziel nice one thanks :)
[21:04] <sdeziel> de-facto: I've also wrote a little nagios check for it: https://github.com/simondeziel/custom-nagios-plugins/blob/master/plugins/check_deleted_libs
[21:05] <de-facto> very neat indeed :)
[21:06] <sdeziel> if you run some some VMs, you might be interested by check_qemu_outdated (now I'm done promoting my stuff ;)
[21:10] <karstensrage> thanks rattking and sdeziel and de-facto
[21:10] <karstensrage> all restarted
[21:10] <de-facto> yeah  def have to look more into automagic stuff :)
[21:11] <teward> Does apache2 on systemd systems (such as 15.10 and 16.04) still install an init.d file to /etc/init.d/ ?
[21:11] <teward> asking because I'm automating a 'restart web services' script to push to my servers and systems via Landscape
[21:12] <teward> it also reloads ssh too (openssl!)
[21:12] <sdeziel> teward: yes, the systemd task is just a wrapper around the init script
[21:12] <teward> s/automating/setting up/
[21:12] <teward> sdeziel: this would apply for other services such as 'ssh' too?
[21:13]  * teward is trying to use the existence of the init script for apache or nginx (the two web servers he would ever use) as criterion for a services restart
[21:13] <sdeziel> teward: for ssh it's a a native task
[21:13] <teward> sdeziel: SSH has a restart command either way - it's the web servers i'm trying to do this with
[21:13] <teward> sdeziel: http://paste.ubuntu.com/16208886/ is the 'current' script logic
[21:13] <teward> but not executed
[21:14] <teward> yet
[21:14] <sdeziel> teward: you might want to use the generic "service" wrapper command
[21:14] <sdeziel> should be working for all versions/daemons
[21:15] <sdeziel> AFK now
[21:15] <teward> sdeziel: OK, will do, thanks
[21:15] <sdeziel> np
[21:24] <teward> apparently when I set up landscape on clients, it doesn't allow scripts as root - is there a way to fix that easily on multiple systems?
[21:24] <teward> without me manually redoing the landscape config
[21:25] <teward> nevermind
[21:28] <teward> i have a temporary fix :P
[21:44] <blizzow> upgrading my ubuntu server to xenial is hanging up during the upgrade of the mdadm package.  Specifically, it looks like it hangs while trying to stop the currently running mdadm service.  Does anyone have an idea how I might get past this hurdle? I tried manually stopping the mdadm service and that's not working.  The OS is on md0, so that has me a little concerned.
[21:51] <autofsckk> hi, i have an isue with the r8169 problem, i think this thing is ancient but it seems to be giving us a lot of problems in some servers, i have strange behaviour on the network, for what ive been investigating i think is because of that module issue
[21:52] <autofsckk> we use very old productions servers 12.04, could it be possible that this module bug could be hurting this servers?  they run 4 of this  05:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 06)
[21:54] <autofsckk> i get a lot of this on dmesg   http://pastie.org/private/ezasoe2ell2kejbaqgfig
[21:58] <genii> !info firmware-realtek
[21:58] <genii> Hm
[22:01] <autofsckk> genii:  i was making some iperf test with a dev server to see if i get some info from it but it seems to be working just fine, but when i look at some logs it seems that some errors come from disconnecting
[22:02] <genii> autofsckk: You might want to try the r8168-dkms package, and see if that helps
[22:02] <genii> !info r8168-dkms
[22:02] <autofsckk> i mean, when i see some logs from some apps running in the server it shows that those error come from bad network or disconnections, any idea on how could i be sure that the actual module is working right?
[22:04] <genii> There's no real foolproof way to be able to tell if it's due to actual network disconnections, or the driver, unfortunately
[22:05] <autofsckk> i couldnt find that r8168-dkms, should i add something to the repos to get it? im sorry i dont know too much ubuntu
[22:10] <autofsckk> genii:  you mean i have to make this?  https://sikpigs.wordpress.com/2013/10/07/realtek-r8168-on-ubuntu-12-04/   i already downloaded the modules from realtek, compiled them and installed it on a dev server here on the office, but i dont know how to test this stuff locally without the real environment :S well thanks for the tip anyway genii :D
[22:18] <genii> autofsckk: Since the package is in the "universe" repository, you need to enable that repository in your /etc/apt/sources.list file for apt to be able to see it's there after a sudo apt-get update
[22:34] <autofsckk> genii:  ah ok, thanks a lot, ill change that then :D