[09:30] <YamakasY> anyone doing some decent kernel management with puppet ?
[09:52] <LeMike> hello. how do I see the inbox of another user via terminal? I can cat his /var/mail/foo file but due to the chunks I can not read it. Is there a command like mailq to read the mails of another user?
[09:59] <caribou> nacc: rbasak: I think I will need help with the new usd-import git repo created for kexec-tools
[10:00] <rbasak> caribou: for a merge?
[10:00] <caribou> rbasak: yes, I'm testing the new procedure from nacc's usd-import tree
[10:01] <rbasak> nacc and I spent a while yesterday figuring out how to do a merge using the new structure. It is non-trivial.
[10:01] <caribou> rbasak: (I've done the merge using the previous method )
[10:01] <rbasak> Our conclusion is that we need a tool that will take the imported tree and produce a branch that has no merge commits that is suitable for rebasing.
[10:02] <caribou> rbasak: for some reason, I'm checking out ubuntu/yakkety, rebasing -i, reset HEAD^ and I get a version that is way past the last merged version
[10:02] <rbasak> It will probably start with a commit from Debian (git merge-base debian/sid ubuntu/yakkety) and then cherry-pick using -m back up to ubuntu/yakkety).
[10:03] <rbasak> Then that should be familiar against the old process.
[10:03] <rbasak> Otherwise rebase gets really confusing since most commits are merge commits. It sounds like that's what you're hitting.
[10:04] <caribou> rbasak: yes, looks like that
[10:05] <caribou> rbasak: will try that
[10:08] <rbasak> caribou: so I did:
[10:08] <rbasak> git checkout -b master abb0083a5ad0b9e25c0f08a8d08758ed0922d553
[10:08] <rbasak> git cherry-pick -m 2 4715d38
[10:08] <rbasak> git cherry-pick c92c1ce
[10:08] <rbasak> That seems to work. As a sanity check, this should produce no output:
[10:08] <rbasak> git diff origin/ubuntu/yakkety
[10:09] <caribou> rbasak: let me try that
[10:10] <rbasak> And now you can "git rebase debian/sid" and it should just work (no --onto needed)
[10:10] <rbasak> I get merge conflicts but that's as expected. To avoid the debian/changelog ones, it's necessary to do the full reconstruct and logical steps.
[10:14] <caribou> rbasak: ok, let me try that method on my kexec-tools merge that I will compare with the previous one I did & document the steps
[10:15] <caribou> rbasak: then you guys can review the document & see if I'm right, then we can append the doc somewhere
[12:25] <jonah> Hi guys I just wondered if anyone could please help. I have just installed 'smartmontools' on my backup server just to keep an eye on the disk errors etc. I've also set to run it as a daemon on startup. The thing is in the config  file it says it will email errors to my root user... Does anyone know how I set mailx to email me on a different email address or set the root user/sudo user (whichever it will email) to have a forwarding email so
[12:25] <jonah> I don't have to ssh into the backup server and type mailx to check for local mail. I instead just want to get it in my normal mail inbox...
[12:25] <jonah> Sorry for the long question
[12:25] <jonah> hope it makes sense...
[12:31] <brigante> jonah, which config file are you referring to? /etc/defaults/smartmontools?
[12:32] <jonah> brigante: yeah that's the one
[12:33] <rbasak> jonah: set up an MTA to forward all mail to some other system. I configure exim for this.
[12:35] <rbasak> jonah: here's my exim4.conf as an example: http://paste.ubuntu.com/16918668/
[12:35] <rbasak> All my non-email servers use this to forward all root mail to me.
[12:36] <brigante> surely you could just change the setting from root to your email address?
[12:36] <brigante> rather than forwarding
[12:36] <brigante> like in crontab
[12:36] <rbasak> That requires a working MTA still.
[12:37] <brigante> sure but, apt install postfix is not exactly difficult
[12:37] <rbasak> Or msmtp or something, but that will lose mail during network outages.
[12:38] <rbasak> You still need to configure it, and you end up with a fully functional email server with mailboxes for every user, etc.
[12:38] <rbasak> I don't want that, so I configure exim to only have a local root user and forward all root mail to me. This will catch mail generated by any daemon (eg. debconf).
[12:39] <rbasak> One could set up root alias in the MTA, of course. I'm effectively doing this but more minimally.
[12:39] <brigante> its a nice lightweight approach rbasak, but does it account for bouncebacks etc?
[12:40] <brigante> when i need a server to email out, i always install postfix. no need tro setup users or anytrhing else
[12:40] <rbasak> What bouncebacks?
[12:40] <brigante> rbasak, you say you're forwarding mail to another mailbox, what if your mailbox is not reachable?
[12:40] <brigante> will it queue the email, and re-try?
[12:41] <rbasak> Yes.
[12:41] <brigante> thats nice
[12:41] <brigante> isn't exim a fully featured MTA anyway?
[12:41] <rbasak> Yes, but its configuration is flexible. I'm configuring it minimally to do almost nothing.
[12:42] <brigante> fyi - you dont NEED to configure postfix either
[12:42] <brigante> out of the box it has a minimal setup
[12:42] <rbasak> You get asked questions.
[12:42] <rbasak> That's configuring it.
[12:42] <brigante> haha
[12:42] <brigante> i guess
[12:43] <brigante> 1) Internet site 2) enter bogus server name... config complete
[12:43] <brigante> but always nice to know of alternatives
[12:43] <rbasak> and that's what leads to stuff like https://bugs.launchpad.net/ubuntu/+source/postfix/+bug/1576978
[12:43] <rbasak> The bug tracker is full of these.
[12:44] <brigante> there will always b users tho
[12:44] <brigante> some bnetter than others
[12:44] <brigante> better*
[12:45] <brigante> i hear what you're saying though, less config is good
[12:45] <brigante> im not sure we helped jonah tho
[12:46] <brigante> rbasak, do you know much about iproute2 or networking in general?
[12:46] <brigante> i have a very strange scenario on one of my servers i'd like to run by someone to check im not going insane
[12:48] <brigante> jonah, type in man smartd.conf
[12:49] <brigante> the -m flag allows you to email any email address, i.e. -m admin@example.com
[12:59] <coreycb> beisner, qemu 1:2.2+dfsg-5expubuntu9.7~cloud4 is ready to be promoted to kilo-updates
[13:03] <caribou> rbasak: nacc: done with the kexec-tools merge using your imported repository. MP is in LP
[13:05] <jonah> brigante: hi sorry I just had to go for a bit there, i'll just read what I've missed!
[13:09] <jonah> brigante: ok thanks for all those comments and the pastebin... so in that conf file could I just change the -m root to -m emailaddress@addres.com etc?
[13:10] <brigante> jonah, that should do the trick mate
[13:10] <brigante> BUT you will still need an MTA (like postfix or exim)
[13:10] <brigante> otherwise you have no running SMTP server, which is required to send email
[13:10] <jonah> brigante: ok I've changed that line and changed root to just an email address, that's awesome thanks. But what about the -M exec bit, will this need removing or does it do anything by default on ubunt server?
[13:11] <brigante> jonah,  /dev/sdc -m admin@example.com -M test
[13:11] <brigante> sends a test to admin@example.com
[13:11] <brigante> im just reading the man pages here, i dont really know
[13:12] <jonah> brigante: thanks, really appreciate the help
[13:12] <brigante> no problem
[13:14] <jonah> brigante: just tried the test but it says "Unrecognized option: m"
[13:14] <jonah> brigante:  is this the right command sudo /dev/sda -m myemail@gmail.com -M test
[13:15] <brigante> no it wont work from the command line
[13:15] <jonah> brigante: or is that because I need to install postfix?
[13:15] <brigante> you have to put it in the smartd.conf
[13:15] <jonah> brigante: ah ok
[13:15] <brigante> then when you run `service smartmontools restart`
[13:15] <brigante> it will send the test email
[13:15] <jonah> brigante: so you put that in smartd.conf instead of the normal command and do the servrice restart, ok will try it thanks
[13:16] <brigante> but you should `apt install postfix` first
[13:16] <rbasak> caribou: process-wise that looks great! How was it for you? Would you like me to review the merge itself?
[13:17] <jonah> brigante: ok I've add postfix and it's asking if it is 'Internet Site' 'Sattelite system' etc do I just use Internet Site?
[13:17] <caribou> rbasak: a review would be nice indeed. not much delta so it should be quick
[13:17] <rbasak> OK I'll do it now.
[13:17] <caribou> rbasak: process-wise, I find it clearer, once I had your clarifications
[13:18] <caribou> rbasak: I just sent a request to import corosync so I can test it again
[13:18] <brigante> jonah, chose internet site
[13:18] <caribou> rbasak: & see if I can figure out how to get started myself
[13:18] <jonah> brigante: ok thanks and just leave the System mail name as the name of the machine?
[13:19] <brigante> yeah thats fine, unless you plan on using a domain name you actually own
[13:19] <brigante> if so, use that
[13:20] <brigante> jonah, when your server sends email out, it will be from [user]@[the domain you use]
[13:21] <caribou> rbasak: it's easier not to have to worry about the ./git ./gitwd directories
[13:22] <jonah> brigante: hmm tried that and now the service smartmontools restart won't work, it just says fail...
[13:23] <brigante> jonah, `tail -f /var/log/syslog` should shed some light as to why
[13:23] <brigante> probably a bad line in your config
[13:24] <brigante> leave tail running, then restart smartmontools in another terminal window
[13:29] <jonah> brigante: I ended up rebooting and now it will restart but weirdly the test email doesn't come through...
[13:30] <caribou> rbasak: btw, regarding the merge proposal, I wasn't sure about the "preference path" to be used for the MP. Used debian/sid as was shown in the screen capture
[13:30] <brigante> jonah, type `mailq`
[13:30] <brigante> any mail queued?
[13:30] <jonah> brigante: mail queue is empty...
[13:31] <jonah> brigante: maybe that test command line is a bit wrong somewhere, I tried both DEVICESCAN /dev/sdc -m myemail@gmail.com -M test
[13:31] <jonah> brigante: and without DEVICESCAN in front too...
[13:31] <rbasak> caribou: you don't have to use xgit git/gitwd if you don't want to. None of the tools require it. I find it handy to not worry about the .git directory being inside the working tree, that's all.
[13:32] <brigante> jonah, you do have a /dev/sdc right?
[13:32] <caribou> rbasak: yeah, I kind of like this fact.
[13:32] <brigante> jonah, type in `lsblk` to show your block devices (disks)
[13:33] <rbasak> caribou: thank you for the note about MP creation. I'll take a look. I've not tried it myself yet so I'm not sure what you mean, but your MP looks reasonable to me. It's showing the new diff against Debian, which seems reasonable. For more diffs I had to clone your repo, but that also seems reasonable.
[13:33] <jonah> brigante: that brings up sda, sdb, sdc and sdc and all the raid partitions and such too as well as md1,md0 etc
[13:33] <brigante> ok kool
[13:33] <brigante> just checking
[13:34] <jonah> brigante: thanks...
[13:34] <brigante> jonah, i would test your MTA is working correctly
[13:34] <caribou> rbasak: btw, you'll see a few results of the multiple questions I asked two weeks ago about removing conffiles, etc
[13:34] <jonah> brigante: right... what's the best way to do a test?
[13:35] <brigante> `mail -s "testing!" your@gmail.com
[13:36] <brigante> then type A BODY OF THE MESSAGE
[13:36] <brigante> THEN PRESS CTRL+D
[13:36] <brigante> oops caps lock
[13:36] <jonah> brigante: ok and it just says EOT
[13:37] <jonah> brigante: ah yes got that email in my gmail ok, just comes through as root@servername
[13:37] <jonah> brigante: so mail must be ok...
[13:37] <brigante> bingo
[13:37] <brigante> so your mta is fine
[13:37] <brigante> must b problem with smartmontools
[13:37] <brigante> the line without DEVICESCAN is the correct one i think
[13:38] <brigante> not sure why its not working for you
[13:40] <jonah> brigante: ah found this: http://serverfault.com/questions/426761/is-smartd-properly-configured-to-send-alerts-by-email
[13:42] <brigante> hmmm seems to differ to the man pages
[13:42] <brigante> i hope you get to the bottom of it
[13:43] <jonah> brigante: could it be because I've just uncommented DEVICESCAN -d removable -n standby -m EMAILADDRESSHERE  -M exec /usr/share/smartmontools/smartd-runner
[13:43] <jonah> brigante: but haven't uncommented anything else? or that that line doesn't work with just "removable" there...?
[13:44] <jrwren> be sure to file a bug if the man page is wrong.
[13:44] <jonah> brigante: maybe I don't need that line at all and should just uncomment this line DEVICESCAN -S on -o on -a -m myemail@mydomain.com -s (S/../.././02|L/../../6/03)
[13:45] <brigante> i would comment out the line with EMAILADDRESSHERE
[13:45] <brigante> and yes, uncomment the latter
[13:45] <brigante> replace the eail address with your own
[13:45] <brigante> email*
[13:46] <jonah> brigante: but the line with EMAILADDRESSHERE was the one that doesn't seem to work for me...
[13:54] <brigante> jonah, comment out means put a # at the start so the program IGNORES the line
[13:54] <brigante> i.e. ignore the line that doesn't work
[13:55] <brigante> uncomment, means remove the hash from the start of the line so the program uses it
[13:56] <jonah> brigante: ah got it working i think...
[13:56] <brigante> with that -M test line it should send a test email everytime the daemon is restarted
[13:57] <jonah> brigante: all it was is that DEVICESCAN doesn't work for me. So I ended up adding a line for each drive like in this tutorial: https://blog.shadypixel.com/monitoring-hard-drive-health-on-linux-with-smartmontools/
[13:57] <jonah> brigante: then when I added the test line at the end of each of those lines I get an email! thanks for helping me out!
[13:58] <brigante> haha excellent
[13:58] <brigante> you're welcome dude
[13:58] <jonah> brigante: thanks again, we got there in the end
[13:59] <brigante> now if only i could get some help with my problem...
[14:07] <doublel93> hey, I installed ubuntu 16.04, shouldn't python be there by default ?
[14:07] <brigante> I have a server with 3x physical network cards. Each card has 1x IP address all in the same network 192.168.0.0/24. Each NIC is wired to the same switch. The problem, when I ping the IP addresses: ALL traffic goes to eth0... wtf
[14:08] <brigante> doublel93, probably
[14:08] <brigante> doublel93, i just installed 16.04 on my laptop and i have python 2.7
[14:08] <jonah> brigante: good luck, I wish I could help you with that one, just one network card for each server for me...
[14:09] <brigante> jonah, thanks mate. I have never run into this situation before. its a real conundrum
[14:10] <doublel93> I come from a centos background, on installation centos ask you for ip, gateway,... of the server, in ubuntu 16.04 it didn't ask me anything about it, how do you go about changing them in ubuntu ? is there a wizzard ?
[14:10] <brigante> doublel93, no - edit /etc/network/interfaces
[14:10] <doublel93> ok thanks
[14:11] <sdeziel> brigante: if you want your 3 NICs to be in the same network, you probably want bonding https://help.ubuntu.com/community/UbuntuBonding
[14:13] <brigante> sdeziel, thanks, would that make my 3x NIC's into 1x 3Gbit NIC?
[14:13] <sdeziel> brigante: kinda ;) It depends on the way you setup your bond
[14:14] <kyle__> How do you stop dnsmasq-dhcp from responding?  service dnsmasq stop claims to kill it, and I see no processes running, but the damndable thing is still consuming dhcp requests as noted in syslog
[14:17] <doublel93> ifconfig -a | grep eth is showing 0 results, this is a new installation, is there something that needs to be configured ?
[14:25] <genii> doublel93: Network devices now get unique names similar in principle to UUID scheme of hard disks. So you will not see names now like eth0 eth1 and so on. You will see a name like enxff99ee88 with 3 letters and then 8 characters of it's physical MAC address, or a similar name to this
[14:25] <doublel93> I found ens32
[14:25] <doublel93> still googling why =p
[14:27] <genii> doublel93: The basic idea is same as why UUID for hard drives. So that each device has a universally unique name and cannot be used by mistake
[14:27] <kyle__> Internally, they're actually still called eth#.  udev renames them by default now.
[14:27] <Odd_Bloke> doublel93: See https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/ :)
[14:27] <ogra_> "predictable"
[14:27]  * ogra_ always found that a good joke :P
[14:27] <kyle__> You can pin which devices goes to which device by MAC address
[14:28] <genii> I blame Dell.
[14:28] <kyle__> ogra_: Like most things the udev/systemd team have pushed, It makes sense in some slim selection of cases, and confuses the hell out of every other one :P
[14:28] <patdk-wk> harddrives don't have uuid
[14:28] <patdk-wk> they have wwn
[14:28] <ogra_> kyle__, yep
[14:28] <kyle__> patdk-wk: I thought wwn was scsi or iscsi only actually....
[14:29] <patdk-wk> wwn is scsi
[14:29] <patdk-wk> harddrives don't have uuid
[14:29] <patdk-wk> sata has nothing
[14:29] <patdk-wk> filesystems have uuid though or something else, depending on the filesystem
[14:30] <patdk-wk> the names of my nics are really annoying now, enp3s133f1
[14:30] <kyle__> sata & ata do have serial numbers.  Not sure if those are always in the same place though.
[14:30] <patdk-wk> they don't ALWAYS have serial numbers
[14:31] <ogra_> patdk-wk, what ? and you didnt *predict* that name ?!?
[14:31]  * ogra_ grins
[14:31] <kyle__> patdk-wk: What's fun is a naming/renaming but in the 14.04 installer, that occasionally fails to rename all the nics if you have > 4.
[14:31] <patdk-wk> nope, that name changes based on the pci slot it's plugged into
[14:31] <patdk-wk> the 133 is the pci slot id
[14:31] <kyle__> "Oh, I have rename7 again, guess i have to reboot the installer and hope for better luck"
[14:31]  * kyle__ grumbles
[14:32] <patdk-wk> I don't understand why there is a renamexxx at all
[14:32] <patdk-wk> it has done nothing but screw up systems since before trusty
[14:33] <patdk-wk> I know I filed a bug about it, it was *fixed*, but problem continued
[14:34] <patdk-wk> think it got marked as a wont fix
[14:34] <Odd_Bloke> If only there were some way that those interfaces could get predictable names so this wasn't an issue. ;)
[14:35] <genii> If you have scripts that used to do something with an interface and relied on the names being eth0 eth1 and so on, it gets pretty annoying
[14:36] <genii> To revert though, you just need net.ifnames=0 in GRUB
[14:38] <brigante> i tend to use udev rules to specify my own names for nics
[14:39] <brigante> bind the name to the mac address
[14:41] <genii> That was the way before, yes, edit /etc/udev/rules.d/70-persistent-net.rules but it no longer works unless you disable the net.ifnames
[14:41] <patdk-wk> I had never had an issue with the old way
[14:41] <patdk-wk> using udev to match the nic name to the mac address
[14:42] <compdoc> 70-persistent-net.rules could get screwed up if you swapped nics, or installed new ones
[14:42] <patdk-wk> how so?
[14:42] <brigante> genii, i just setup an 16.04 server and used udev rules without the grub flag set... seems to work fine
[14:42] <patdk-wk> never had it happen
[14:42] <patdk-wk> if you swap nics, just update the mac with the new one, reboot, done
[14:42] <brigante> patdk-wk, maybe if you remove the nic before updating udev, i guess it might hang on boot?
[14:42] <patdk-wk> it won't hang at boot
[14:43] <brigante> ofc it just wont rename the missing nic
[14:43] <patdk-wk> yep
[14:43] <brigante> so whats the point of the grub flag again?
[14:43] <patdk-wk> if it hangs cause a nic is missing, your system has very bad boot scripts installed
[14:44] <brigante> agreed
[14:44] <genii> brigante: Because the interface naming is now in the kernel, it turns that off
[14:45] <patdk-wk> kernel? or systemd?
[14:45] <brigante> so if i dont set udev rules, eth0 is renamed by the kernal to enps0lmnop32xyz ?
[14:45] <brigante> and if i set the grub flag, it will remain as eth0 ?
[14:46] <teward> this may sound a little odd, but is there any decent guide for setting up an email server, complete with antispam and antimalware solutions, from a pure Ubuntu box?  Would more or less be a 'to learn how everything works' thing, but need to start somewhere :P
[14:46] <brigante> teward, check out the tutorials on digitalocean.com
[14:46] <genii> patdk-wk: It was coming in before upstart->systemd, so it's not something systemd
[14:46] <patdk-wk> but that used to be the biosdev??? package before
[14:46] <patdk-wk> I haven't looked in 16.04 though
[14:47] <genii> patdk-wk: Yes, but now it is not
[14:47] <patdk-wk> but in 14.04 you just uninstall that, and done
[14:48] <genii> biosdevname is a package you could uninstall or tell GRUB biosdevname=-1 or =0. But now it's in kernel, so you need either edit sysctl.conf or GRUB
[14:48] <patdk-wk> sysctl.conf would be seriously late to adjust a kernel eth rename
[14:49] <patdk-wk> I still don't believe it's kernel thing though
[14:49] <brigante> interesting
[14:49] <patdk-wk> cause my 16.04 system boots with eth?? devices
[14:49] <patdk-wk> and it renamed them later
[14:49] <patdk-wk> if it was kernel, the eth?? would have never existed to start
[14:50] <brigante> i got one for you guys: i recently tried to move my LVM pv's into an encrypted partition which used luks, i rsync'd my boot partition out of LVM to /dev/sda1, ran grub-install, everything seemed OK. But on reboot, I was not prompted to unlock the crypt and grub sat there asking "where's the root drive?"
[14:50] <brigante> how do you tell ubuntu that its root is inside a crypt?
[14:50] <patdk-wk> brigante, did you bother to tell grub it is?
[14:50] <patdk-wk> there is some good adjustments you have to make :)
[14:50] <nacc> rbasak: caribou: i've got the script working with our exim4 example, at least
[14:51] <nacc> rbasak: caribou: didn't get a chance to push last night, as i was still testing the parent overrides, let me finish that and push it up and you should be able to use the same script with corosync
[14:51] <brigante> patdk-wk, perhaps this is what i'm missing...
[14:51] <patdk-wk> says here, the nic renaming is systemd
[14:51] <nacc> rbasak: the override seems to have worked for clamav (not exim4 as mentioned above, sorry!) so that's good!
[14:51] <caribou> nacc: I've already merged i the traditional way (only one patch delta) so it'll be quick to test
[14:52] <brigante> whats the deal with systemd in ubuntu anyway, it seems half in half out
[14:52] <nacc> caribou: yep, was just reading the backlog :)
[14:52] <nacc> caribou: the script basically encapsulates the manual commands rbasak gave for git
[14:52] <brigante> no netctl for example (that is a systemd component right?)
[14:53] <nacc> netctl is an ... arch linux thing?
[14:53] <brigante> nacc, im fuzzy on it tbh, i've only ever used it in arch so maybe
[14:53] <nacc> brigante: systemd is (aiui) 'fully' in ubuntu...
[14:53] <patdk-wk> systemd-networkd to be specific
[14:53] <brigante> i must have my wires crossed then
[14:53] <brigante> thanks nacc
[14:54] <brigante> i thought systemd was taking over everything
[14:54] <brigante> systemd alarm clock etc
[14:54] <patdk-wk> I thought it already did
[14:54] <patdk-wk>  Ican't believe how many CVE's it *reopened* for past systems that solved these problems years ago
[14:54]  * patdk-wk looks at systemd dns resolver
[14:56] <Odd_Bloke> genii: patdk-wk: The renaming is systemd, but it isn't networkd (which Ubuntu does not yet use).
[14:56] <patdk-wk> hmm
[14:56] <patdk-wk> ya, I knew it couldn't be the kernel atleast
[14:56] <Odd_Bloke> (It may be in the networkd code, but it doesn't require networkd to be running)
[14:57] <brigante> Odd_Bloke, does networkd contain the netctl command? and is part of the systemd family?
[14:58] <genii> Interesting
[14:58] <brigante> or is netctl specifically an "arch thing"
[14:58] <brigante> im still fuzzy
[14:59] <brigante> i quite like netctl
[15:00] <brigante> ok i just checked, netctl is arch specific and nothing to do with systemd/networkd
[15:04] <doublel93> I modified the "interfaces" file and installed openssh-server, how can I see if my server actually has the ip I set ?
[15:05] <brigante> doublel93, run `ip addr`
[15:05] <brigante> doublel93, also run `ss -ntlp` to see if SSH is listening
[15:07] <doublel93> brigante: it is listening on *:22,  instead my ens32 interface says inet 192.168.1.161/21 brd 192.168.1.255 scope global ens32 (I tried to set it to 192.168.1.48)
[15:08] <EmilienM> coreycb, jamespage: FYI: module some issues we ignore now (documented on https://etherpad.openstack.org/p/puppet-openstack-xenial) - we now gate on Xenial (and not trusty anymore) for our master (current newton). But still deploy Mitaka.
[15:08] <brigante> doublel93, you may need to `ifdown ens32` and then `ifup ens32` to get the new IP
[15:08] <EmilienM> coreycb, jamespage: any ETA on "be able to test newton repo"?
[15:09] <doublel93> brigante: nothing changed
[15:09] <brigante> doublel93, try a reboot
[15:10] <brigante> thats one thing i dislike about ubuntu/debian... how the fudge do you restart networking properly
[15:10] <brigante> `service networking restart` seamingly does nothing
[15:11] <Odd_Bloke> brigante: Well, it does an ifdown/ifup. :p
[15:11] <brigante> i always end up rebooting... someone tell me there is a better way? im not talking about setting up manually using iproute2 commands either... how do you tell ubuntu to reload the /etc/network/interfaces config?
[15:11] <brigante> Odd_Bloke, lol
[15:11] <coreycb> EmilienM, \o/ on the xenila move.  newton pkgs should be in proposed a few days after upstream releases newton b1, which should happen any day now.
[15:11] <doublel93> brigante: even after a full reboot it didn't change, I just came to ubuntu from centos, such a pain
[15:12] <brigante> doublel93, is that a DHCP address?
[15:12] <EmilienM> coreycb: ok cool
[15:12] <doublel93> brigante: this image is what I did https://mjmckinnon.files.wordpress.com/2016/05/edit-interfaces-file.png?w=776&h=454  ,  except I used the ens32 name
[15:12] <coreycb> xenial, that is
[15:13] <brigante> doublel93, that looks fine to me
[15:13] <brigante> double check your config doesn't have dhcp instead of static defined
[15:13] <doublel93> brigante: address is the one that has to become 192.168.1.48 right ?
[15:13] <brigante> yup
[15:14] <brigante> Odd_Bloke, so how do you reload the /etc/network/interfaces config? or do you just reboot like me?
[15:15] <brigante> ubuntu seems to be stuck in the past when it comes to network configs, probably the fault of debian
[15:15] <doublel93> brigante: ok I feel stupid, that was it dhcp was left
[15:15] <doublel93> but now I have 2 in ip addr, global and global secondary, is that good ?
[15:15] <nacc> brigante: keep opinion to #offtopic, please :)
[15:16] <nacc> brigante: rebooting to restart networking is ... not sane, I would be surprised if that was actually required
[15:17] <brigante> nacc, apologies i will refrain from opinion comments
[15:17] <brigante> nacc, so what is the advised way to reload the networking config?
[15:18] <brigante> ifup, ifdown, service networking restart all seemingly do nothing
[15:18] <nacc> brigante: does `service networking status` indicate it is active?
[15:18] <brigante> other than using iproute2 to specify the same thing as the config i dont see a way...
[15:18] <brigante> nacc, yes, active
[15:19] <brigante> active (exited) to be precise
[15:19] <brigante> perhaps it requires the `reload` argument
[15:20] <brigante> rather than restart?
[15:20] <doublel93> ok I have finally solved that problem thanks a lot, how can I doublecheck that I'm using my dns server now ?
[15:20] <brigante> doublel93, use the dig command
[15:20] <nacc> brigante: then i expect a `service networking restart`  to dtrt, but you can laways check that
[15:20] <nacc> stop followed by start does work
[15:21] <nacc> in my case, in a simple lxc container, stop brings down the interfacae and a start brings it back up
[15:21] <brigante> nacc, yikes, good luck doing that remotely :)
[15:21] <nacc> brigante: i assume if you're mucking with networking, you have another means of access (serial console or otherwise)
[15:22] <doublel93> brigante: I do not see any of my dns, (the dns server is configured on another one), dns-nameservers 192.168.1.67  was the right option to change to point to that ?
[15:22] <brigante> not always, i have rented dedicated servers in the past without serial access
[15:22] <brigante> alas, thats why virtual machines are a god send tho
[15:23] <brigante> still seems odd that a `restart` doesn't actually do anything
[15:23] <nacc> brigante: it does seem like `service networking reload` will do what you want
[15:23] <brigante> aha
[15:23] <brigante> well thats good to know
[15:23] <nacc> brigante: but i don't see why that is safer if you've chagne the config
[15:23] <nacc> for remote modifications
[15:24] <brigante> if you `service networking stop` you cant start it again as ssh will drop...
[15:24] <brigante> hence reload is much safer imho
[15:24] <nacc> also, `service networking restart` absolutely does something here, so i'm not sure what your config is like
[15:24] <brigante> hmmm perhaps its fixed in latest ubuntu
[15:25] <brigante> i've been using ubuntu since v6 and it has bugged me for years
[15:25] <nacc> brigante: what version are you on?
[15:25] <brigante> 16.04 currently, but still manage servers with 12.04 and 14,04
[15:26] <nacc> 16.04 is the latest, and was what i was testing on (in a container)
[15:26] <brigante> i will test now too
[15:28] <JanC> teward: https://help.ubuntu.com/16.04/serverguide/email-services.html has the basics
[15:28] <teward> JanC: thanks
[15:31] <brigante> nacc, nope. it does nothing, either `reload` or `restart` doesn't update the interface
[15:32] <brigante> i incremented the last octet in the static ip address by 1, tried to reload/restart, and I get the same IP i had on boot
[15:33] <brigante> for a minute i thought i had been a fool for many years...
[15:33] <brigante> i'd happily be the fool if someone can tell me the correct way to do this without a reboot or using iproute2 commands
[15:35] <brigante> nacc, what test did you run?
[15:37] <nacc> brigante: both reload and restart and checked what syslog said
[15:40] <brigante> nacc, try the test I just did. change the ip address.
[15:40] <brigante> this is what im talking about. reloading the config
[15:41] <brigante> just cos syslog says "it worked" doesn't mean your IP address has changed
[15:41] <nacc> no ... but asserting it doesn't do anything is also false ... :)
[15:41] <nacc> that's all i was checking
[15:41] <brigante> lol fair enough
[15:41] <brigante> i'll chose my words more carefully, blimey...
[15:45] <nacc> brigante: :)
[15:45] <nacc> brigante: let me change the lxc config
[15:46] <brigante> i have not tested or tried lxc
[15:53] <nacc> brigante: well, odd, lxc started with dhcp, so it is refusing to give up the lease, but adding a static gave the device a second ip as configured on networking restart
[15:53] <nacc> brigante: unfortunately don't have more time to spend on it right now
[15:54] <brigante> nacc, well thanks for clarifying. this "bug" has occured in every version of debian and ubuntu i have ever used
[15:54] <nacc> brigante: and i've never experienced it :)
[15:55] <nacc> brigante: which isn't a good case, just saying i've not seen that issue
[15:55] <brigante> it seems most people haven't
[15:56] <brigante> who knew that changing an IP address was such a specialist thing lol
[15:57] <nacc> given the nature of cloud deployments, i would also think many configs are using dhcp or don't really care about the networking details, as long as its routing properly -- dunno
[15:58] <brigante> nacc, perhaps dude. it's just always struck me as odd
[16:00] <brigante> anyway - im off home now, thanks to everyone for their help today
[16:00] <ddellav> coreycb can you take a look at keystone, It builds successfully for me after I updated the d/control for oslo-utils: lp:~ddellav/ubuntu/+source/keystone
[16:01] <coreycb> ddellav, sure
[16:04] <Yuri4_> Hi, guys! I have 2 servers. There is a folder /mount that has wordpress files that are syncronized accross servers via SMB 3.0 protocol. What is the best way to sync that folder with my Wordpress that is located at /var/www/html ?
[16:05] <coreycb> ddellav, that should work
[16:05] <ddellav> coreycb ok, ironicclient is also ready for review: lp:~ddellav/ubuntu/+source/python-ironicclient
[16:05] <coreycb> ddellav, just changing back to 0ubuntu1
[16:06] <coreycb> ddellav, ok
[16:06] <ddellav> coreycb oh ok, i never know if i should step on someone elses commit message
[16:07] <coreycb> ddellav, you wouldn't be
[16:07] <coreycb> ddellav, if you clone it and it's UNRELEASED then that version wasn't ever uploaded
[16:08] <ddellav> coreycb ok,  in keystone my changelog message was in the same block as yours and jamespages so i wasn't sure if updating d/control was enough to modify the version number
[16:08] <ddellav> coreycb or did dch change it to 2
[16:12] <coreycb> ddellav, keystone pushed, want to manually retrigger a build in jenkins?
[16:12] <coreycb> ddellav, one for xenial and one for yakkety
[16:15] <jrwren> Yuri4_: rsync or unison both sync folders very well.
[16:16] <Yuri4_> jrwren, which one is better? Don't have rsync expirience and this is the first I hear of unison
[16:17] <jrwren> Yuri4_: rsync is more widely used and it sounds like you are syncing 1 direction. use it.
[16:17] <jrwren> Yuri4_: unison is better when changes might be in both places, a 2 way sync.
[16:19] <Yuri4_> jrwren, yes! Thank you! The source folder /mount might become disconnected from time to time. Might that be an issue? Also, should I create a script that rsyncs changes every 20 seconds, or can rsync run forever?
[16:22] <jrwren> Yuri4_: cron is a good tool for running periodic jobs.
[16:22] <jrwren> Yuri4_: it is probably best to check that the folder is mounted and to do nothing if it is not mounted.
[16:23] <Yuri4_> jrwren, no, I want it to be run constantly. But if the folder drops rsync won't delete everything, would it?
[16:24] <Yuri4_> jrwren, basically it's for a WordPress website running on 2 linux VMs in Azure. From time to time Azure shuts down VMs for 30 minutes. I'm building a web site that should not go down
[16:26] <jrwren> Yuri4_: there is no such thing as "constantly" unless you want "while true ; do rsync... ; done"  which I don't recommend.
[16:26] <Yuri4_> jrwren, why not?
[16:26] <jrwren> Yuri4_: IIRC there are wordprss plugins that move teh static storage out of the local filesystem. there maybe one for azure blobstorage. I'd recommend that.
[16:26] <ddellav> coreycb ok i'll try heh
[16:27] <Yuri4_> jrwren, thank you
[16:27] <jrwren> Yuri4_: its a waste of cycles. Ideally you'd not do anything unless soething is written and changed. You could use inotify for that.
[16:27] <jrwren> Yuri4_: but yes, look into not using local filesystem for those files at all. Its a great solution.
[19:42] <wedgwood> I can't tell from https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1573231 when that patch will be released in AMI form. Anyone here have any idea?
[20:07] <dannf> hey hallyn: would you mind taking a look at LP: #1566564 and letting me know if you have any objections to me uploading xenial SRUs?
[20:16] <coreycb> ddellav, I'm working on a new python-hacking, that should fix up neutron for newton
[20:17] <ddellav> coreycb awesome
[20:30] <coreycb> ddellav, it looks like trove just needs it's patches rebased
[20:31] <coreycb> ddellav, I'll let you do that one
[20:31] <ddellav> coreycb ok, i'll take care of it
[20:31] <coreycb> ddellav, thanks
[21:02] <jdeler> hello guys i have registered a vhost on port 3000 locally and how can i start the server
[21:02] <jdeler> i am lill new here for apache2 server
[21:14] <Sling> jdeler: service apache2 start/stop/restart
[21:14] <Sling> just like any other service
[22:58] <sdeziel> nacc: if you could check my new debdiff attached to LP: #1570472, I'd really appreciate :)
[22:59] <sarnold> lol "upstart is required to confuse Puppet"
[22:59] <sdeziel> sarnold: don't laugh, I was really confused when my lxc container wouldn’t let me reproduce the issue ;)
[23:00] <sarnold> sdeziel: apparently it's also suitable for confusing sdeziel and amusing sarnold :)
[23:00] <sdeziel> lol
[23:01] <sdeziel> next time, I'll try to reproduce an issue before looking into patching it ;)
[23:01] <sarnold> hehe
[23:01] <sarnold> strange that the specific version numbers are listed, rather than allowing open end points
[23:02] <sarnold> that's just going to need to be edited again in four months..
[23:02] <sarnold> but hey looks like that's the way they chose to go
[23:02] <sdeziel> I guess that's because they don't support "greater than" when there is a "." in the version number
[23:06] <sdeziel> hmm, I'll update that debdiff to fix the changelog address
[23:30] <ahi2> apache2 runs on different ports but only accessible from web on 80? anyone know why?
[23:32] <sarnold> ahi2: check netstat -tnlp to see which addresses it is bound to -- and check your firewall, if one is installed, to see if those other ports / addresses are open
[23:33] <ahi2> apache listening on 7950
[23:35] <ahi2> firewall rule good