[00:20] Hrmm, I'm trying to hide a .git folder from a samba share and it doesn't seem to be working, which is kindof baffling me. [00:21] I would've assumed, particularly since it's in the root of the share, that [00:21] hide files = /.git/ [00:21] would've been sufficient, but apparently not? [00:22] Maybe I'll just try the "hide dot files" option and be done with it. [00:23] Wait, *that* didn't work? [00:28] * keithzg reads the documentation further, figures out he actually wants to "veto" the folder, aha [00:48] cpaelzer, \o/ libvirt [01:22] hello. I have apache2 up and running and just installed PHP7 via apt. If I want to use the latter with the former, do I install the `libapache2-mod-php7.0` package, or what's the usual way to do it? [02:30] adrian_1908, personally I recommend using php-fpm, but it's not nearly as simple and easy as mod-php [02:32] patdk-lap: yeah I need easy, I'm a total beginner :) [02:32] well, just install the mod-php then, it should do everything needed by itself [02:34] Yes, I already did and am busy with it. It activated itself in Apache and appears to have changed the MPM model (I was already wondering about how I'd accomplish that myself). Very convenient. [02:55] Damn, the `mysql-server` package sure has a lot of dependencies (perl stuff). Didn't expect it too be that large. [02:58] hi [02:58] anyone know alot about LXD? [02:59] i'm just trying LXD on 16.04 first time running on a ZFS root, looking to set up LXD with MACvlan based bridging [03:00] i installed it using DIR as storage, because the underlying ilesystem is already ZFS [03:01] but does that mean LXD can't use zfs snapshots, instead it plays around with compressed files and rsync? [03:02] and does that impact performance? i'm referring to the comments section in this article: https://bayton.org/2016/05/lxd-zfs-and-bridged-networking-on-ubuntu-16-04-lts/ [03:03] secondly, the concept on maCvlan is new to me, and i'm looking to try to learn if it could be used to replace IOMMU/PCI passthrough. Anyone here know alot about MACvlan/MACvtap? === rxc is now known as Guest55099 === ReScO_ is now known as ReScO === cmh is now known as cmh_ [06:20] thanks xnox [08:06] Good morning. [09:11] anyone with apt module error "FATAL -> Failed to fork" http://vpaste.net/izPXA don't say with apt-get install -f ... it emits the same error messages [09:22] failed to fork? how much free RAM do you have left? [09:27] hello guys im having an issue with ubuntu 12.04 with ipv6. after some time my routing tables are getting full and then ipv6 stops working [09:36] why are you routing tables "full' [09:36] it would be a HUGE networking route table to fill a host [09:39] little odc http://vpaste.net/ZPYp6 any suggests? [09:41] ikonia i dont know why they are getting full i think its a bug [09:42] ws2k3: are they actually "full" or do they just look full [09:42] ikonia i have around 100 ipv6 hosts in my network and i have a /64 [09:42] eg: how are you determaining they are "full" [09:42] ikonia how can i check that? [09:42] ws2k3: hang on [09:42] do you know if they are full or not [09:43] fishcooker: suggestion on what ? [09:43] fishcooker: Try to free up some ram any try again. [09:44] ikonia wc -l /proc/net/ipv6_route shows me 21864 [09:44] and on some machines im already getting wc: /proc/net/ipv6_route: Cannot allocate memory [09:44] 0 /proc/net/ipv6_route [09:46] ws2k3: so you're out of memeory - is that because of the routes though [09:46] ws2k3: what's allocating these routes [09:48] ikonia i dont know exacly [09:48] ws2k3: are they 1:1 host routes [09:49] or network routes [09:49] ikonia i do know that im getting the canot allocate memory [09:49] ikonia they are just webserver with 1 ipv6 adres and 1 ipv6 default gateway [09:49] ws2k3: what are the routes to then [09:50] can you actually display the routing table ? [09:50] yes i can i think you mean ip -6 route show cache? [09:51] i sended my ip -6 a in private message ikonia [09:52] not cache [09:52] ws2k3: based on what you've just pasted your networking setup is broken [09:52] you have dead routes on the loopback interface [09:52] ikonia i know they are for a load balacer [09:52] you also have conflicting routes [09:53] you have the same network routing out of two interfaces [09:53] your network setup appears to be screwed up looking at that route table [09:53] i think i need to delete this one 2001:1aa8:1850::/64 dev eth0 proto kernel metric 256 expires 1906251sec [09:54] as I don't know your networking setup, I can't comment on what's valid, what's not valid, but you had dead routes, you have conflicting routes and you have bad routes - those 3 things are not good and suggest you have overall network problems [09:57] ikonia the dead routes should not be an issues cause its just a local loopback adres [09:57] and the conflicting route is a different subnet [09:57] it is an issue though [09:57] ws2k3: [09:57] ws2k3> fe80::/64 dev eth0 proto kernel metric 256 [09:57] that is NOT a differnt subnet [09:57] 09:51 fe80::/64 dev eth1 proto kernel metric 256 [09:58] look how wide that network is too [09:58] (I hope that is acceptable as it's not secret information on a subnet that wide) [10:00] fe80 is link local, each ipv6 nic should have that. [10:00] Though wikipedia lists it as fe80::/10 [10:01] right [10:01] hence why I'm saying you've got conflict [10:01] as it's set to route both that out of two different physical interfaces and look how wide it is [10:01] I don't know the network setup on this host so I don't know what's needed/not, however looking at the route table it's not good [10:02] Must say I have the fe80::/64 for each interface too. [10:03] really ? [10:03] that seems very odd [10:03] ikonia: You don't? [10:04] no [10:04] but then the only ipv6 network I'm on is really simple [10:07] by now on allmost all my servers ipv6 is broken [10:09] ws2k3: just because of the number of routes [10:09] or something else [10:09] ikonia my machines are getting into trouble after i get this wc: /proc/net/ipv6_route: Cannot allocate memory [10:10] right - I get that you're out of ram [10:10] but are you out of ram due to the number of routes or just out of ram [10:10] ikonia machines have plenty of ram free [10:10] ok, so if you look at the route table you pasted there are only a few individual routes [10:11] so what's causing the population of that ? the dead routes ? [10:11] ikonia thats why im trying to find out [10:11] What do you get when you try to ping hosts? [10:11] lordievader most of them dont reply [10:11] And on the hosts themselves? [10:12] lordievader also no reply [10:12] lordievader i assume you mean ping from the host to the internet for example? [10:12] Do you see ipv6 traffic from them on your gateway? [10:12] For example, yes. [10:13] lordievader i dont have access to the gateway so i cant check [10:13] lordievader but i cant ping the gateway nor any other host [10:13] i have now run sysctl -w net.ipv6.route.max_size=8388608 and then it temperarly works again [10:14] ws2k3: hang on [10:14] ws2k3: metric 256 !!! [10:14] ws2k3: thats 256 hops it has to know to get to that route [10:16] ikonia as far as i know metric is the priority not the amount of hops [10:16] maybe thats my mistake then [10:17] your right, my apologies [10:17] ahhh the metric can include number of hops [10:17] so yes, I am right, but thats not the "factual" answer in terms of a "solid number" [10:17] sorry about that [10:18] ikonia np im happy with all the help i can get with this strange issue [10:18] But, if I understand correctly, the host just adds routes as if they are candy? [10:18] it's a large number of routes in /proc for sure, but looking at the route table the individual routes as few [10:21] ikonia and thats exacly the problem [10:22] ws2k3: how are the routes getting populated ? [10:22] ikonia i dont know [10:22] (the ones that are in the route table no the 10000000 other ones) [10:22] ikonia i just have one default gateway and one ipv6 address and thats it [10:23] Are privacy extensions enabled? [10:23] ikonia it looks like im having this issue https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1065159 [10:23] Launchpad bug 1065159 in linux (Ubuntu) "ipv6 routing memory leak" [Medium,Expired] [10:23] wow - a memory leak [10:23] medium "expired" [10:24] ikonia this looks like the same behavior that im having [10:25] ws2k3: What kernel are you running? [10:26] lordievader 3.2.0-57-generic #87-Ubuntu SMP [10:27] ws2k3: Well you could do the same as in the comments, run a newer kernel and see if the problem persists. [10:27] 3.2 is old anyways ;) [10:27] on newer kernels i am not having this issue [10:27] ws2k3: parts of it certainly map to that bug in terms of behaviour [11:38] buy some RAM ikonia :p [11:38] noted lordievader... it works === giraffe is now known as Guest73411 === madwizar1 is now known as madwizard [12:17] Buy? Download: http://www.downloadmoreram.com/download.html [13:48] coreycb: neutron is blocked on pyroute2 so I think we might want to drop it for now to get b2 things going and then re-add it [13:49] zul, i think we're ok. we can test b2 in proposed. [13:49] coreycb: ok [13:50] coreycb: you uploaded sahara? [13:50] zul, yep and a few others [13:51] coreycb: okie dokie [14:43] does pm-suspend work in qemu-kvm virtual machines? [14:43] it did nothing for me =( [14:46] xnox: my xenial guest goes into state "pmsuspended" [14:46] cpaelzer, cool. what about trusty? =) with xenial kernel? [14:46] * xnox is weird, I know [14:46] xnox: umm I can try [14:47] cpaelzer, please, as i'm on a frankenstein trusty. [14:52] xnox: T on 3.13.0-105-generic working as well, no rebooting into linux-generic-lts-xenial [14:53] xnox: the same - working [14:53] xnox: also wakeup works a la "virsh dompmwakeup testguest-T-Xkernel-pmsuspend" [14:53] cpaelzer, nice.... now i just need to fix why it doesn't for me =/ (as in what did I break) [14:53] cpaelzer, you suspend using $ sudo pm-suspend ? [14:54] yes [14:54] cool. [14:55] FYI /var/log/pm-powersave.log http://paste.ubuntu.com/23746485/ [14:55] FYI /var/log/pm-suspend.log http://paste.ubuntu.com/23746488/ [14:55] xnox: ^^ [14:57] for me it returns 1 and doesn't log anything in those files =( [14:57] xnox: my host is xenial [14:57] ditto [14:58] xnox: set -x in /usr/lib/pm-utils/bin/pm-action ? [15:00] better [15:00] LOGGING=1 pm-suspend [15:00] pm-utils does not know how to suspend on this system [15:03] Why suspend a vm when you can do a managed save? [15:04] lordievader, just testing if my userspace changes as to how suspend is initiated breaks things or not. To make sure I don't break real hw. [15:04] Ah, I see. [15:14] cpaelzer, $ cat /sys/power/state [15:15] has for me only "freeze disk" and no "mem" do you have "mem" in your VMs there? [15:15] * xnox ponders if I need to add more ram. [15:23] cpaelzer, my Ubuntu Trusty Desktop says pm-is-supported --suspend -> exit code 1, as in not supported, and I have no idea why. [15:27] cpaelzer, launched VM manually via qemu command line, instead of GUI point and click virt-manager and suspend is supported. [15:28] i guess virt-manager is goofing with me. [15:29] Or the xml doesn't contain the necessary config. [15:30] xnox: wait I'll check [15:32] xnox: reeze mem disk [15:32] +f [15:32] xnox: so yes I have mem [15:32] pm-is-supported --suspend is rc=0 [15:33] cpaelzer, yeah, doing a basic VM launch using qemu -enable-kvm -m 2048 -hda foo.qcow2 results in a suspendable VM, whatever xml virt-manager generated for the foo machine does not hibernate =/ [15:33] does not suspend that is. [16:09] coreycb: ping can you rediff your openstack-charm-testing branch i would like to test out the other stuff as well [16:09] zul, this? https://code.launchpad.net/~corey.bryant/openstack-charm-testing/add-new-svcs-to-sparse/+merge/313156 [16:10] coreycb: thats the one [16:10] zul, ok [18:40] hi can anyone please help. my server seems to have lost internet connection and can no longer be reached. From the server itself I can't ping google.com either... [18:40] jonah: what is the error message you get [18:41] ikonia: everything worked great, then all of a sudden a few hours ago I couldn't reached my website any more. Tried to ssh into the server and couldn't do that either. So I've plugged a keyboard in and accessed directly and the box won't ping google.com or other sites [18:41] ikonia: no idea what happened! [18:50] jonah, looks like networking is down === wwalker_ is now known as wwalker [19:19] pmatulis: really weird issue. all firewalls off etc. when i reboot the server internet/sites load up no problem. then when it has been on for maybe 15 seconds it can't connect all of a sudden again [19:20] jonah, i would concentrate on the fact that you cannot ssh to it, if that's still the case [19:24] pmatulis: I can ssh in initially when it first boots but as I say 15 seconds later the internet on that box drops and then I can't ssh in any more as there is no access in [19:24] jonah, and what kind of server is this? a cloud instance? [19:27] beisner, hello can you please promote all of newton-staging to newton-proposed? [19:31] pmatulis: i think it is due to resolv.conf only having 127.0.0.1 nameserver [19:32] when i add another nameserver to resolv.conf it wipes it on reboot... [19:32] jonah, right, that's normal [19:33] beisner, also neutron is ready to promote from mitaka-proposed and newton-proposed to -updates [19:37] hi coreycb, newton staging-to-proposed done [19:42] coreycb, neutron promoted to updates for mitaka and newton [19:51] beisner, thanks [20:03] is there a wiki somewhere that shows what debian versions ubuntu releases are based on? [20:04] cmh_: No, since they arent based on specific debian releases. [20:05] or, they're all based on the debian "testing" release. not very useful answer either. :) [20:05] (or is the sync against unstable? thought it was testing) [20:05] since 14.04 it's unstable i think [20:06] not 100% [20:06] maybe LTS -> testing , non-LTS -> unstable [20:07] https://wiki.ubuntu.com/LTS says Starting with the 14.04 LTS development cycle, automatic full package import is performed from Debian unstable [20:07] otherwise I think it's as pmatulis said [20:07] ok, so, thanks. easy then. all ubuntu releases are based on Debian Sid. :) [20:07] So its unstable for both LTS and non LTS :) [20:08] unstable = sid? [20:08] cmh_: yeah [20:16] bekks, since 14.04 yeah. i didn't know that [20:17] take a look at /etc/debian_version , for xenial (LTS) it shows 'stretch/sid' === Guest55099 is now known as rxc [22:51] is it possible to install php 5.2.anything on ubuntu server 16.04? [22:51] 5.6 is already installed but I need 5.2 [22:52] I keep getting erros like this: php-apcu : Conflicts: php-xcache but 3.2.0-2+deb.sury.org~xenial+1 is to be installed [22:52] how do I tell apt to only install the one from the xenial repo? [22:53] ninjai: not officially, no [22:53] ninjai: also, 5.2 is incredibly old and even unsupported upstream, afaict [22:54] While I was expecting someone to say that, I'm well aware but it's what the client wants to support their ancient web server [22:54] so they're fine with (potentially) CVEs...? [22:54] just trying to move it off of a physical freebsd box from 2007 or so, to a linux VM... [22:54] seems like a terrible choice to me, but not mine to make :) [22:54] not mine either [22:55] I just need to get it on and make it work [22:56] ninjai: as to your conflicts, you've add ondrej's ppa, which will conflict with the archive. And installing php5, period, on 16.04, will lead to issues with the archive (afaict), as ondrej has to provide his version of the deps in the ppa [22:56] ok [22:56] any way I can find out which version of ubuntu server had the option to install php 5.2? [22:57] ninjai: even 12.04 (the oldest currently supported) has 5.3 [22:57] ninjai: you could look at hte publishing history for php5 to go back further [22:58] ok thanks [23:12] hello, I'm a bit unsure what good practices are for folders & files inside `/var/www/` regarding ownership and permissions. Could someone help clarify this for me? For static websites I can probably just let root or my user own them, but what about sites like WordPress that need to write content? [23:12] I've found conflicting advice online. [23:12] (I'm using apache2 btw) [23:15] adrian_1908: it's not a hard set of rules, it's an evaluation of the setup [23:15] adrian_1908: for example the wordpress referenc eyou use, actually needs very little write access and it's very controlled of only a few directories [23:15] adrian_1908: thats very different than "/var/www" [23:16] (just for example) [23:17] ikonia: yeah, wordpress would be very different and it's the part I'm unsure about. I don't want to create some security risk. My thinking would be to set ownership inside the wordpress dir to `myuser:www-data` and give writable folders 775 permissions. [23:18] adrian_1908: it's not just that simple [23:18] the whole structure needs to be evaluated, [23:18] but you're approach of varition depending on needs is correct