[00:01] Anybody here ever tried to use automount inside a NAT'd LXC container to mount home directories shared via NFS? Can't get it to work on a NATd container [00:01] Waaaay back in 10.04, if you installed screen, it would pull in update-notifier-common, which would change the MOTD to remind you if there was a newer kernel to reboot into. [00:02] Can anyone tell me offhand if that's still the case? [00:02] That is: if I install update-notifier-common, will MOTD start warning about pending kernel updates [00:02] (I suspect this used to happen via byobu, and I have no idea what's happened in byobu.) [00:12] twb: yeah, update-notifier-common still does the motd updating stuff [00:12] Cool, so I can install just that and get that feature [00:13] http://packages.ubuntu.com/xenial/all/update-notifier-common/filelist [00:13] Thanks [00:14] twb: I think you might also need update-motd installed, if you don't already have it (looks like it's recommended by libpam-modules) [00:15] I'll find out if this change request gets approved :-) [00:37] hello [00:39] im newbie playing with ubuntu server 12.04 [00:40] i want to setup dns server but /etc/init.d/bind9 start fail [00:41] Unless you need to stick with 12.04, it is a good idea to install a newer release --- 14.04 will be supported for longer than 12.04. And I guess 16.04 is due out next month. === mikal is now known as chair_occupier_t === chair_occupier_t is now known as chair_occupier_3 [00:43] the issue with bind9 is: bridge.grumpy-troll.org/2012/05/pangolin-update/ [00:44] so if i upgrade the distro, it will work without no fix/workaround needed? [00:57] Don't know :-) [00:57] I use nsd3 rather than bind. [00:57] (But I'm weird, and I'm not a regular here.) [00:58] From the URL you gave, it's not clear to me what the issue is === Lcawte|Away is now known as Lcawte === Lcawte is now known as Lcawte|Away [01:57] how do you guys deal with updating distros [01:57] for many many servers [01:57] if $boss has his way, we just don't :/ [01:58] are LTS -> LTS updates reliable? [01:58] Assuming you use do-release-upgrade, they should be. [01:59] I've blown my foot off once or twice by doing something it didn't expect, like writing "auto lo eth0" in interfaces(5), which the upgrade script didn't understand. [02:00] its seems like there could me a lot of stuff that has to manually updated if you change anything [02:01] yeah, mmight be a bit more difficult if you have a lot of customization [02:01] In Debian, weird things are just documented. In Ubuntu, do-release-upgrade tries to automate it with scripts. [02:01] The weirder you are, the more likely the scripts are to fall over. [02:01] Because I'm very weird, I prefer the Debian approach :-) [02:02] I also had problems where if do-release-upgrade crashed halfway, it didn't make any attempt to roll back [02:03] Since most of my gear is stuck of 10.04, I can't comment on recent upgrades [02:05] seriously? [02:05] ya: 12:57 if $boss has his way, we just don't :/ === Lcawte|Away is now known as Lcawte [02:06] i get so much shit from the packaging community about wanting to put my packages on natty and saucy [02:07] i keep telling them its not realistic to expect everyone to be on the latest [02:07] i really dont like packaging === Lcawte is now known as Lcawte|Away [02:16] "it's behind a firewall so it doesn't matter if it's EOLd" === chair_occupier_3 is now known as mikal === n0Str3s5-AW is now known as n0Str3s5 [07:14] can anyone please help http://stackoverflow.com/questions/35221886/export-foreman-workers-from-rails-app-to-systemd ? [07:15] "foreman" is some kind of software? [07:16] yes, it's awesome [07:17] What version of systemd do you have? systemctl --version [07:18] Unfortunately systemd's "No such file or directory" doesn't tell you *which* file it failed to access. It could be seriapp.target, or some underlying file. [07:18] The fact that "systemctl start seriapp" reports seriapp.target instead of seriapp.service is suspicious. === Algorithm is now known as ^King [07:19] What systemd units are you using / have you written? [07:19] twb: systemd 215 [07:20] Also you wrote "export systemd /etc/init", which seems wrong to me -- /etc/init is for upstart, not systemd [07:20] systemd jobs typically go into /etc/systemd/system/ [07:20] twb: I've done it wrong but now rvmsudo foreman export systemd -a seriapp -u deploy says I must specify a location [07:20] twb: yes I removed that [07:20] hmmm should I try with /etc/systemd/system ? [07:21] I don't know. It is not clear to me what your setup currently looks like. [07:21] It sounds like you made several mistakes at once, fixed some, and some are still there [07:21] Can you pastebin a summary of what your system looks like right now? e.g. "systemctl", "systemctl cat seriapp.service" "systemctl cat seriapp.target", &c [07:22] twb: I 've looked for a log file but it seems there isn't any in /var/log [07:22] systemd writes to a "journal", this starts out in /run and moves into /var/log/ later. You read it with "journalctl" rather than by catting files in /var/log/. [07:22] If you are using rsyslog (or similar), you might ALSO have logs copied back out of journald into traditional logfiles [07:23] twb: systemctl cat seriapp.target doesn't return anything [07:23] journalctl -> No journal files were found. [07:24] That is interesting. [07:24] Hi. Is there software to automaticaly restart script that exit. I know you can use bash for this, but is there some package for that? [07:24] nocontrol: what environment is this? e.g. a regular Ubuntu 14.04 install, or what [07:24] tiblock: there are many ways to achieve that, which to use depends on your specific case [07:24] twb: tbh it's a raspbian [07:25] nocontrol: for raspbian, this is the wrong place [07:25] !raspbian [07:25] twb, i just need to start binary file again when he crashes. [07:25] twb: does it have any difference ? [07:25] On #debian it says: 18:25 Raspbian is a distribution made specifically for the . Raspbian is not Debian and it is not supported in #debian. Please use #raspbian on irc.freenode.net for support. http://www.raspbian.org/ [07:26] nocontrol: I don't actually know! That makes it hard to support :-) [07:26] tiblock: if it is a daemon, you can tell upstart or systemd to restart it [07:26] tiblock: if it is something like rtorrent, then the best way is different [07:27] nocontrol: you could also try asking the foreman or ruby communities, or #systemd. [07:27] twb, its not daemon, its handmade thing. I mean if i do "./prog" then when it chashes nothing happens. I want "relaunch ./prog" so it will restart on crash. Is there such "relaunch" software? [07:28] tiblock: hrm... I don't know of a good one [07:28] tiblock: here is one I wrote years ago: http://cyber.com.au/~twb/.bin/twb-loop [07:28] twb, okay, will use .sh scripts. Thank you [07:28] The simplest is: while ! ./prog; do sleep 1; done [07:29] But that does not do exponential backoff, logging &c [07:30] tiblock: you can ask #bash for help with writing scripts like this, but be aware they can be grumpy and rude :-) === Algorithm is now known as ^King === jelly-home is now known as jelly [08:33] Good morning. === Lcawte|Away is now known as Lcawte [11:15] Anybody here who can help me with an autofs/lxc issue: http://askubuntu.com/questions/729950/automouting-nfs-share-inside-natd-lxc-container-fails ? [13:12] hi folks [13:19] http://pastebin.com/UEyE83Ak [13:25] sb_9: iptables and ip6tables need to be configured individually [13:25] sb_9: so for permitted v6 addresses, you have to add ACCEPT rules for those sources going to that port === Lcawte is now known as Lcawte|Away [13:26] sb_9: i.e. ip6tables -I INPUT 1 -s ipv6-address -p tcp --dport 2070 -j ACCEPT [13:26] it looks like you already have one such rule in place though [13:26] teward: configured ipv6 tables to allow the known ipv6 address. But DROP Rule from Ipv4 fireall configuration is rejecting the connection. [13:27] sb_9: no, because v6 goes through ip6tables rules [13:27] NOT iptables [13:27] trust me, assuming iptables handled both v4 and v6 is how I got breached through the v6 IP on one of my VPSes [13:27] sb_9: are you getting timeouts or just "connection refused"? [13:28] connection refused could mean it's not listening on v6 [13:28] teward: it allowed to make a connection when i removed the DROP rule from iptables. [13:28] sb_9: that's likely because you're not using v6 as expected [13:28] sb_9: if the iptables rule is blocking, then you're going over v4, apparently [13:28] teward: did i configured ipv6 wrongly. [13:28] ip6tables is the one that would be handling v6 [13:29] sb_9: i can't say, i'm not at your location to do diagnostics - you have to make absolutely certain that the computer initiating the connection to the system is in fact using IPv6 [13:29] for that connection attempt [13:29] teward: i have configured ipv6, but it is not working as i expected. [13:30] sb_9: configured on the server, or the system you're using to reach to port 2070 on the target server? [13:31] teward: configured on the server. [13:31] i'm not talking about the server [13:31] i'm talking about the client you're on [13:31] I.E. [13:31] which system is connecting to the server at port 2070 [13:32] sb_9: telnet -6 SERVERIPV6ADDRESS 2070 [13:32] after putting that iptables rule back in place [13:32] if it works, then you know it's not an issue with the v6 on the server [13:33] (yes it may not be telnet, it's just one way to see if there's a connection that can get through) [13:33] teward: then what should be the issue. do you mean client system is trying with ipv4 connection? [13:33] yes that's exactly what i'm saying [13:33] sb_9: if the client system doesn't have proper v6 it may be falling back to v4 [13:34] teward: i am testing with telnet only. but didn't tried option -6 . [13:34] in which case ip6tables wouldn't see it and iptables would [13:34] sb_9: use -6 - it forces IPv6 [13:34] sb_9: the client system you're using - does it have proper working v6? [13:35] teward: since my client is trying to connect it. i just asked to send him the ip address shown by Google. "my ip address". [13:36] make sure it's showing an ipv6 address [13:36] and if it is, then use `telnet -6` to *force* Ipv6 [13:36] if it's still not working, then it could be a client-side configuratin [13:37] but if you don't *know* that you have properly-working IPv6, then that may be the issue [13:37] teward: it is showing ipv6 address only. [13:37] sb_9: with -6 does it try and connect? [13:37] (you never answered that heh) [13:38] teward: not yet. need to do. But if it fails? [13:38] then remove the ip6tables DROP rule and try again [13:38] teward: to test that I need to connect my client system [13:38] if it still fails, it may be your client system that's not doing v6 right, through whatever connection is in place [13:39] teward: okay. let me try it. [13:39] teward: thanks for your suggestion. [13:42] whee, power outage >.> [14:09] I have a server here at my lab which keeps getting zombie processes. [14:10] How can a process become a zombie process and not be collected by init? [14:10] Can the application itself cause that? === my-name is now known as kulyzu === catalase is now known as puppymonkeybaby [15:41] alright, i have a huge issue with postfix and virtualmin. I've got a virtual and canonical mapping going, and I want to restrict sending email to ONLY be allowed from addresses in virtual/canonical db [15:41] currently i've used smtpd_sender_restrictions = reject_unauth_pipelining permit_mynetworks permit_sasl_authenticated reject_unauth_destination reject_unlisted_sender reject_rbl_client cbl.abuseat.org reject_rbl_client zen.spamhaus.org reject_rbl_client bl.spamcop.net [15:41] i know a bit about how postfix works, but this one has me tied [15:41] i've read the ubuntu docs, and the postfix manuals including postconf(5) [15:45] necrophcodr: you could probably add a check_sender_access (http://www.postfix.org/postconf.5.html#check_sender_access) to your smtpd_sender_restrictions === Lcawte|Away is now known as Lcawte [17:10] alright, so having a canonical and virtual db setup with postfix, how do you use that to reject emails being sent from the server, not listed in those db? [17:11] it can be as restrictive as possible or not, as long as only those in canonical/virtual maps are allowed to send [17:14] necrophcodr: make the DB lookup via the check_sender_access restriction [17:15] sdeziel: is that really absolutely the only way? i mean if it is, it's what i'm gonna have to do, although that requires a different db [17:19] necrophcodr: maybe this would be better suited: http://www.postfix.org/postconf.5.html#reject_unlisted_sender [17:22] sdeziel: the following is what i've been using [17:22] smtpd_sender_restrictions = reject_unlisted_sender, reject_non_fqdn_sender, reject_unknown_sender_domain, reject_unauth_pipelining, permit_mynetworks, permit_sasl_authenticated, reject_unauth_destination, reject [17:22] it wasn't enough [17:28] necrophcodr: odd. Maybe you could try debugging this with debug_peer_list. I'd also recommend checking postfix own IRC channel (if any) and/or the mailing list as I'm not familiar with postfix DB integration [17:29] sdeziel: thanks, i'll give both a shot. [18:01] when i run netstat i saw some PID/program-name has a -, even if it's run with sudo, why? [18:02] !crosspost | kuly-zu [18:02] kuly-zu: Please don't ask the same question in multiple Ubuntu channels at the same time. Many helpers are in more than one channel and it's not fair to them or the other people seeking support. [18:04] kuly-zu: connections opened by the kernel shows has having a prog name of "-" [18:13] hey guys - is there a good guide for the tuning of a server for high network activity. We have a micro http service that is getting around 150k req/sec then makes at least one db request per to our db cluster over a second nic. [18:14] what is your current network through put [18:14] how big is each request [18:14] what does the processing of that request [18:15] looks like 38MiB RX, 31MiB tx on the http bound interface, then 15MiB rx and 26MiB txt on the DB facing interface [18:15] ikonia, the http requests are simple get requests [18:15] yes but what do they do [18:15] processed by a netty app [18:15] whats serving them [18:16] what is load thats creating [18:16] how many requests are those stats for [18:16] load right now is about 33% of capacity [18:16] 150 req/sec is the stats above [18:16] what makes you think it won't scale up ? [18:16] just trying to look into tuning the network side same point we're tuning our app more [18:17] I think you'd need to look at a problem to "fix" rather than "generic" tuning [18:17] as what you change for A impacts B [18:18] true makes sense [18:19] i guess one question would be what to look for that points to network issues vs app level [18:20] it will really depend on the behaviour [18:20] depending on what's happening you'll get a few tell tail signs [18:20] but it depends on the situation, === Luke_ is now known as Luke