[06:25] <JayPi> Hi everyone! I would like to build a public ftp server (or http) that can manage a queue-like system. By example, if I give the link www.my-server.com/test.zip, I would like to create a *queue* in case of many users trying to download it at the same time... So, the first one got the download directly, the second has to wait, etc.
[06:29] <JayPi> No idea, anyone?
[06:32] <sarnold> JayPi: that's not how ftp works; you can delay a bit but sooner or later the ftp client is going to time out and fail. either then they re-try or they give up.
[06:33] <JayPi> sarnold, ok. but is there a way to have like a *waiting line* for a download http request?
[06:33] <sarnold> JayPi: better is to reject their connection outright if someone else is already connected. that way their client can just pound away reconnecting like a monkey until eventually you let them in.
[06:34] <sarnold> JayPi: there might be bandwidth shaping modules for e.g. apache that let you say you handle at most three connections...
[06:34] <JayPi> I would like to create a webpage with their download informations like : you are 2nd in waiting line. Your approx. waiting time is : 2 min 45 s.
[06:35] <sarnold> ah sure you can dothat
[06:36] <sarnold> if you don't mind having people land on a landing page, and only give them access to a long randomly generated URL after a while, that'd be easy enough
[06:36] <sarnold> probably an afternoon's work for someone who knows python / ruby..
[06:36] <JayPi> If I have 45 requests, I would like my server to give them a number from 1 to 45 (their priority)... How can I achieve that?
[06:36] <JayPi> and if I don't know python or ruby lol ?
[06:37] <sarnold> well, I guess you could use erlang or perl ... probably not as easy to pull off though
[06:38] <JayPi> I don't know any programming language
[06:38] <sarnold> this would be my starting point if I wanted to do it: http://www.sinatrarb.com/
[06:38] <sarnold> this will be a fun project to learn :D
[06:38]  * Abhijit smells money
[06:40] <Abhijit> :-(
[09:41] <foo357> Hello, I'm setting up apache on a ubuntu 14.04 machine, I've installed apache from the repository. When I try to browse to the server I get no connection. "telnet <ip> 80" on the server results in a connection, but not if I try from another machine
[09:42] <obi12341> maybe you have a firewall installed?
[09:42] <obi12341> try: ufw status
[09:42] <foo357> On the machine, hmm could be possible.
[09:43] <obi12341> and look in your iptables
[09:43] <foo357> obi12341: ufw is inactive
[09:43] <obi12341> iptables --list
[09:44] <obi12341> and maybe your server has a hardware firewall
[09:45] <foo357> obi12341: iptables looks empty to me. http://pastebin.com/9KfzdrXi
[09:46] <obi12341> yes so no software firewall
[09:46] <obi12341> what is the ip of your server?
[09:46] <foo357> obi12341: sorry, I can't tell you the exact address
[09:47] <obi12341> okay, then you have to look on your own. looks like you have a hardware firewall
[09:47] <foo357> obi12341: but if I can do a telnet from within the server (specifying the ip), I should reasonably be able to do that from any other machine as well?
[09:48] <obi12341> no the firewall can detect that the connection is from the "inside" so he allows the packages
[09:49] <obi12341> the firewall sees the source IP
[09:49] <foo357> obi12341: of course, what I meant was that the server itself shouldn't stop it.
[09:50] <obi12341> your telenet is successful isn't it?
[09:50] <foo357> obi12341: yes, but only from the server, the local machine.
[09:50] <obi12341> yes so you do: telnet 127.0.0.1 80
[09:50] <obi12341> and telnet <ip> 80
[09:51] <foo357> obi12341: no I've always tried with the eth0 interface
[09:51] <obi12341> the firewall recognize the source ip (your server ip) and allows these packages
[09:51] <obi12341> yes so telnet <ip> 80
[09:53] <foo357> obi12341: ok, thanks for your help. I guess there's no configuration on the server that should cause this problem then.
[09:54] <obi12341> just take a deeper look in your contract with your hoster and you will see you have a firewall in front of your server
[09:54] <obi12341> i guess no configuration issue
[09:54] <foo357> obi12341: thanks
[09:54] <obi12341> no problem
[09:57] <YamakasY> I forgot... why can I remote login without password with username@hostname but not as username@fqdn
[11:09] <pmatulis> morning
[11:10] <pmatulis> YamakasY: check ~/.ssh/config
[12:31] <YamakasY> pmatulis: already fixed! thanks
[12:44] <pmatulis> YamakasY: keep on truckin'
[12:55] <YamakasY> pmatulis: broaaaammmm!
[12:56] <pmatulis> :)
[13:02] <ddsf> Are extra packages required to do iptables MASQUERADE on -server?
[13:20] <rbasak> ddsf: I don't think so. Though IIRC masquerade is intended for dialup (dynamic IPs). Usually on a server SNAT is used instead.
[13:27] <patdk-wk> masquerade is snat :)
[13:29] <jamespage> coreycb, did arges reject nova? don't want to put two in the queue
[13:30] <coreycb> jamespage, hmm, not sure
[13:30]  * coreycb checks
[13:32] <jamespage> coreycb, https://launchpad.net/ubuntu/trusty/+queue?queue_state=1&queue_text=
[13:32] <jamespage> no
[13:33] <coreycb> jamespage, ok
[13:35] <arges> jamespage: i can reject the older one if necessary
[13:35] <jamespage> arges, please
[13:35] <jamespage> arges, just a tweak to the changelog
[13:35] <arges> jamespage: i'll look at those soon
[13:35] <jamespage> arges, thanks!
[13:42] <zagaza> hi, what's a healthy memory usage percentage for a ubuntu vps?
[13:42] <kklimonda> that's too general of a question - it all depends on what you are planning on running
[13:43] <zagaza> just 4-5 wordpress sites
[13:44] <kklimonda> again, it depends on the workload - misconfigured wordpress is known to use excessive resources
[13:44] <kklimonda> I'd probably start with the 512MB VPS and see if it's enough
[13:44] <kklimonda> if not, you can always upgrade
[13:45] <zagaza> I use a 512mb one, it uses up 40-50% mem
[13:46] <kklimonda> well, memory is there to be used :)
[13:46] <zagaza> I guess my question is: is there any upper breakpoint where it gets critical?
[13:46] <zagaza> generally
[13:47] <kklimonda> well, at some point your sites are going to take too long to respond to queries, then it's critical :)
[13:47] <kklimonda> you can monitor system load
[13:47] <kklimonda> in general, once your server starts swapping (if you have swap enabled) it's a good sign you should start thinking about upgrading ram
[13:48] <zagaza> I'm a hardcase newbie on vps'es and servers, so I don't even know what swapping is :) but gonna read upon it
[13:49] <kklimonda> yeah, it's a good idea to read about that stuff if you want to manage a server.
[13:49] <kklimonda> also, if you just need a place to host a couple of wordpress sites, maybe you should take a look at some hosted solutions
[13:50] <zagaza> well I know some basics, I managed to set up a wordpress site on it, mysql, phpmyadmin, nginx etc.. so Im not completely out of it :) just need to get a broader grip of it
[13:50] <kklimonda> mhm
[13:51] <zagaza> I ad a hosted solution for wp, but it was more expensive than this digitalocean vps I grapped so I thought it was pretty much pointless
[13:51] <zagaza> although this is way more work
[13:51] <zagaza> had*
[13:52] <kklimonda> yeah, that's why a hosted solution is more expensive - when you have to managed stuff yourself, it takes time - especially when you are still learning.
[13:54] <zagaza> yep yep, nonetheless its healthy to learn this stuff if one is into webdev, that's mainly why I picked up a vps
[13:54] <zagaza> but I like to learn by doing, that's why I know some more difficult stuff, but I might have missed out on some of the basics lol
[13:56] <arges> jamespage: ok nova rejected, upload the new one when ready ( also you can upload multiple versions and I can reject them after the fact too)
[13:56] <jamespage> arges, will do
[13:57] <kklimonda> zagaza: well, yeah - knowing more about the stack you are running is always beneficial.
[13:57] <jamespage> arges, new version uploaded
[13:58] <jamespage> I think that covers neutron, nova and cinder revisions
[13:58] <arges> jamespage: ok i'll go through those now
[13:58] <jamespage> ta
[14:24] <funcoland1> does anyone know how i can do a dist-upgrade on 14.04 to 14.04.1 without the internet?
[14:26] <cfhowlett> funcoland1,   http://jeffhoogland.blogspot.com/2010/06/howto-installing-ubuntu-packages.html          or  techspalace.blogspot.com/2009/04/offline-update-ubuntu.html
[14:26] <arges> jamespage: fyi, for nova it would be good to have a line in the changelog saying "dropped Fixed-rbd-backend-* patch", but probably not a big deal now
[14:27] <funcoland1> thanks cfhowlett
[14:27] <cfhowlett> funcoland1, happy2help
[14:27] <funcoland1> is it as easy as just grabbing all of the new packages and just running an apt-get update then?
[14:28] <funcoland1> the thing is, i've grabbed packages off of a 14.04.1 disc and then tried to run an update... what ends up happening is the number when i first log in at the top does not change to 14.04.1
[14:28] <funcoland1> it still says "Ubuntu 14.04 LTS"
[14:28] <cfhowlett> funcoland1, I used it years ago --- it worked.  Don't remember details
[14:29] <funcoland1> alright
[14:30] <cfhowlett> funcoland1, that won't work.  the alternate install (no longer made) allowed one to use the Ubuntu CD as a software source as you desribe.  not any more
[14:31] <funcoland1> ahhhh, is that written up anywhere on the ubuntu site? I'd like to read more into that
[14:31] <cfhowlett> funcoland1, as I said, the alternate CD method no longer works, no longer available
[14:34] <funcoland1> alright, thanks sir
[14:39] <arges> jamespage: ok got them all in proposed. I see heat listed as a task, but I don't see an upload for that btw
[16:07] <linuxgeek_> on a 14.04 server, i want to configure vlan on em1 device.
[16:08] <linuxgeek_> i have installed the vlan package
[16:11] <linuxgeek_> http://paste.ubuntu.com/8037328/ is the interfaces file
[16:12] <linuxgeek_> the issue is the kvm host [14.04] server does not get a dhcp ip
[16:12] <linuxgeek_> if i remove the vlan config in the interfaces file and use just em1, it gets the dhcp ip
[18:10] <tgm4883> I have a 12.04 server with LDAP, MySQL and Jasig CAS installed and I'm running into an intermittant issue where CAS will stop working and throwing errors indicating that it cannot reach mysql anymore. Attempting to connect to Mysql or ldap during this time also fails until mysql or ldap is restarted. Looking at netstat, It doesn't appear that I'm hitting any
[18:10] <tgm4883> connection limits, and during this time load on the machine is pretty low. Any ideas where to keep looking?
[18:11] <sarnold> "fails" why? connection reset? timeout? ..
[18:11] <tgm4883> timeout
[18:12] <tgm4883> Yesterday this happened 3 times, it hasn't happened today though
[18:12] <sarnold> how long is the timeout? is it reasonable?
[18:13] <tgm4883> it seems to timeout after 10 or so seconds. I'm a bit limited in the troubleshooting I can do on this since it's a production server and needs to be back up :/
[18:28] <RoyK> how does update-initramfs -u know which files to put in the initrd.img-something file?
[19:55] <michele> hi there. I run do-release-upgrade on an ubuntu server. The upgrade failed with grub, with the following error: http://pastebin.com/NeShL1sw - I did not reboot, so I am still inside the machine. How do I solve?
[20:00] <rberg> huh /boot is a LV? you probably want to install grub to the underlying device
[20:12] <TJ-> michele: The issue is that GRUB2 has problems with fake-raid controllers, and in your case it looks like an SIL fake-raid controller disk's partition #1 has the /boot/ file-system on it. Depending on how the SIL controller presents the underlying device to the system via BIOS calls, GRUB may be able to access partition #1 on the inderlying device. It depends on whether the SIL controller puts its RAID meta-data at the beginning of the drive, or not
[20:14] <TJ-> michele: "grub-install --disk-module=biosdisk ..." *may* be a workaround
[20:16] <TJ-> michele: Otherwise, if the underlying device doesn't have RAID meta-data at the beginning (meaning that the underlying device and the Device Mapper device have identical sector 0's, you may be able to do "grub-install /dev/sdX" where X is the underlying device letter
[20:18] <michele> TJ-: thanks for the detailed answer. I will try with module=biosdisk, but what should I put after?
[20:19] <michele> root@zulu:/boot/grub# grub-install --disk-module=biosdisk
[20:19] <michele> grub-install: --disk-module: (PROGRAM ERROR) Option should have been recognized!?
[20:19] <michele> Try 'grub-install --help' or 'grub-install --usage' for more information.
[20:19] <TJ-> michele: the device to install grub to, so with no other changes it'd be "grub-install --disk-module=bios /dev/mapper/sil_bhbiaicbfaah1" but I doubt the BIOS option will work
[20:19] <TJ-> michele: oops, typo! "grub-install --disk-module=biosdisk /dev/mapper/sil_bhbiaicbfaah1"
[20:20] <michele> same as above
[20:20] <michele> root@zulu:~# grub-install --disk-module=biosdisk /dev/mapper/sil_bhbiaicbfaah1
[20:20] <michele> grub-install: --disk-module: (PROGRAM ERROR) Option should have been recognized!?
[20:21] <TJ-> michele: OK, so that option isn't viable. Move on to the 2nd suggestion.
[20:21] <michele> root@zulu:~# grub-install /dev/sda
[20:21] <michele> Installing for i386-pc platform.
[20:21] <michele> grub-install: error: cannot find a GRUB drive for /dev/mapper/sil_bhbiaicbfaah1.  Check your device.map.
[20:22] <TJ-> michele: You'll need to identify if the kernel can see the raw device underlying /dev/mapper/sil_bhbiaicbfaah
[20:22] <TJ-> michele: Then you'll need to compare sector 0 on each to discover if they are identical.
[20:22] <TJ-> michele: If so, you can "grub-install <underlying-device>"
[20:50] <RoyK> how does update-initramfs -u know which files to put in the initrd.img-something file?
[20:51] <RoyK> linuxgeek_: I guess the dhcp server isn't on that VLAN...
[21:02] <TJ-> RoyK: it calls hook scripts installed by packages that require an initrd.img presence
[21:12] <RoyK> TJ-: where can I read more about that?
[21:55] <StolenToast> whenever i log into my ubuntu server now and it asks me to log in I want to type in groot