[00:41] <wilee-nilee> Not sure if this is appropriate channel, however I'm a fairly experienced ubuntu user except in this area.  So I have a smart tv, I have minidlna working with home and, video, but can't get it to read an external usb HD, not sure the pathway. I
[02:33] <crass> weird, I have been able to use resize2fs on a live filesystem in the past
[02:33] <crass> but now it appears to be requiring that the block device is not in use...
[02:34] <crass> any ideas why now resize2fs fails when opening the device because its in use?
[05:44] <gartral> alright all, i've been fighting this all night, even with ddrescue I can't get the superblock of my booting partition from my server to transfer to a new drive, any advice?
[06:15] <sarnold> gartral: (just a quick note before I go to bed...) if you really can't get the superblock there tend to be superblock backups made around 8192 or 8193 or something, there's a whole string of superblock backups. Probably you can't get your master boot record or partition boot record to copy over instead, but those might be easier to create anew anyhow. Hope this helps :) have fun
[06:47] <darkXplo_> hi guysss
[06:48] <darkXplo_> can anyone help me with UPW
[06:49] <darkXplo_> https://gist.github.com/anonymous/39e070dbc70c45c48502
[06:50] <darkXplo_> UFW
[07:48] <Jinxed-> Anyone know of a video conferencing server for a private network that would allow android mobile devices to connect
[08:01] <gartral> sarnold: yea, none were working
[08:10] <gartral> Jinxed-: asterisk should allow that
[09:54] <chmac> Added eth1 to /etc/network/interfaces with a static ip for a private subnet (10.0.0.0/24), but the route for that subnet is not being created...
[09:54] <chmac> Do I need to manually add the route somewhere?
[09:54] <chmac> I think on my other machine it "just works"...
[09:56] <Svetlana> http://multipath-tcp.org/pmwiki.php/Users/ConfigureRouting has some details about how this is done. Is it how you do it on both machines?
[09:58] <Svetlana> Looking at http://manpages.ubuntu.com/manpages/precise/man5/interfaces.5.html I think you can add a command into interfaces file directly... like 'IFACE OPTIONS' section says. Is this how you do it?
[10:00] <chmac> Svetlana: http://pastie.org/8381455
[10:00] <chmac> Svetlana: Yeah, I'm reading here and there that I could add a `post up route add foo`
[10:01] <chmac> But that seems like a hack, and I'm not sure why it's working on one machine but not another
[10:01] <chmac> These are both ubuntu-server instances running inside virtualbox
[10:02] <Svetlana> Is this how your interfaces file looks under both machines? Does any of them have different files in /etc/network/if-up.d/?
[10:03] <Svetlana> (I gave you two links. https://wiki.merproject.org/wiki/Nemo/USBNetworking is an example for the second one: you would have to replace the stuff after 'up' and 'down' with 'route bla bla' commands.)
[10:03] <chmac> Svetlana: The machine that works has a postfix file in there, let me see if there's anything useful in it...
[10:04] <chmac> Otherwise, the md5s match on all the files on both machines
[10:04] <chmac> Svetlana Nothing about routing in the postfix file
[10:04] <Svetlana> And while the route is not bring created, is the connection activated at all? 'sudo ifconfig' shows 'inet addr' line for the active ones.
[10:04] <chmac> I think I could solve it with the `post up route add` option, but it's a Sunday morning and I'm inclined to spend a little time to see if I can figure out the issue!
[10:05] <chmac> Svetlana: Hmm, let me restart the vm and see...
[10:05] <chmac> Svetlana: Yeah, the interface is up
[10:06] <Svetlana> what IP address does ifconfig show?
[10:06] <chmac> 10.1.0.2
[10:06] <Svetlana> Okay, that looks fine.
[10:06] <chmac> Shows a netmask of 255.255.255.255 though...
[10:06] <chmac> Aha, while my other machine shows 255.255.255.0
[10:07] <chmac> Crazy, typo in the interfaces file, I wrote netmasT instead of netmasK :-)
[10:07] <chmac> Svetlana: Thanks a lot for your help, I really appreciate it
[10:08] <Svetlana> http://manpages.ubuntu.com/manpages/precise/man5/interfaces.5.html in 'static method' it does not list 'network' rule, but you have it. It is a bit strange.
[10:09] <chmac> Svetlana: Hmm, good point, probably ignores it, I guess it ignored the netmasT line! :-)
[10:10] <Svetlana> When you said 'Nothing about routing in the postfix file', you meant the interfaces file, right? Or were you referring to the if-up.d/* files, which you compared?
[11:45] <eagles0513875_> hey guys i am trying to change the php upload size and for some reason even after changing the upload_max_filesize and restarting apache the filesize has not changed
[15:48] <AntelopeSalad> hey guys, i'm setting up my first server and i'm curious if init.d and upstart scripts will conflict with each other
[15:48] <AntelopeSalad> basically i've installed things like nginx through apt-get (after updating my source list) and that seems to setup an init.d script but i've been following guides that create a custom upstart config
[15:49] <AntelopeSalad> is there a way to somehow disable the init.d version so it doesn't run double the processes?
[15:50] <qman__> most packages which have upstart scripts install a dummy init.d script, verify first that there isn't an upstart script already
[15:50] <qman__> that said, if the package includes an init.d script and not an upstart script, adding your own in will conflict with the package contents
[15:51] <AntelopeSalad> qman__: what can i do to fix it? i made the upstart config afterwards
[15:51] <qman__> it's possible to manually remove/disable the init.d script and install the upstart script, but this will break every time you update the package
[15:51] <qman__> the correct way to change this is to rebuild the package using the upstart script instead
[15:51] <AntelopeSalad> what do you think is the most sane way to ensure that the process stays up?
[15:52] <qman__> the most sane way is to stick to the distro default and not touch it
[15:52] <AntelopeSalad> i didn't build anything, all i did was apt install nginx
[15:52] <AntelopeSalad> but i wasn't sure if that came with any type of process monitoring
[15:52] <qman__> it doesn't; neither does upstart
[15:53] <AntelopeSalad> but i setup process monitoring with upstart (i think)
[15:53] <qman__> if you need process monitoring, implement that separately, either with a simple cron job or to full on systems like nagios
[15:54] <AntelopeSalad> this is what i setup by following some tutorial http://pastie.org/8381932
[15:55] <AntelopeSalad> i thought upstart was the standard way to make sure stuff loads on startup and doesn't die
[15:55] <qman__> upstart provides service management the same as traditional sysvinit did; it has different features
[15:55] <qman__> it is not a monitoring system
[15:56] <AntelopeSalad> maybe i'm using the wrong termonology
[15:56] <qman__> it can be used by your monitoring system, though
[15:56] <AntelopeSalad> all i want to do is make sure if nginx dies that it gets restarted
[15:56] <AntelopeSalad> and that it starts when the system boots up
[15:56] <qman__> in a traditional init script you'd manually implement respawn on failure in the service itself
[15:56] <qman__> where upstart provides rudimentary functionality to respawn on failure
[15:57] <AntelopeSalad> so i should use the default init.d script and not touch it, and that will take care of it loading on startup?
[15:57] <qman__> yes
[15:57] <AntelopeSalad> but then use something other than upstart to make sure it doesn't die?
[15:57] <qman__> yes, if you care about monitoring, implement it
[15:58] <qman__> that can be anything from a simple cron job that checks if it's running to a full featured monitoring system like nagios
[15:58] <AntelopeSalad> if i delete the config file in the /init/ dir is that good enough to completely reverse the mistake i made?
[15:58] <qman__> if that's all you created, then yes
[15:58] <AntelopeSalad> (and reboot the machine of course)
[15:58] <qman__> no reboot necessary
[16:00] <qman__> while converting to upstart is a good thing long term, sticking to the system defaults wherever possible is the most sane way to manage a system, because then updates won't break things
[16:00] <qman__> nginx will more than likely be converted to upstart in a future release if it hasn't already
[16:00] <AntelopeSalad> ok thanks, so it's really up to the tool to handle supplying the startup scripts
[16:00] <qman__> correct; you can do it yourself but there's a lot to take into consideration
[16:01] <qman__> and it involves building your own packages
[16:01] <AntelopeSalad> that sounds like the opposite of what i want to do haha
[16:01] <qman__> if you want set and forget, sticking to defaults is best
[16:01] <AntelopeSalad> btw is it standard to run "service nginx start" or should i be using sudo?
[16:02] <qman__> you generally need root privileges to run the service command
[16:02] <AntelopeSalad> ok, i actually didn't try it without it but i wasn't sure if running a process as root was a good idea
[16:02] <qman__> it doesn't run nginx as root
[16:03] <qman__> or at least it shouldn't, that would be a very bad bug
[16:03] <qman__> generally the init script handles deescalation of privileges
[16:03] <AntelopeSalad> i think it's running it as root
[16:04] <AntelopeSalad> http://pastie.org/8381963
[16:04] <qman__> ah, I see
[16:05] <qman__> only the main binary is using root, the subprocesses are running unprivileged
[16:05] <AntelopeSalad> i'm not sure though
[16:05] <qman__> which looks to be by design
[16:05] <AntelopeSalad> yeah
[16:05] <qman__> I'm not intimately familiar with nginx
[16:05] <AntelopeSalad> the worker is what gets touched by the public maybe?
[16:05] <qman__> but that's a sane setup
[16:05] <qman__> right
[16:05] <AntelopeSalad> because in the nginx conf i had to supply the # of workers
[16:05] <AntelopeSalad> i chose 1 in this case
[16:06] <qman__> apache does the same thing
[16:06] <qman__> there's one root process and everything else is unprivileged
[16:07] <AntelopeSalad> ok, thanks for the advice, i'll continue to install everything as i did before and look into nagios for making sure they don't explode
[16:07] <AntelopeSalad> (minus the upstart confs)
[16:07] <qman__> http://pastie.org/8381969
[16:07] <qman__> there's lots of ways to do it, nagios is just one popular one
[16:08] <AntelopeSalad> is it lightweight?
[16:08] <qman__> not really, it's made to be comprehensive
[16:08] <AntelopeSalad> i'm running quite a bit on a micro instance, coming really close to the ram limitation
[16:08] <qman__> well there's two main ways to run nagios
[16:08] <qman__> you can run with a client, or clientless
[16:09] <qman__> clientless, your nagios server just makes connection attempts to see if the service is up
[16:09] <qman__> you can http request and based on the result, either say it's up, or failed
[16:09] <AntelopeSalad> oh, so i can't even run nagois on the same server as the deployed web app?
[16:09] <qman__> right
[16:09] <AntelopeSalad> yeah that's definitely not going to work in this case
[16:10] <qman__> well, you can, but it's not the best design setup
[16:10] <qman__> if you just want local monitoring, you'd be better off with a simple shell script
[16:10] <AntelopeSalad> atm i'm dealing with 1 instance
[16:10] <AntelopeSalad> maybe i should look into something tied into the language the app is written in?
[16:12] <qman__> this is about the most basic script: http://pastie.org/8381973
[16:12] <qman__> you'd then set that script up to run as root in the crontab every X minutes depending on your need
[16:13] <AntelopeSalad> oh
[16:13] <qman__> you can get more complicated by checking pids or logs
[16:14] <qman__> or add a line to email you when it's down
[16:14] <qman__> whatever you need
[16:14] <AntelopeSalad> would the pid be in /var/run/nginx.pid?
[16:15] <AntelopeSalad> i'm comparing your script to that upstart conf and it seems similar, except the upstart one handles the workers
[16:15] <AntelopeSalad> i'm not sure what happens if nginx crashes, it might be possible for it to be alive but with no workers in which case that's bad
[16:15] <qman__> probably, I don't have any nginx servers to check
[16:15] <qman__> well, one thing you can do
[16:16] <qman__> http://pastie.org/8381986
[16:16] <qman__> er, want that to say restart
[16:17] <AntelopeSalad> is restart smart enough to also start if it can't find something to restart?
[16:18] <qman__> should be, you can test it
[16:18] <qman__> kill nginx without stopping the service, then try
[16:19] <qman__> if it isn't, you could stop then start
[16:19] <qman__> then if stop failed it would continue to start
[16:19] <AntelopeSalad> it seems to be ok, i killed the master manually
[16:19] <AntelopeSalad> and it restarted
[16:19] <AntelopeSalad> yep it works, i killed both the master+worker manually and restart still worked
[18:13] <AntelopeSalad> a sources.list.d file is saying to use "squeeze" for an ubuntu code name for a certain tool i'm trying to install, they don't seem to have one for raring or precise -- is this going to be bad if i use squeeze?
[18:13] <AntelopeSalad> i'm using ubuntu 13.x
[18:17] <genii> AntelopeSalad: What exactly are you trying to install?
[18:17] <AntelopeSalad> genii: redis, and after some googling it seems like the dotdeb.org list is my best hope for a recent version
[18:18] <TJ-> AntelopeSalad: The biggest challenge can be what the package depends on (other packages and their versions) because they may not be correct in Ubuntu. The other issue is slight difference in  package config installation practices / locations between Debian and Ubuntu
[18:19] <AntelopeSalad> i decided to take a risk and just tried it using "stable" as the code and it seems to have installed correctly, do you think i'm in the clear if it happened to complete without errors?
[18:20] <TJ-> AntelopeSalad: Well you've managed to avoid the dependency issues so just monitor it, its log-files, and the system log-files until you're satisfied
[18:20] <AntelopeSalad> ok thanks
[18:20] <genii> There seems to be a popular PPA for it which is maintained
[18:20] <AntelopeSalad> genii: which?
[18:21] <genii> AntelopeSalad: https://launchpad.net/~rwky/+archive/redis
[18:21] <genii> Last build was 5 weeks ago which is pretty recent
[18:21] <AntelopeSalad> ah yes, i saw this earlier but it didn't contain the steps on how to get it into the sources list
[18:23] <genii> AntelopeSalad: sudo add-apt-repository ppa:rwky/redis     ..in this case. DISCLAIMER: USE PPAs AT YOUR OWN RISK
[18:24] <AntelopeSalad> yeah i got hung up on doing the key step
[18:24] <AntelopeSalad> most other tools i've used had a url that had the key
[18:24] <AntelopeSalad> this one links to the key but it's not something i can just wget
[18:24] <genii> But in this scenario, the PPA is far less dangerous than adding Debian repositories
[18:25] <AntelopeSalad> oh, so this ppa method bypasses having to do the sources.list steps?
[18:25] <genii> AntelopeSalad: It makes an entry in /etc/apt/sources.list.d/ directory with the PPA name
[18:26] <AntelopeSalad> if i delete that file in the sources.list.d dir does that completely wipe it out after i run apt-get update?
[18:26] <genii> AntelopeSalad: This way no foreign entries make it into /etc/apt/sources.list main file and you can always later do ppa-purge command if it was problemmatic
[18:26] <AntelopeSalad> because if this launchpad version works i'd like to delete the dotdev one
[18:27] <genii> AntelopeSalad: No, just deleting the file there doesn't remove the app if it was already installed. You need to do the ppa-purge as described. Alternately with ppa name, like: sudo ppa-purge <name>
[18:31] <AntelopeSalad> ok
[18:31] <AntelopeSalad> it seems to have worked using the launchpad's ppa, thanks
[18:32] <AntelopeSalad> that version it has is the latest stable one too
[18:35] <genii> AntelopeSalad: Just remember that if you experience some issue with it, you probably won't find support here and will need to talk to the PPA maintainer or to the redis people instead :)
[18:36] <AntelopeSalad> genii: yeah
[18:37] <AntelopeSalad> also 1 one last thing if you don't mind...
[18:37] <AntelopeSalad> if i do a dpkg -i foobar.deb , this still gives me an init.d script right?
[18:38] <genii> AntelopeSalad: If the package installs one
[18:38] <AntelopeSalad> ok, so that's up to the package -- gotcha
[18:39] <AntelopeSalad> thanks again, in this i guess i got lucky and it did (elasticsearch)
[19:59] <zexcriz> what is the difference between hosting our data on our VPS and in the other in the cloud ? i doubt there is any difference
[20:05] <bekks> zexcriz: How important is that data?
[20:06] <zexcriz> bekks, currently hosting a VPS and now i was curious what are advantages or disadvantages if any if go for cloud
[20:06] <zexcriz> so just need some advice, what to go for as i am totaly confused over here.
[20:06] <bekks> I'd not use a cloud for anything that might be important.
[20:07] <soren> Uh..
[20:07] <soren> So you'd not put anything important in the cloud, but a VPS is fine?
[20:07] <zexcriz> bekks, data security in cloud is a prob ?
[20:07] <bekks> soren: I am not using a VPS either.
[20:07] <soren> bekks: Then your argument is nonsense.
[20:08] <bekks> soren: If you think so: please take over the conversation.
[20:10] <zexcriz> so thinking about efficiency and security which is the best option ?
[20:12] <pmatulis_> bekks: why do you say the cloud is not good for important data?
[20:13] <bekks> pmatulis_: Because there is no guarantee whatsoever that you actually can access your data when you need it. There is somethin like 99.9xy% -- but thats lower than 100%.
[20:13] <soren> Same for anything else.
[20:13] <soren> Even stuff you control yourself.
[20:14] <soren> But certainly for any other type of storage managed by someone else, whether it's called "cloud" or just "VPS" or whatnot.
[20:16] <soren> zexcriz: "Cloud" basically means that you can control "it" with an API. It also typically means that you're charged in very small increments and only for exactly how much you're actually consuming.
[20:16] <soren> zexcriz: In terms of security for instance, cloud and VPS is the same thing.
[20:18] <zexcriz> soren, getting your point.
[20:18] <soren> zexcriz: Being able to control things with an API means that you can very easily acquire (and get rid of) resources. If your resource requirements are almost entirely static, you might as well stay with what you've got.
[20:19] <zexcriz> soren, ok :)
[20:19] <soren> If, on the other hand, your resource demand varies (over the course of a day, or over the course of a week or over the course of a year, for instance) and you applications scale well horizontally, "cloud" might be a good option for you.
[20:19]  * soren has to run
[20:20] <pmatulis_> good intro to cloud ^^^
[20:21] <zexcriz> yeah that was really a good intro..!! :) thanks soren :)
[20:23] <pmatulis_> a main point is trust.  you implicitly place your trust in a cloud provider.  as opposed to building your own (private) cloud
[20:24] <zexcriz> imp point.
[20:26] <ersi> Magic eightball says: Outlook cloudy.