|
Linux on the desktop: well yeah
|
# ? Jun 25, 2016 07:14 |
|
|
# ? Apr 29, 2024 12:11 |
|
uhh it's totally giving you a valid explanation: *barfs up some halfassed libcurl error message like a cat on the floor*
|
# ? Jun 25, 2016 10:55 |
|
ahmeni posted:uhh it's totally giving you a valid explanation: linux.txt
|
# ? Jun 25, 2016 12:04 |
It's the unix philosophy, we don't do file download because curl does it better, so here's curl's explanation for what broke
|
|
# ? Jun 25, 2016 15:52 |
|
that is not an error message from libcurl and was in fact hand-written by the rpm/dnf team. they do not know how to write error messages.
|
# ? Jun 25, 2016 16:20 |
|
Apocadall posted:i am having such a bitch of a time getting a yealink voip phone talking through a linux openvpn server to an asterisk pbx Sngrep. You're welcome.
|
# ? Jun 25, 2016 16:40 |
|
Apocadall posted:i am having such a bitch of a time getting a yealink voip phone talking through a linux openvpn server to an asterisk pbx sounds hosed up
|
# ? Jun 25, 2016 17:15 |
|
ratbert90 posted:Sngrep. You're welcome. thanks! trying to watch two linux servers using iptraf and grepping through logs is a real ballache
|
# ? Jun 25, 2016 17:49 |
|
Apocadall posted:thanks! trying to watch two linux servers using iptraf and grepping through logs is a real ballache Yeah, that app has saved me so much time in the past. My job is a lot easier when I can tell Avaya that it's their issue even quicker!
|
# ? Jun 25, 2016 18:00 |
|
Dnf is a tool for sysadmins anyway, you should be using gnome software and Chrome should provide flatpaks for their poo poo. But that tech is all very new and of course loving Shuttleworth had to go front-run it with his own lovely garbage alternative (see also: Mir, political manouvering around the existing Upstart project) and do so in his usual dishonest fashion. (Whoosh go the goalposts etc)
|
# ? Jun 25, 2016 18:41 |
|
Just installed Fedora 24 over Fedora 23. Only kept my home and opt partitions... everything else was a fresh install. Compared with moving from ubuntu to fedora 23, this was much less painful. The only unexpected thing was adding myself to the ecryptfs group. Oh, and the desktop icons are now comically big! Gonna boot into wayland next, see if I can summon the magic smoke...
|
# ? Jun 26, 2016 08:49 |
|
cant get plex to see my drat thing. mayube im retarded. it can see it in /media, which i gavbe my dude permissions to edit etc, but uh. i mouinted my old big media drive (ntfs) and its /dev/sdb2 and i cant figure out how to get plex to see it. whatever. agg. fail
|
# ? Jun 26, 2016 08:50 |
|
im a fucktard with a small ssd and a spinny, you see.
|
# ? Jun 26, 2016 08:51 |
|
Smythe posted:im a fucktard with a small ssd and a spinny, you see. that's what I do too 64 gb ssd root drive 500 gb slow home drive I'd try to help but I know nothing about plex
|
# ? Jun 26, 2016 08:59 |
|
celeron 300a posted:Just installed Fedora 24 over Fedora 23. Only kept my home and opt partitions... everything else was a fresh install. posting from wayland now, with a geforce 210 running nouveau drivers all of my desktop icons have mysteriously disappeared other than that, it seems to work fine.
|
# ? Jun 26, 2016 09:02 |
|
celeron 300a posted:that's what I do too you might consider keeping most of /home on the ssd too -- time to read things like desktop environment config and web browser profiles can be pretty significant in time from boot to usable desktop, and opening web browsers and such i keep things like 'vmware', 'Downloads', steam library, etc. on a big spinny drive and symlink/mount things into the right place in my home directory as appropriate
|
# ? Jun 26, 2016 15:02 |
|
celeron 300a posted:that's what I do too same but 240/1000
|
# ? Jun 26, 2016 15:03 |
|
same but 2tb/36tb on nas over 40gbe some rubberized pink weight weak poo poo itt, compute big or go home
|
# ? Jun 26, 2016 17:24 |
|
Well, alright .. I guess *dejectedly collects belongings into jansport bag and leaves*
|
# ? Jun 26, 2016 19:10 |
|
i can't tell if it's linux being a shitpile, or unraid, or docker but a plex client now takes 10+ seconds to start a video when before it took 2 seconds max running the server on windows it's me, the idiot who thought switching was a good idea
|
# ? Jun 26, 2016 22:02 |
|
wait, you're running plex in docker?
|
# ? Jun 27, 2016 03:45 |
|
DID SOMEBODY SAY DOCKER??? https://www.youtube.com/watch?v=PivpCKEiQOQ&hd=1
|
# ? Jun 27, 2016 03:52 |
I love an incredibly specific Hitler parody. ugh. I don't know kubernetes from borg. my poo poo goes to an autoscaling group on AWS and that's it and that's all Coffee Jones fucked around with this message at 06:09 on Jun 27, 2016 |
|
# ? Jun 27, 2016 05:56 |
|
I haven't used docker but after hearing about it, there is something I don't get. If people are just creating these distribution independent things without virtualization, then how does it account for api changes between libc versions and kernel versions? I understand that with cgroups, you can have a process see only a specific set of processes and files, but they're sharing the same kernel at the very least. Also, wouldn't an enterprise deployment be using nfs or iscsi (for each container) to mount a working directory, instead of using the same mounted filesystem, to reduce or eliminate the chance of an errant container bringing down a whole kernel and fudging the disk?
|
# ? Jun 27, 2016 06:22 |
|
I think I made fun of people for talking about containers in the desktop linux thread or was that configiration management? either way I declare myself a hypocrite
|
# ? Jun 27, 2016 06:23 |
|
Desktop linux chat. I put ansible on my fedora 24 workstation and now I have a recipe to do most everything except install a bios grub password. Feels good.
|
# ? Jun 27, 2016 06:25 |
|
Progressive JPEG posted:wait, you're running plex in docker? unraid has an app store that installs in docker containers. it's actually nice in a lot of ways. If an image needs to be updated, just restart and pull but it's also another layer making things harder to debug when there's a problem there's always a problem.
|
# ? Jun 27, 2016 07:08 |
|
Mr Dog posted:Dnf is a tool for sysadmins anyway, you should be using gnome software I was. I think. I was using whatever graphical tool you get when as a normal person you try to install a software by clicking on a download link. it had reviews and star ratings and everything. that's why I was complaining. those error messages would have been fine for a command line sysadmin tool.
|
# ? Jun 27, 2016 09:38 |
|
increased containerisation on the desktop is a nice thing to see cause it's basically btrfs and cgroups but for humans
|
# ? Jun 27, 2016 12:57 |
|
Soricidus posted:I was. I think. I was using whatever graphical tool you get when as a normal person you try to install a software by clicking on a download link. it had reviews and star ratings and everything. then yeah that kinda sucks, hopefully it will improve as it matures because it's a rather new project at the moment.
|
# ? Jun 27, 2016 17:31 |
|
so i'm kind of new at this whole pbx/vpn/godonlyknows systems admin thing, i did get the openvpn passing traffic through to an asterisk pbx server and we got two phones connecting to each other over it. my boss wants to now set up a couple hundred on it how screwed am i? should i start worrying about codecs and overhead or is it still small for what these systems can handle?
|
# ? Jun 27, 2016 22:07 |
|
start drinking
|
# ? Jun 27, 2016 22:12 |
|
PBX systems are universally awful, have fun janitoring insane configs until the end of time
|
# ? Jun 27, 2016 22:16 |
|
celeron 300a posted:I haven't used docker but after hearing about it, there is something I don't get. If people are just creating these distribution independent things without virtualization, then how does it account for api changes between libc versions and kernel versions? I understand that with cgroups, you can have a process see only a specific set of processes and files, but they're sharing the same kernel at the very least. kernel guarantees abi compatibility, there isn't anything fundamentally being broken between it and libc. cgroups does indeed lead to a lot of corner cases that have resulted in security bugs celeron 300a posted:Also, wouldn't an enterprise deployment be using nfs or iscsi (for each container) to mount a working directory, instead of using the same mounted filesystem, to reduce or eliminate the chance of an errant container bringing down a whole kernel and fudging the disk? yes, it's hard to allocate I/O since it's typically allocated per-device rather than per-process/container*, the easiest workaround therefore is to have multiple devices * can depend on the io scheduler you're configured with
|
# ? Jun 27, 2016 22:52 |
|
for docker specifically, you can easily have a running docker image consume your whole filesystem ama about running graphite for a couple hours
|
# ? Jun 27, 2016 22:54 |
|
docker is bad, this seems like a much better approach https://lwn.net/Articles/644675/ idk how much traction it will get though
|
# ? Jun 27, 2016 23:00 |
|
Progressive JPEG posted:kernel guarantees abi compatibility, there isn't anything fundamentally being broken between it and libc. cgroups does indeed lead to a lot of corner cases that have resulted in security bugs thanks, this clears it up The kernel abi is stable now, I guess? I guess we aren't gonna have any new system calls until the end of time or at least they will be backwards compatible. Progressive JPEG posted:yes, it's hard to allocate I/O since it's typically allocated per-device rather than per-process/container*, the easiest workaround therefore is to have multiple devices thanks.... I'm gonna go see how much of this is already commonplace in standard deployment instructions
|
# ? Jun 28, 2016 04:27 |
|
does anyone remember from 10+ years ago when usermode linux was a thing?
|
# ? Jun 28, 2016 04:28 |
|
Apocadall posted:so i'm kind of new at this whole pbx/vpn/godonlyknows systems admin thing, i did get the openvpn passing traffic through to an asterisk pbx server and we got two phones connecting to each other over it. my boss wants to now set up a couple hundred on it is everybody going to call into the conference room at the same time?
|
# ? Jun 28, 2016 04:34 |
|
|
# ? Apr 29, 2024 12:11 |
|
Apocadall posted:so i'm kind of new at this whole pbx/vpn/godonlyknows systems admin thing, i did get the openvpn passing traffic through to an asterisk pbx server and we got two phones connecting to each other over it. my boss wants to now set up a couple hundred on it probably wont be a problem unless you're doing a lot of call recordings on the switch and even then the problem will be disk I/O. Something you should understand here is that unless asterisk needs to participate in the audio path (ex: playing an IVR or recording a call) it wont be in the audio path and will just tell the participants how to talk to each other directly. This is generally fine, but then you need to make sure your clients can actually reach each other. If you're going thru a vpn, depending on your OS, vpn client, and sip client, you may run into a 1 way audio problem where audio gets from the remote (VPN) machine to the destination but not back to the remote client. This is because the sip client you're using cant figure out what its VPN address is and its sending its local IP address as its contact address. I've never seen it happen in windows because most windows VPN clients create a virtual NIC to handle the VPN connection and that makes it easy for the SIP client to discover the correct address. It will happen with OSX depending on the vpn client you're using. Any of the cisco clients don't create virtual adapters under osx so whatever sip client you're using cant figure out what its vpn address is. The best way to fix this is to setup a STUN server inside your network. When the SIP client connects to your asterisk server (or to peers) it will ignore the ip address reported by the OS and use the IP address as seen by the STUN server which will be the VPN address.
|
# ? Jun 28, 2016 04:39 |