The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

«‹›1028 »

jaegerx: Sep 10, 2012; Maybe this post will get me on your ignore list!

Cidrick posted:

Yeah, I haven't actually looked at it, but maybe it's possible to set up a service account and use curl to POST a login and then scrape screen output or something.

I was hoping there was a way via an RPC command or something to remotely query info about an NFS export that I just didn't know about :|

Ugh maybe windows NFS tools via powershell might be able to but I can't think of anything from the Linux side that's available. Why is there no snmp on this thing?

# ? Jun 27, 2014 04:32

Adbot: ADBOT LOVES YOU

# ? May 10, 2024 00:12

Cidrick: Jun 10, 2001; Praise the siamese

jaegerx posted:

Ugh maybe windows NFS tools via powershell might be able to but I can't think of anything from the Linux side that's available. Why is there no snmp on this thing?

Because apparently HDS wants you to pay for that luxury?

Although in some limited research it looks like this guy did some acrobatics to get catci working. I'll pass this onto our monitoring guys.

# ? Jun 27, 2014 04:55

evol262: Nov 30, 2010; #!/usr/bin/perl

joe944 posted:

Yeah, that would become quite unmanageable after a while. Best bet is to group your classes into "roles" so that you only need to include the role for a particular node. It gets a lot better when you store all of your server info in a database and start using hiera to classify the nodes.

Yes, but that's pretty overblown for a single node environment.

A mix of classes, roles, classifiers, and potentially custom providers is great for an environment large enough to need hiera and a database and maybe mcollective if you hate yourself. It's totally unnecessary for one

# ? Jun 27, 2014 04:56

joe944: Jan 31, 2004; What does not destroy me makes me stronger.

ZippySLC posted:

I'm trying to teach myself Puppet, with the goal of managing the VPS I run with it. Additionally, when I get a better grasp on things, I want to deploy it at work. So while my home deployment will be small, I want to learn how to "do it right."

This is mainly what I was responding to, I assumed he would be using more than one server at work.

Of course, it's not like I've done all of those things for my home lab for learning purposes..

# ? Jun 27, 2014 05:12

evol262: Nov 30, 2010; #!/usr/bin/perl

joe944 posted:

This is mainly what I was responding to, I assumed he would be using more than one server at work.

Of course, it's not like I've done all of those things for my home lab for learning purposes..

Ah, I missed the "eventually deploy it at work" part.

# ? Jun 27, 2014 18:58

Doctor w-rw-rw-: Jun 24, 2008

What's the current wisdom on best practices for provisioning and managing cloud VM instances? I'm looking into Chef for instance configuration, and took a look at Netflix Asgard for the provisioning very briefly, but can't tell if it's overkill or not.

Additional info: Node.js server (with compiled modules required), and a RethinkDB database.

Doctor w-rw-rw- fucked around with this message at 12:40 on Jun 28, 2014

# ? Jun 28, 2014 12:38

RFC2324: Jun 7, 2012; http 418

Don't forget the multifactor authentication!

# ? Jun 28, 2014 12:48

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

Doctor w-rw-rw- posted:

What's the current wisdom on best practices for provisioning and managing cloud VM instances? I'm looking into Chef for instance configuration, and took a look at Netflix Asgard for the provisioning very briefly, but can't tell if it's overkill or not.

Additional info: Node.js server (with compiled modules required), and a RethinkDB database.

If you're going to be using Chef as your node provisioner, Asgard is way overkill. Take a look at Chef Metal.

# ? Jun 28, 2014 15:22

Megaman: May 8, 2004; I didn't read the thread BUT...

I'm a Debian guy, and I've always installed testing and SID on every computer since I started using Linux, but I've never known the "correct" way to install testing or SID. I've noticed that sometimes when I install it it just doesn't install correctly, or it's broken sometimes (either repo keys, or installing a critical package like xorg or chromium). What is the tried and true method of installing testing/sid that everyone here uses? I've been burning stable mini.iso to a usb key, installing just base testing/sid with no extra packages, then installing xorg and whatever other packages on top of it. What does everyone else do to consistently get a working build up and running?

# ? Jun 29, 2014 02:19

kujeger: Feb 19, 2004; OH YES HA HA

Megaman posted:

I'm a Debian guy, and I've always installed testing and SID on every computer since I started using Linux, but I've never known the "correct" way to install testing or SID. I've noticed that sometimes when I install it it just doesn't install correctly, or it's broken sometimes (either repo keys, or installing a critical package like xorg or chromium). What is the tried and true method of installing testing/sid that everyone here uses? I've been burning stable mini.iso to a usb key, installing just base testing/sid with no extra packages, then installing xorg and whatever other packages on top of it. What does everyone else do to consistently get a working build up and running?

I've always used the daily/weekly testing netinstall, then running a dist-upgrade to get it to sid. Except for once when the actual installer was broken, I've never had problems with it.

e: and if you need non-free firmware to get network up and running, there's this: http://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/

kujeger fucked around with this message at 12:20 on Jun 29, 2014

# ? Jun 29, 2014 12:17

hummingbird hoedown: Sep 23, 2004; IS THAT A STUPID NEWBIE AVATAR? FUCK NO, YOU'RE GETTING A PENTAR

SKILCRAFT KREW Reppin' Quality Blind Made Products

kujeger posted:

I've always used the daily/weekly testing netinstall, then running a dist-upgrade to get it to sid. Except for once when the actual installer was broken, I've never had problems with it.

So does this mean Debian testing isn't really a rolling release? Or it kind of is? Once the testing version becomes the stable version, do you have to install the next testing version or will a dist-upgrade do that for you?

# ? Jun 29, 2014 22:59

waffle iron: Jan 16, 2004

Hummer Driving human being posted:

So does this mean Debian testing isn't really a rolling release? Or it kind of is? Once the testing version becomes the stable version, do you have to install the next testing version or will a dist-upgrade do that for you?

It depends on your /etc/apt/sources.list file. I think by default the installer writes the codename of the release, but instead you could find/replace "jessie" with "testing" and you'd always track testing.

# ? Jun 30, 2014 00:19

fuf: Sep 12, 2004; haha

Does anyone use Vifm?

This is a dumb question but I can't find the answer anywhere. You can type :shell within vifm to open a shell, but then how do I get back to vifm from the shell?

Also, is it possible to quit vifm but return to a shell with PWD set to wherever I last was in vifm? i.e:

code:

~$ vifm
[navigate to ~/test in vifm, then quit with :q]
~/test$

# ? Jun 30, 2014 14:42

Experto Crede: Aug 19, 2008; Keep on Truckin'

I bought an old 486 laptop on the cheap and I'm planning to use DSL and I'm just wondering if anyone has any experience getting DSL/linux in general running on a low spec device? I'll be installing from a CF PCMCIA adaptor and I'm almost certain the BIOS won't support booting from this, but I'm hoping the boot floppy method will work?

# ? Jul 1, 2014 21:09

fletcher: Jun 27, 2003; ken park is my favorite movie; Cybernetic Crumb

Experto Crede posted:

I bought an old 486 laptop on the cheap and I'm planning to use DSL and I'm just wondering if anyone has any experience getting DSL/linux in general running on a low spec device? I'll be installing from a CF PCMCIA adaptor and I'm almost certain the BIOS won't support booting from this, but I'm hoping the boot floppy method will work?

Why would you pay any amount of money for a 486 laptop? I would assume the only money involved with a piece of equipment like that is for paying to dispose of it. Why not get a used netbook for <$100 instead?

# ? Jul 1, 2014 21:33

Suspicious Dish: Sep 24, 2011; 2020 is the year of linux on the desktop, bro; Fun Shoe

Experto Crede posted:

I bought an old 486 laptop on the cheap and I'm planning to use DSL and I'm just wondering if anyone has any experience getting DSL/linux in general running on a low spec device? I'll be installing from a CF PCMCIA adaptor and I'm almost certain the BIOS won't support booting from this, but I'm hoping the boot floppy method will work?

You're an idiot.

# ? Jul 1, 2014 21:41

Experto Crede: Aug 19, 2008; Keep on Truckin'

fletcher posted:

Why would you pay any amount of money for a 486 laptop? I would assume the only money involved with a piece of equipment like that is for paying to dispose of it. Why not get a used netbook for <$100 instead?

This isn't for actual use, really just to play around with, see what you can do with a modern OS within an old system. If I can't get linux running properly I can just fall back to the standard DOS setup for old games.

Suspicious Dish posted:

You're an idiot.

Pretty much just a mix of too much free time and being a bit of a massive goon.

# ? Jul 1, 2014 21:45

spankmeister: Jun 15, 2008

You will not be able to run any of the current graphical shells.

Maybe something like twm

E: also, get a cf to ide adapter instead of the pcmcia one, you could boot off it that way

# ? Jul 1, 2014 21:58

fletcher: Jun 27, 2003; ken park is my favorite movie; Cybernetic Crumb

Experto Crede posted:

This isn't for actual use, really just to play around with, see what you can do with a modern OS within an old system. If I can't get linux running properly I can just fall back to the standard DOS setup for old games.

Pretty much just a mix of too much free time and being a bit of a massive goon.

Oh ok, nothing wrong with that then. Just sounds like an exercise in frustration to me, but more power to ya if you have the patience. I guess start with the Boot Floppies and go from there?

# ? Jul 1, 2014 22:03

JHVH-1: Jun 28, 2002

Experto Crede posted:

This isn't for actual use, really just to play around with, see what you can do with a modern OS within an old system. If I can't get linux running properly I can just fall back to the standard DOS setup for old games.

Pretty much just a mix of too much free time and being a bit of a massive goon.

Please tell me you saw George R. R. Martin on Conan talking about how he writes on an old MS Dos machine and that is what made you decide to do this.

# ? Jul 1, 2014 22:42

evol262: Nov 30, 2010; #!/usr/bin/perl

Experto Crede posted:

I bought an old 486 laptop on the cheap and I'm planning to use DSL and I'm just wondering if anyone has any experience getting DSL/linux in general running on a low spec device? I'll be installing from a CF PCMCIA adaptor and I'm almost certain the BIOS won't support booting from this, but I'm hoping the boot floppy method will work?

Honestly, you'd be better off with NetBSD or OpenBSD if you insist on doing this.

# ? Jul 2, 2014 00:22

CaptainSarcastic: Jul 6, 2013

Experto Crede posted:

I bought an old 486 laptop on the cheap and I'm planning to use DSL and I'm just wondering if anyone has any experience getting DSL/linux in general running on a low spec device? I'll be installing from a CF PCMCIA adaptor and I'm almost certain the BIOS won't support booting from this, but I'm hoping the boot floppy method will work?

Sorry - the oldest hardware I've run Linux on is a Pentium II Toshiba laptop. Puppy actually made it surprisingly usable, aside from being stuck with the crappy screen resolution the laptop was built with.

# ? Jul 2, 2014 01:33

thebigcow: Jan 3, 2001; Bully!

Don't you need a processor with physical addresses extensions? Or is DSL an ancient kernel?

# ? Jul 2, 2014 02:52

Doctor w-rw-rw-: Jun 24, 2008

Per the previous suggestions I've now got a chef-metal setup working to where I can get it to boot up the EC2 instances and configure chef clients on them. Thanks!

Now, a couple of other questions:
1. Should I stick with local recipes on the chef-metal box, or run a chef server on EC2?
2. The source is on GitHub, what's a sensible deploy strategy? I don't think that running "git pull" on each server is necessarily the best idea.
3. Ideally, each individual server would have zero filesystem writes, dealing only with the database and S3. Is EC2 going to make a new EBS volume for each instance, or can I somehow run off of read-only volumes and deduplicate?

# ? Jul 2, 2014 08:09

pipebomb: May 12, 2001; Dear God, what is it like in your funny little brains?
It must be so boring.

Hey all. I have a DigitalOcean server (in SF, hosting my sites) and am thinking about setting up another in NYC or Amsterdam in order to circumvent sports blackouts (MLB) as well as proxy some traffic that I don't want my ISP monitoring. I figure I can also serve some site content from it as well for European visitors.

Any thoughts on the best vpn/proxy setup to implement - preferred servers, apps, etc?

# ? Jul 2, 2014 15:48

fletcher: Jun 27, 2003; ken park is my favorite movie; Cybernetic Crumb

Doctor w-rw-rw- posted:

Per the previous suggestions I've now got a chef-metal setup working to where I can get it to boot up the EC2 instances and configure chef clients on them. Thanks!

Now, a couple of other questions:
1. Should I stick with local recipes on the chef-metal box, or run a chef server on EC2?
2. The source is on GitHub, what's a sensible deploy strategy? I don't think that running "git pull" on each server is necessarily the best idea.
3. Ideally, each individual server would have zero filesystem writes, dealing only with the database and S3. Is EC2 going to make a new EBS volume for each instance, or can I somehow run off of read-only volumes and deduplicate?

I still haven't bothered with a chef server, I just use fabric to git pull my chef repo onto the machine and then run chef solo. It works well enough but I only have a couple machines, I could see wanting to move to chef server if I had a few more boxes.

On the third point, what about log files and such? EC2 won't make any EBS volumes for each instance unless you tell it to. Pretty much all the instance types come with instance storage though (except for the free tier I think). You don't have to use the instance storage beyond the 8GB root volume for the OS, but I don't think you save anything by not using it, it's still built into the hourly price of that instance.

pipebomb posted:

Hey all. I have a DigitalOcean server (in SF, hosting my sites) and am thinking about setting up another in NYC or Amsterdam in order to circumvent sports blackouts (MLB) as well as proxy some traffic that I don't want my ISP monitoring. I figure I can also serve some site content from it as well for European visitors.

Any thoughts on the best vpn/proxy setup to implement - preferred servers, apps, etc?

I used SSH to setup a SOCKS proxy to get around the NHL blackouts. It was crazy easy, I think it was just ssh -D 8080 <server> and then went to Firefox settings and tell it to use localhost:8080 as the SOCKS proxy. However I did try to do this recently with a friend for MLB and he said it was still blacked out even with the proxy. Maybe they are wising up to what people are doing?

# ? Jul 2, 2014 20:28

pipebomb: May 12, 2001; Dear God, what is it like in your funny little brains?
It must be so boring.

For mobile, mlb.tv do a gps lookup. For desktop/set-top, it is IP based. I figure if I run something out of Singapore or Amsterdam for $5 per month, I can avoid the blackouts - they don't BO foreign traffic, even for AllStar/World Series. Not to mention having that redundancy.

I'll look at what you mentioned, thanks.

# ? Jul 2, 2014 21:37

evol262: Nov 30, 2010; #!/usr/bin/perl

pipebomb posted:

For mobile, mlb.tv do a gps lookup. For desktop/set-top, it is IP based. I figure if I run something out of Singapore or Amsterdam for $5 per month, I can avoid the blackouts - they don't BO foreign traffic, even for AllStar/World Series. Not to mention having that redundancy.

I'll look at what you mentioned, thanks.

Just pay for ipvanish.

# ? Jul 2, 2014 22:22

spankmeister: Jun 15, 2008

pipebomb posted:

Hey all. I have a DigitalOcean server (in SF, hosting my sites) and am thinking about setting up another in NYC or Amsterdam in order to circumvent sports blackouts (MLB) as well as proxy some traffic that I don't want my ISP monitoring. I figure I can also serve some site content from it as well for European visitors.

Any thoughts on the best vpn/proxy setup to implement - preferred servers, apps, etc?

Use openvpn

# ? Jul 2, 2014 23:22

Hollow Talk: Feb 2, 2014

spankmeister posted:

Use openvpn

Seconding using OpenVPN. You can simply use it to run all your traffic through the vpn and thus through your European system, including DNS requests etc (you might have to push dns servers via the vpn server). Otherwise you might be "leaking" traffic, since the SOCKS proxy only works for applications you specifically instructed to make use of it.

OpenVPN is also surprisingly light on hardware utilisation if only a handful of people are using it, and it gives you the added bonus of decent key management. I use a VPS in a similar way (though not with DO) and I really like it.

# ? Jul 3, 2014 01:32

spankmeister: Jun 15, 2008

Yeah. If you do go the socks route make sure to set up your browser to forward dns queries through the proxy as well.

But really just use openvpn.

# ? Jul 3, 2014 06:47

taqueso: Mar 8, 2004

I'm playing around with making my older intel mac mini into an equalizer, basically setting up something like this. I installed Ubuntu 14.04 Server on it. Pretty much everything seems to be good to go, but I'm not very experienced with wifi on the command line and I'm having trouble figuring out how to automatically switch between network interfaces based on availability.

Wireless is working, I can connect to the network/etc. If I boot without wired ethernet connected, it hangs for awhile (2-3 minutes) "Waiting for network configuration...", but after that everything functions as expected. If I remove eth0 from /etc/network/interfaces, there is no delay on boot.

If I boot with eth0 connected and both interfaces enabled, I end up with a default route over eth0. If I unplug eth0, the route never changes.

What do I need to do to get the system to automatically switch between networks, change the default route, etc?

edit: Got it going, wicd does what I wanted.

taqueso fucked around with this message at 07:47 on Jul 4, 2014

# ? Jul 4, 2014 00:52

My Rhythmic Crotch: Jan 13, 2011

I am having one hell of a time figuring out why netatalk (afpd) is not working on a Fedora 20 machine.

code:

[root@helium ~]# systemctl status  selinux.service
selinux.service
   Loaded: not-found (Reason: No such file or directory)
   Active: inactive (dead)

[root@helium ~]# systemctl status iptables
iptables.service
   Loaded: not-found (Reason: No such file or directory)
   Active: inactive (dead)

[root@helium ~]# systemctl status netatalk
netatalk.service - File and Printer sharing for Macintosh clients
   Loaded: loaded (/usr/lib/systemd/system/netatalk.service; enabled)
   Active: active (exited) since Fri 2014-07-04 14:18:40 EDT; 3h 4min ago
Process: 18709 ExecStart=/bin/sh -c exec /usr/libexec/netatalk/netatalk.sh (code=exited, status=0/SUCCESS)
 Main PID: 18709 (code=exited, status=0/SUCCESS)
   CGroup: /system.slice/netatalk.service
&#9500;&#9472;18715 /usr/sbin/cnid_metad -l log_note
&#9492;&#9472;18718 /usr/sbin/afpd -U uams_guest.so,uams_dhx.so,uams_dhx2.so -g nobody -c 20 -n helium

Jul 04 14:18:40 helium.localdomain systemd[1]: Starting File and Printer sharing for Macintosh clients...
Jul 04 14:18:40 helium.localdomain systemd[1]: Started File and Printer sharing for Macintosh clients.
Jul 04 14:18:40 helium.localdomain afpd[18718]: AFP/TCP started, advertising 10.0.0.100:548 (2.2.3)

[root@helium ~]# systemctl status avahi-daemon
avahi-daemon.service - Avahi mDNS/DNS-SD Stack
   Loaded: loaded (/usr/lib/systemd/system/avahi-daemon.service; enabled)
   Active: active (running) since Fri 2014-07-04 14:18:35 EDT; 3h 4min ago
 Main PID: 18693 (avahi-daemon)
   Status: "Server startup complete. Host name is helium.local. Local service cookie is 67214613."
   CGroup: /system.slice/avahi-daemon.service
           &#9500;&#9472;18693 avahi-daemon: running [helium.local]
           &#9492;&#9472;18695 avahi-daemon: chroot helper

...: Successfully dropped remaining capabilities.
...: Loading service file /services/afp.service.
...: Joining mDNS multicast group on interface em1.IPv4 with address 10.0.0.100.
...: New relevant interface em1.IPv4 for mDNS.
...: Network interface enumeration completed.
...: Registering new address record for fe80::76d4:35ff:fe1c:e0d7 on em1.*.
...: Registering new address record for 10.0.0.100 on em1.IPv4.
...: Registering HINFO record with values 'X86_64'/'LINUX'.
...: Server startup complete. Host name is helium.local. Local service cookie i...14613.
...: Service "helium share" (/services/afp.service) successfully established.
Hint: Some lines were ellipsized, use -l to show in full.

So I've got iptables off, selinux off, and the daemons have been configured and are running. However I get this cryptic message when trying to mount:

code:

mount_afp -i afp://10.0.0.100/share /mnt/share
User: derp
Password: 
mount_afp: AFPMountURL returned error -1069, errno is -1069

I'm not having any luck figuring out the error message. I've set this up on several other machines with no problems, but I'm really stuck here :raise:

My Rhythmic Crotch fucked around with this message at 22:29 on Jul 4, 2014

# ? Jul 4, 2014 22:26

Doctor w-rw-rw-: Jun 24, 2008

My Rhythmic Crotch posted:

I'm not having any luck figuring out the error message. I've set this up on several other machines with no problems, but I'm really stuck here

http://www.opensource.apple.com/source/CarbonHeaders/CarbonHeaders-18.1/MacErrors.h

code:

  aspNoServers                  = -1069, /*No servers at that address*/

# ? Jul 4, 2014 22:49

hifi: Jul 25, 2012

My Rhythmic Crotch posted:

I am having one hell of a time figuring out why netatalk (afpd) is not working on a Fedora 20 machine.
code:
[root@helium ~]# systemctl status  selinux.service
selinux.service
   Loaded: not-found (Reason: No such file or directory)
   Active: inactive (dead)

[root@helium ~]# systemctl status iptables
iptables.service
   Loaded: not-found (Reason: No such file or directory)
   Active: inactive (dead)
So I've got iptables off, selinux off

You should be checking for those with "getenforce" and "iptables -L" (the latter command as root). If iptables shows something then it's probably getting fed by firewalld, the default fedora firewall. You can disable it through systemctl or easily add a rule to it with "firewall-cmd --add-port XXX (and add --permanent to make the rule permanent). nmap is also probably really useful in this situation (use from both ends of the connection).

# ? Jul 5, 2014 06:05

My Rhythmic Crotch: Jan 13, 2011

It was firewalld. Thank you for the advice!

# ? Jul 6, 2014 18:02

Thermopyle: Jul 1, 2003; ...the stupid are cocksure while the intelligent are full of doubt. �Bertrand Russell

Every once in awhile I hit Ctrl-Alt-F10 on my Ubuntu 14.04 running in a VMWare Player because it's close to another hotkey I use quite often.

When I do this the guest screen goes black, and I cannot do anything to switch back to tty7 or any other virtual terminal. Note that this isn't reserved to tty10, the same thing happens when I switch to tty1-6 as well. I don't get any terminal...just a black screen.

Some Googlin led me to try using Ctrl-Alt-Space-F7 instead of just Alt-F7, but that has no effect.

What's going on here?

# ? Jul 7, 2014 20:12

a slime: Apr 11, 2005

I have a server set up that only allows connections from a specific address, which corresponds to a VPS of mine. I would like to tunnel traffic from my laptop through my VPS and ultimately to this restricted server. Normally I would do this with OpenSSH,

code:

ssh -L 50340:restrictedserver:9000 slime@myvps -N

However, my VPS has multiple addresses, and the restricted server only accepts access from one of them. How can I tell OpenSSH to bind to a specific remote address on my VPS?

a slime fucked around with this message at 15:14 on Jul 8, 2014

# ? Jul 8, 2014 15:12

fuf: Sep 12, 2004; haha

I may well not be understanding your question but could you put something in an ssh config file?

~/.ssh/config:

code:

Host myVPS
   Hostname 123.123.123.123
   Port 2211

# ? Jul 8, 2014 15:48

Adbot: ADBOT LOVES YOU

# ? May 10, 2024 00:12

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

a slime posted:

I have a server set up that only allows connections from a specific address, which corresponds to a VPS of mine. I would like to tunnel traffic from my laptop through my VPS and ultimately to this restricted server. Normally I would do this with OpenSSH,
code:
ssh -L 50340:restrictedserver:9000 slime@myvps -N
However, my VPS has multiple addresses, and the restricted server only accepts access from one of them. How can I tell OpenSSH to bind to a specific remote address on my VPS?

If you're saying you want to use a specific interface/IP address to make connections to your remote server, you're describing routing. You'll want to update your routing table so traffic to a particular endpoint always uses the egress interface you've chosen.

# ? Jul 8, 2014 16:26

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

«‹›1028 »