|
|
Line 271: |
Line 271: |
| * source ~/.profile | | * source ~/.profile |
| * nvm ls-remote | | * nvm ls-remote |
| * nvm install ##### # (latest version, eg 5.8.0) | | * nvm install node # to get the latest release (stable by default) - you can pick a release too if needed, eg 5.8.0) |
| * nvm alias default stable | | * nvm alias default stable |
| * (update .bashrc to configure node on even non-interactive shells - see another for example) | | * (update .bashrc to configure node on even non-interactive shells - see another for example) |
Revision as of 15:36, 4 June 2016
Mediawiki - Wordpress - Phabricator
CodeLite - Eclipse - Sublime - Scite - Emacs
LibreOffice - Gimp - i3
Chromium - Firefox - IceWeasel
Linux software raid - Apache
OS installation
|
Create and boot from Ubuntu USB
|
There should always be a boot USB for this in my set, but it needs recreation on new Ubuntu versions...
- Download the latest 64-bit Ubuntu desktop iso
- Format a USB drive as FAT (NOT exFAT or NTFS)
- Burn the iso to the USB, providing a GB of space (we want to add the nvidia driver once booted)
sudo usb-creator-gtk
- Boot with it
- On startup, select the USB EFI boot option in refind, select "Try Ubuntu", hit e and add [ nouveau.noaccel=1] to grub line, hit F10 to start
- Once it is running, start System Settings, select Software, enable proprietary drivers
|
Ubuntu repo management
|
To see what repos Ubuntu is currently using:
cat /etc/apt/sources.list
|
Set up a new Pi in 10 minutes
|
The BEST thing to do is to copy the MicroSD from Carambola (marked with a black 'O'):
- open a terminal so we can watch the MicroSD /dev/sd{#} assignments
sudo tail -f /var/log/syslog
- take the carambola MicroSD card out of the Pi and put it into an Anker hub (gets less hot than the small MicroSD sleeves - and it WILL get hot!)
- put another new MicroSD card into another Anker hub
- open another terminal
dcfldd bs=4M if=/dev/sd{Letter of carambola} of=/dev/sd{Letter of new card}
- They will get HOT... I don't know how to make sure they don't get TOO hot yet... cross your fingers I guess...
- Drop the new card into the new Pi, boot
- set up a new config folder
- rm ~/config
- cd development/config
- cp -rp carambola lime && git add lime
- cd ~ && ln -s development/config/lime config
- change /etc/hostname
- change name of exfat "share" partition
exfatlabel /dev/disk/by-label/carambola_share lime_share
- edit /etc/fstab to update partition name
- (optional) change the uuid of partitions as desired (otherwise you may get kernel/userspace warnings about conflicts when mounting more than one card in an ubuntu host) - note that I've never actually done this...
tune2fs /dev/sdaX -U random
|
Set up OpenWRT on buffalo
|
choices (3 is the only sensible!):
1) hardcode all wan info and hope your network doesn't over-assign (this sucks)
2) DCHP WAN, bridge lan so ports just become another switch
3) DCHP WAN, serve up lan on different range than WAN
this is AWESOME, you can immediately admin from anything that you hardwire up to the LAN ports
set up a WAN static IP using WAN MAC if you can
otherwise, to get WAN IP:
i can connect laptop to LAN port and get a 192.168.1 address on laptop from router DHCP
then i can connect PA LAN to router WAN port and get router IP from
steps:
* reset buffalo as needed!
it always starts with LAN DHCP support for 192.168.1 range, yeah baby
wire laptop into LAN port and browse to 192.168.1.1
* set up to get WAN IP via DHCP; make note of it using LAN connection: 192.168.50.57
* Allow ssh from WAN IPs to router
openwrt admin page->Network->Firewall->Traffic rules->"open ports on router"
name: allow-wan-ssh
Protocol: TCP+UDP
external port: 22 (i could make it non-standard...)
ADD
then you can ssh to the WAN DHCP port, if you know it! for now, it's:
ssh root@192.168.50.57
* leave LAN support of 192.168.1 ON
remember you can simply wire anything into LAN ports to get an address!
and then you can browse to http://192.168.1.1 to admin the router
---
now i can ssh to it from wallee (or anywhere on PA LAN)
ssh root@192.168.50.57 root/p*
---
open https port too! let's admin from wallee
you have to open WAN port 443 in firewall config
you also have to install support for this!
opkg update
opkg install luci-ssl
/etc/init.d/uhttpd restart
but others have commented that this exposes your router and it WILL get hacked. good point.
skip for now
|
OpenELEC multi-boot install
|
The easiest way is to add a new drive just for OpenElec and install OpenElec to it with the installer. But if you want to SHARE ONE DRIVE with other boots, DO NOT DO THAT :-) Do this instead:
- On an existing refind-booted system, set up two new ext4 partitions
- one about 2GB in size, labeled [SYSTEM], and marked as bootable
- the other with 10GB or more, labeled [STORAGE]
- prep the drives (no journal, ssd trim)
- download openelec and mount; there are a couple ways:
- get the img and install to a thumb
- get the img and mount (see below)
- Set up OE drives to mount in other OSes to /openelec-system and /openelec-storage
/dev/disk/by-label/SYSTEM /openelec-system ext4 noatime 0 0
/dev/disk/by-label/STORAGE /openelec-storage ext4 noatime 0 0
- copy target/KERNEL and target/SYSTEM to SYSTEM
cp OpenELEC_img/target/KERNEL /openelec-system/
cp OpenELEC_img/target/SYSTEM /openelec-system/
- set up UEFI boot
- subl /boot/efi/EFI/refind/refind.conf (and add this block)
# MDM Trying this, from: http://openelec.tv/forum/64-installation/70783-how-to-efi-booting-openelec-on-new-pc-s-nuc-s
# Only I had to change BOOT to SYSTEM. and quiet to debugging.
menuentry OpenELEC {
icon EFI/refind/icons/os_openelec.png
volume SYSTEM
ostype Linux
loader KERNEL
# options "boot=LABEL=SYSTEM disk=LABEL=STORAGE debugging"
options "boot=LABEL=SYSTEM disk=LABEL=STORAGE quiet"
}
- I THINK you need one or more of these too, not sure!
cp target/boot/bootx64.efi /openelec-system/BOOT
cp target/boot/bootx64.efi /boot/efi/EFI/ubuntu/
cp target/boot/bootx64.efi /boot/efi/EFI/ubuntu/
|
OpenELEC boot from thumb
|
NOTE: I have the image already on a white stick with red lettering... anyway...
- Get the latest dev version (stable didn't work for me although this may change)
- dd it onto a thumb
- reboot and select to boot to the thumb in BIOS
- when the boot: line comes up, type "live" to get run a live Kodi rather than run the crufty old installer
|
Update gentoo kernel
Windows 10 new setup
|
C:\Users\Michael\AppData\Roaming\Sublime Text 3\Packages\User
|
|
Cloud and AWS tasks
|
AWS clone
|
- create a snapshot of an instance
- create an instance of a snapshot
Example...
instances -> pa-jenkins-win-test -> right-click, Create Image ->
{
name: pa-jenkins-win-test,
AMI ID: ami-57cadc3d
}
This creates a new "AMI Image"- WARNING: The source machine will be rebooted!
monitoring it, waiting for it to be ready... ready. man that took a while.
Select AMI -> rightclick -> Launch...
instance type: t2.medium (same as original; 2 vCPUs, 4GB mem)
config: SELECT THE SAME SUBNET (us-east-1d) so on the same LAN and can ping each other; (all other defaults)
storage: (default)
tag instance: { name: "pa-jenkins-win-test-older" }
security: select existing { "pa-shared immed_inno_group raleigh-windows-dev" } (as per Brad, previously - note that "raleigh" is no longer available)
LAUNCH
---
select an existing key pair or create a new pair
let's use the existing ones since this is a shared resource
2014Dec25_air.pem (it's in my docs repo)
----
monitor it, wait for it to be created...
because it was cloned from a machine that had the password changed, you can't use the existing key to regenerate
i didn't realize it was changed??
anyway, it should be the exact same way to connect to test, but with a diff ip/hostname (*-older)
yep. switched IP to 172.30.3.208 and it's good to go!
|
Expand AWS storage
|
stop instance
in instance details, right-click the block device link (typically /dev/sd1)
click the EBS ID
snapshot the volume, wait for it to get to "complete"
right-click snapshot, create volume
use the same availability zone: us-east-1d
on the volume created screen, click to show the volume
vol-b88c3c45
get the orig volume:
vol-5af8f0a0
https://console.aws.amazon.com/ec2/v2/home?region=us-east-1#Volumes:search=vol-5af8f0a0;sort=size
type this into volume filter to see them both:
vol-b88c3c45|vol-5af8f0a0
make note of the attachment info for the old 30GB drive:
Attachment information
i-ad45941d (pa-jenkins-win-test):/dev/sda1 (attached)
detach the volume! holy cow.
i thought about restarting the instance and shutting it down first...
but i think stopping it already did that! MOVE ON
right click new volume, "attach volume"
start typing instance name (pa-jenkins-win-test worked)
CHANGE TEH SELECTED DRIVE NAME (xsdf or something) to the original! /dev/sda1
geez that's scary, wonder if it will work...
YEP!!! AW#ESOMENWESSSS##$$
type "disk management", right-click the 30gb image, select Extend
DONE!!!
|
|
Ubuntu set up networking
|
Install NetworkManager, as the wpagui UI sucks
- sudo apt-get install network-manager-gnome
- YOU MUST remove interfaces from /etc/network/interfaces so wpa gives them up to nm-applet
- add nm-applet to startup if needed - i don't think it is needed as it seems to start up automatically now - try rebooting first
|
Ubuntu install Sublime
|
Sublime editor seems way better than anything else out there on linux:
- sudo add-apt-repository ppa:webupd8team/sublime-text-3
- sudo apt-get update
- sudo apt-get install sublime-text-installer
- git config --global core.editor "subl -n -w" # to use sublime with -n(ew window) and -w(ait for exit)
|
Install Node.js
|
Windows
Linux
- install Node.js using the "Node.js Version Manager" nvm details
- source ~/.profile
- nvm ls-remote
- nvm install node # to get the latest release (stable by default) - you can pick a release too if needed, eg 5.8.0)
- nvm alias default stable
- (update .bashrc to configure node on even non-interactive shells - see another for example)
- also: nvm use 4.2.1; node -v; nvm ls; nvm alias default 0.11.13; nvm use default
- also: You can create an .nvmrc file containing version number in the project root directory and it will default to that version
- example: npm install -g rad-scripts # to set the package manager to use a globally shared location
|
Ubuntu install Java
|
Install Oracle JDK 8 (9 is scheduled for release in 2016)
su -
add-apt-repository ppa:webupd8team/java
apt-get update
apt-get install oracle-java8-installer
apt-get install oracle-java8-set-default
update-alternatives --config java # make note of path
update-alternatives --config javac # make note of path
sudo subl /etc/environment
JAVA_HOME="/usr/lib/jvm/java-8-oracle/jre/bin/java" # actually, use the path from above, of course
|