Home / user-126501

gluk47's questions

Martin Hope
gluk47
Asked: 2014-09-13 08:47:20 +0800 CST
  • 0

I'm designing a systemd service file for some plugin. I need to include two environment files there: first for the main program (let it be main.env), second — for the plugin (plugin.env). The plugin needs some variable VAR to be defined before it is started.

Sometimes users may want to define this variable in main.env, sometimes this variable may be missing, and actually I have no control on the environment file of the main program, this variable may be removed from it in the next release as non-obligatory.

So I want to provide a fallback value for the variable, but only in case it is not set yet. Also I want this fallback value to be actually defined in plugin.env so that users may change this value locally without losing it after update, and I use this variable at the same file to set next variables.

In bash, what I want looks like: : ${VAR:=fallback_value}. Is it possible to write this in systemd environment file?

I've looked through the manual (here), but, alas, found nothing relevant.

systemd
Martin Hope
gluk47
Asked: 2013-09-19 12:59:36 +0800 CST
  • 3

I use mtr to count number of hops to a host and pass that number to a script. I discovered -l (or --raw) option which produces an output ready for parsing. However, I'd like to know where to find this raw format description? Unfortunately, man and google give me only this:

-l

   --raw
          Use  this option to tell mtr to use the raw output format. This format is better suited for archival of the measure‐
          ment results. It could be parsed to be presented into any of the other display methods.
bash
Martin Hope
gluk47
Asked: 2013-09-16 05:52:27 +0800 CST
  • 2

I have many servers, where I deploy various binaries. For this purpose I use a script that packs them into archives, publishes to the local repository and then ssh-es to the target servers and installs.

For now I have a script that looks like this:

#!/bin/bash
packs=(app1 app2 app3 great_app main_server and_so_on)
for pack in ${packs[@]}; do
    cd "$pack"
    svn up
    make redist-package
    SERVER="$(cat ./server)"
    ssh "$SERVER" <<< "wget 'http://dev-server/$pack'; install './$pack'" # ssh uses keys to authenticate
    cd - &>/dev/null
done

here, install is a special script that unpacks and installs, our hand-made little emerge. I removed all error checking to avoid script complication.

The problem now is: I have to sign all the binaries using bsign. I want a user to enter a password once and then to pass it to bsign. Otherwise it is a pain to enter the same password individually for each of the packages.

Unfortunately, bsign has no ability such as sudo -S to read passwords from stdin. So… I tried using socat without any luck. I used this line from socat's manual, without result:

socat - EXEC:'bsign -si .',pty,setsid,ctty <<< 'My password'

bsign just shows me the password prompt as if there were no socat. After I enter the password, it works.

My guess is that it's because bsign runs gpg that in turn asks a password. Is it possible for socat to affect that gpg as well as bsign?

Update: I found bsign's option -P --passphrase-fd0 which should tell gpg to read password from fd 0 i.e. stdin. Will try it tomorrow.

Update: Failed, -P does not work. It may be caused by a buggy (custom-patched) bsign I have to use, though.

bash