BIMP-Plugin für GIMP unter Opensuse LEAP

Manchmal benötigt man ein Tool, um unter GIMP typische Bild-Manipulations-Operationen wie etwa das Skalieren auf einen Stapel von mehreren ausgewählten Bildern anzuwenden.

Es gibt für das Skalieren oder andere elementare Bildoperationen unter Linux natürlich auch andere Werkzeuge zur Bild-Batchverarbeitung. Gwenview z.B. bietet unter dem Menüpunkt "Module >> Stapelverarbeitung" das Meiste an, was der Normal-Anwender so braucht. Aber warum nicht auch ein einfach zu bedienendes Plugin unter GIMP zur Verfügung haben?

BIMP heißt eine der möglichen Antworten. BIMP ist einfach zu installieren; der Prozess ist unter der Github-Adresse im README.md beschrieben. Man lädt z.B. von Github
https://github.com/alessandrofrancesconi/gimp-plugin-bimp
die gesamte Verzeichnisstruktur als ZIP-file herunter (grüner Knopf rechts oben: "Clone or download") und expandiert es in ein Verzeichnis seiner Wahl.

Dann mittels eines Terminalfensters (Kommandozeile!) zum Verzeichnis navigieren und dort

make && make install

ausführen. Die Installation der kompilierten Plugin-Dateien erfolgt dann unter dem "/home/USER/gimp-2.8/plug-ins"-Verzeichnis. Eine Installation für alle User muss man dagegen als Root mit

make && sudo make install-admin

in die Wege leiten.

Fehlende Bibliotheken?

Leider klappt das angegebene Rezept auf einer Leap-(x86_64)-Standardinstalltion nicht unmittelbar, wenn man nicht noch weitere Bibliotheken lädt. Die im README angegebenen Bibliotheken gibt es aber unter Opensuse Leap nicht mit der für Ubuntu/Debian gültigen Bezeichnung.

Treten während der Kompilation Fehlermeldungen auf, etwa zu einem nicht gefundenen "gimptool-2.0", so muss man unter Leap (z.B. mit YaST) vielmehr das RPM-Paket "gimp-devel" laden. Bei anderen Fehlermeldungen ggf. auch noch das Paket "pcre-devel".
(Ansonsten waren bei mir neben dem "gimp"-Paket installiert "gimp", "libgimp-2_0-0", "libgimpui-2_0-0", "libpcre1", "libpcre16-0", "libpcrecpp0", "libpcreposix0"). Und weitere Bibliotheken, die hier aber nichts zur Sache tun.

Danach läßt sich BIMP kompilieren. Die auftretenden Warnungen habei ich tapfer ignoriert. Anschließend startet man GIMP und öffnet dort das BIMP-Interface über den Menüpunkt "Datei >> Batch Image Manipulation" öffnen. Die Bedienung ist logisch und simpel. Sehr schön ist,

  1. dass man eine Serie von hintereinander geschalteten Bearbeitungsschritten definieren kann
  2. und dass man jede unter GIMP verfügbare Bildverarbeitungsaktion bei Bedarf aus einer Liste verfügbarer Verfahren auswählen und parameterieren kann.

Die Zuordnung einer GIMP-Bildverarbeitungsaktion zu einem definierten Schritt erfolgt über ein Auswahlmenü mit Standardaktionen bzw. dem Punkt "Andere GIMP-Anwendung".

bimp1
bimp2

Zudem kann man eine Serie von Verarbeitungsschritten auch abspeichern.

Einfach, gut und nützlich! Herzlichen Dank an den Entwickler Alessandro Francesconi!

Open Visual Traceroute, Opensuse Leap 42.1 and sudo – howto – I

IP GeoLocation is useful - not only in the context of professional penetration testing, but also for answering very legitimate questions of both system administrators and normal users. E.g.: Where does my web hosting provider really host my website or all my cloud data? In Germany with relatively strict laws regarding personal data protection, the EU or somewhere else where privacy may not be respected to the full extent I expect? The German provider 1&1, for example, is very active on the French market, too, but at least very many, if not all French websites seem to be hosted on servers in German computation centers. Good to know for a French customer, both legally and also in case of technical problems. Also sometimes admins may need to clarify why and where there are latency problems, when a user tries to connect to e.g. web or database servers over the Internet. Another field where IP GeoLocation is very helpful is the analysis of the origin of spam mails from the IP information in mail headers. And, and ...

Of course, there are GeoLocation services available on the Internet - but the documentation of result data very often is pretty inconvenient. Now, you could write e.g. a python program for GeoLocation on your own - which is a nice exercise, but you will soon find out that you must have a profound knowledge about networking protocol details. So, from the point of view of a standard Linux user or admin it would be nice to have an easy to use, graphical and open-source solution for IP GeoLocation available on the desktop. Such a program is "Open Visual Traceroute" [OVT] of Leo Lewis.

OVT is Java based; for IP GeoLocation it uses the CityLight-database of the company Mindmax. It has a nice graphical interface for the geographical display of the results of tracerouting. It has some simple but helpful log and documentation functionality - and even allows for some "sniffing" inspection into data traffic on network interfaces. Note that using the latter capability has legal implications - so, always (!) get the explicit allowance from your company owners, if you intend to use this functionality of OVT - e.g. in the course of penetration test. In general: Consider possible legal restrictions of the use of OVT (especially in Germany) - and do it carefully.

This having said: Can we technically use OVT on an Opensuse Leap 42.1 system? The answer is yes, but .... The "BUT" in this case does not refer to functionality, but to general security considerations, which we shall discuss in a later article.

However, let us first have a look at some elementary obstacles which prevent the successful execution of the program on Opensuse Leap desktop. Starting the application as a normal user leads to a variety of different error messages from the Java side, problems with X11 access and right problems regarding the essential package capturing functionality. One reason and problematic aspect of the program is that it requires root rights for network analysis capabilities. You think of "sudo" ? Yeah, but .... Actually one can learn a bit about "sudo" when dealing with OVT.

In this article I shall describe 3 simple ways of using OVT on Opensuse Leap 42.1. Note that these approaches are only reasonable for single user machines! Actually, due to its special sniffing capabilities OVT is a program to which only selected users should get access to on a multiuser system. In addition running Java and network package analysis as root leaves us with an uneasy feeling .... But let us get it running at all first.

In a second article we make the installation and availability of the program to selected users a bit more convenient. During the discussion I touch some general "sudo" issues, which may be useful in other contexts. A third article then argues for using the program in a chroot jail or virtual environment to improve security.

I have posted some of my approaches also in the discussion forum of OVT on Sourceforge (https://sourceforge.net/p/openvisualtrace/discussion/1799338/thread/76f2b798/).

Download / Installation

Installation of OVT is quite simple and basically means to place a folder into some reasonable branch of the folder hierarchy. A shell script is used to start the program.

The web site http://visualtraceroute.net/ offers the download of different installation packages for different operating systems. For our Opensuse system the most reasonable choice is "Download Universal". The zip-file behind it expands on our SuSE machine into a folder "OpenVisualTraceRoute1.6.3" with a shell script "ovtr.sh" at the highest level. The additional "README.txt" file there informs about the present and older versions of OVT - in my case version 1.6.3.

On my system I have placed the contents of the mentioned folder into a directory "/opt/ovt/". But you may take another choice. We abbreviate the full path to this installation directory during the rest of this article by "/PATH/TO/OVT/" and call it "OVT-directory". The OVT-directory contains a "lib" subdirectory which supplies the required "jar" archives.

In principle the OVT program could be started from the OVT-directory by executing the "ovtr.sh" script there. However, this does not work on OS Leap 42.1 - and results in a variety of error messages.

Requirements for running OVT with full functionality

All steps to overcome the cascade of different errors can better be understood by the regarding the four requirements that must be fulfilled to get access to OVT's capabilities:

  1. The program must find the "libpcap" library. Actually, in the present version OVT assumes that this library can be found in form of a file "libpcap.so.0.8" in some directory of the PATH environment variable. However, such a file does not exist on Opensuse Leap!
  2. The program needs to run with root rights (or certain network privileges on a system that support file capabilities). The reason is that it uses libpcap, i.e. network packet capturing on detected network interfaces.
  3. OVT must find and read the shell environment variable DISPLAY - which under some circumstances is not completely trivial.
  4. OVT must get the right to access the X11 display of the non-privileged user which started the X11 desktop.

This is the result of some trial and error testing plus a look at some discussions at Sourceforge.

Make OVT start at all as a non-privileged user - but with reduced functionality

To get the program running at all for a non-privileged user we must modify the startup script on Opensuse. So, please, change the content of "ovtr.sh" to

#!/bin/bash
#sudo java -Xmx512m -jar org.leo.traceroute.jar
java -Xmx512m -jar org.leo.traceroute.jar

I.e., omit the "sudo" command. The reason for this will become clearer later on. In addition you must the assign execution right to "ovtr.sh". You then get OVT running in the following form:

me@mytux:/opt/ovt> ./ovtr.sh
13:01:33.276 [main] INFO  org.leo.traceroute.Main - Open Visual Traceroute 1.6.3 
13:01:33.284 [main] INFO  org.leo.traceroute.install.Env - Java run-time version: 1.8.0_91
13:01:33.285 [main] INFO  org.leo.traceroute.install.Env - NASA World Wind Java 2.0 2.0.0
13:01:33.285 [main] INFO  org.leo.traceroute.install.Env - /usr/java/packages/lib/amd64:/usr/lib64:/lib64:/lib:/usr/lib
13:01:33.285 [main] INFO  org.leo.traceroute.install.Env - OS:Linux / arch:amd64
Locale en_GB
13:01:36.063 [SwingWorker-pool-1-thread-1] INFO  o.leo.traceroute.core.geo.GeoService - Use geoip db /home/rmo/ovtr/GeoLiteCity.dat which is 8 day(s) old
13:01:36.772 [SwingWorker-pool-1-thread-1] WARN  o.leo.traceroute.core.ServiceFactory - Cannot find a suitable network device for tracing.
java.lang.UnsatisfiedLinkError: /opt/ovt/native/linux/x64/libjpcap.so: libpcap.so.0.8: Kann die Shared-Object-Datei nicht öffnen: Datei oder Verzeichnis nicht gefunden
        at java.lang.ClassLoader$NativeLibrary.load(Native Method) ~[na:1.8.0_91]
        at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1941) ~[na:1.8.0_91]
        at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1857) ~[na:1.8.0_91]
        at java.lang.Runtime.loadLibrary0(Runtime.java:870) ~[na:1.8.0_91]
        at java.lang.System.loadLibrary(System.java:1122) ~[na:1.8.0_91]
        at jpcap.JpcapCaptor.<clinit>(JpcapCaptor.java:251) ~[jpcap.jar:na]
        at org.leo.traceroute.core.network.JPcapNetworkService.init(JPcapNetworkService.java:90) ~[org.leo.traceroute.jar:na]
        at org.leo.traceroute.core.ServiceFactory.init(ServiceFactory.java:110) [org.leo.traceroute.jar:na]
        at org.leo.traceroute.Main$4.doInBackground(Main.java:115) [org.leo.traceroute.jar:na]
        at org.leo.traceroute.Main$4.doInBackground(Main.java:111) [org.leo.traceroute.jar:na]
        at javax.swing.SwingWorker$1.call(SwingWorker.java:295) [na:1.8.0_91]
        at java.util.concurrent.FutureTask.run(FutureTask.java:266) [na:1.8.0_91]
        at javax.swing.SwingWorker.run(SwingWorker.java:334) [na:1.8.0_91]
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) [na:1.8.0_91]
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [na:1.8.0_91]
        at java.lang.Thread.run(Thread.java:745) [na:1.8.0_91]
13:01:36.776 [SwingWorker-pool-1-thread-1] WARN  o.l.t.c.n.JNetCapNetworkService - Cannot find a suitable network device for tracing.
java.lang.UnsatisfiedLinkError: /opt/ovt/native/linux/x64/libjnetpcap.so: libpcap.so.0.8: Kann die Shared-Object-Datei nicht öffnen: Datei oder Verzeichnis nicht gefunden
        at java.lang.ClassLoader$NativeLibrary.load(Native Method) ~[na:1.8.0_91]
        at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1941) ~[na:1.8.0_91]
        at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1857) ~[na:1.8.0_91]
        at java.lang.Runtime.loadLibrary0(Runtime.java:870) ~[na:1.8.0_91]
        at java.lang.System.loadLibrary(System.java:1122) ~[na:1.8.0_91]
        at org.jnetpcap.Pcap.<clinit>(Unknown Source) ~[jnetpcap.jar:1.3.0]
        at org.leo.traceroute.core.network.JNetCapNetworkService.init(JNetCapNetworkService.java:71) ~[org.leo.traceroute.jar:na]
        at org.leo.traceroute.core.ServiceFactory.init(ServiceFactory.java:111) [org.leo.traceroute.jar:na]
        at org.leo.traceroute.Main$4.doInBackground(Main.java:115) [org.leo.traceroute.jar:na]
        at org.leo.traceroute.Main$4.doInBackground(Main.java:111) [org.leo.traceroute.jar:na]
        at javax.swing.SwingWorker$1.call(SwingWorker.java:295) [na:1.8.0_91]
        at java.util.concurrent.FutureTask.run(FutureTask.java:266) [na:1.8.0_91]
        at javax.swing.SwingWorker.run(SwingWorker.java:334) [na:1.8.0_91]
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) [na:1.8.0_91]
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [na:1.8.0_91]
        at java.lang.Thread.run(Thread.java:745) [na:1.8.0_91]
13:01:37.272 [AWT-EventQueue-0] INFO  org.leo.traceroute.Main - Startup completed in 4001ms

 
Despite the many warnings the application opens and displays the following graphical elements, which are quite self-explanatory.
ovt1

However, trying to traceroute some web-address by pressing the orange button results in the displayed error. The reason is that "/usr/sbin" is not included in the PATH variable. Actually, in http://visualtraceroute.net/installation it is recommend to also include

export PATH="$PATH:/usr/sbin/"; java -Xmx512m -jar -Djava.awt.headless=false org.leo.traceroute.jar

in "ovtr.sh". Lets do it

#!/bin/bash
export export PATH="$PATH:/usr/sbin/"; java -Xmx512m -jar org.leo.traceroute.jar

 
Afterwards, we can run "traceroute" ....
ovt3

However, if we compare our screenshot with illustrations on the site http://visualtraceroute.net/ we detect that our interface lacks a button, e.g. for sniffing. We come back to this a bit later.

Why did we need the "/usr/sbin" in our PATH? Because OVT actually uses the "OS's traceroute" command as a fallback - if it cannot analyze network packages itself another way. We can conclude this fact from OVT's "settings" options (tools button on the right):

ovt2

The option "Use OS traceroute" is set to "Yes" in the combobox - and, under the given conditions, this option cannot be changed.

Give OVT the required libpcap !

Actually, the native Linux traceroute command is a bit limited in its capabilities. Therefore, its results are sometimes not convincing - in very many examples the traceroute analysis is stopped by restrictions in intermediate computation centers. A look at the discussions on OVT at Sourceforge and the version history reveals that OVT, therefore, has its own traceroute program which is based on capturing anad analyzing network protocol packets:

The access to "libpcap" is not only required for sniffing, but also for tracerouting!

The feedback on our terminal from the program start shows 2 times that "libpcap.so.0.8" is missing. In my opinion for Linux systems it would have been more clever to search for "libpcap.so.1". But things are easy to remedy. As root we search for "libpcap.so*", find it under "/usr/lib64" and simply add a new soft link there:

mytux:/usr/lib64 # ln -s /usr/lib64/libpcap.so.1 libpcap.so.0.8

Actually, "libpcap.so.1" is in turn linked to the real present version on Opensuse systems, namely "libpcap.so.1.5.3":

mytux:/usr/lib64 # ls -lisa | grep libpcap
 793187     0 lrwxrwxrwx   1 root root       23 Aug 12 15:00 libpcap.so.0.8 -> /usr/lib64/libpcap.so.1
 797513     0 lrwxrwxrwx   1 root root       16 Oct 29  2015 libpcap.so.1 -> libpcap.so.1.5.3
 793262   276 -rwxr-xr-x   1 root root   279312 Oct 25  2015 libpcap.so.1.5.3  

 

A new OVT start now results in clean messages without any errors or warnings

me@rux:/opt/ovt> ./ovtr.sh
15:03:33.154 [main] INFO  org.leo.traceroute.Main - Open Visual Traceroute 1.6.3 
15:03:33.166 [main] INFO  org.leo.traceroute.install.Env - Java run-time version: 1.8.0_91
15:03:33.167 [main] INFO  org.leo.traceroute.install.Env - NASA World Wind Java 2.0 2.0.0
15:03:33.167 [main] INFO  org.leo.traceroute.install.Env - /usr/java/packages/lib/amd64:/usr/lib64:/lib64:/lib:/usr/lib
15:03:33.168 [main] INFO  org.leo.traceroute.install.Env - OS:Linux / arch:amd64
Locale en_GB
15:03:35.468 [SwingWorker-pool-1-thread-1] INFO  o.leo.traceroute.core.geo.GeoService - Use geoip db /home/rmo/ovtr/GeoLiteCity.dat which is 8 day(s) old
15:03:36.906 [AWT-EventQueue-0] INFO  org.leo.traceroute.Main - Startup completed in 3757ms

 
However, still no sniffing button and no change in the OVT settings - OVT still uses "/usr/sbin/traceroute".

Does OVT work for user root?

The reason why OVT still does not work as expected is that packet capturing requires root rights or - on systems supporting file capabilities - the setting of some special rights for certain files (we come back to this in the last article of this mini series). Ever tried "wireshark" or "tcpdump"? For these programs - or parts of them - special rights are required, too. To get a step further we now again take the original, unmodified "ovtr.sh" with the sudo command

#!/bin/bash
sudo java -Xmx512m -jar org.leo.traceroute.jar

and try to run it as user "root". The "sudo" included in the script should do no harm now as we already are root. Or, at least we think so ...

So, as an unprivileged user we open a root terminal on our desktop, type in the root password to get shell access or use "su -" in a standard terminal. But, unfortunately, we get a new error:

  
mysystem:/opt/ovt # ./ovtr.sh
15:26:39.153 [main] INFO  org.leo.traceroute.Main - Open Visual Traceroute 1.6.3 
15:26:39.161 [main] INFO  org.leo.traceroute.install.Env - Java run-time version: 1.8.0_91
15:26:39.161 [main] INFO  org.leo.traceroute.install.Env - NASA World Wind Java 2.0 2.0.0
15:26:39.162 [main] INFO  org.leo.traceroute.install.Env - /usr/java/packages/lib/amd64:/usr/lib64:/lib64:/lib:/usr/lib
15:26:39.162 [main] INFO  org.leo.traceroute.install.Env - OS:Linux / arch:amd64
Locale en_GB
15:26:39.636 [main] ERROR org.leo.traceroute.Main - Uncaught error
java.awt.HeadlessException: 
No X11 DISPLAY variable was set, but this program performed an operation which requires it.
        at java.awt.GraphicsEnvironment.checkHeadless(GraphicsEnvironment.java:204) ~[na:1.8.0_91]
        at java.awt.Window.<init>(Window.java:536) ~[na:1.8.0_91]
        at java.awt.Frame.<init>(Frame.java:420) ~[na:1.8.0_91]
        at javax.swing.JFrame.<init>(JFrame.java:232) ~[na:1.8.0_91]
        at org.leo.traceroute.ui.TraceRouteFrame.<init>(TraceRouteFrame.java:52) ~[org.leo.traceroute.jar:na]
        at org.leo.traceroute.Main.main(Main.java:87) ~[org.leo.traceroute.jar:na]

 
What is the reason for this failure? No X11 DISPLAY variable? An "env" command shows:

mysystem:/opt/ovt # env | grep DISPL
DISPLAY=:0

It took me a bit to see it: The origin of the error actually is the sudo-command in "ovtr.sh" ! Why? The important thing to be aware of is:

"sudo" does not invoke any kind of shell. So, even if the user root had defined special startup scripts for login or execution shells, which specified the export of DISPLAY, such scripts would not be invoked by a "sudo" and as a result the DISPLAY variable would be missing. Therefore, Java complains about the X11 DISPLAY variable!

But, as we know, on Opensuse systems a graphical program started from a root terminal/shell under KDE does know the DISPLAY variable and does start on the unprivileged user's X11 desktop.
[ Off topic: On OS LEAP 42.1 KDE programs are actually started from the root terminal with ignoring basic KDE settings (font sizes, etc.) of the unprivileged and the root user - which is a mess in general, but it is unimportant here as Java's AWT is responsible for the graphics, fonts, etc of the OVT interface.]

So, again: What about turning off the "sudo"?

Solution 1: Run OVT as user root - but without "sudo" in the start script ovtr.sh

Actually, any normal user on a standard (!) unmodified Opensuse system would have to know the root password in case he/she wanted to perform a command with "sudo". Reason is a standard entry in the file "/etc/sudoers":

Defaults        targetpw 

Note: On other systems (as Kali) this entry may not be active after installation. But with this entry you may as well start OVT directly from a root shell. "sudo" would ask you for the root password anyway. So, our first working solution will be to start OVT as user root from a root shell, but without the "sudo" in the start script "ovtr.sh":

#!/bin/bash
java -Xmx512m -jar org.leo.traceroute.jar

And really:

  
mysystem:/opt/ovt # ./ovtr.sh
19:17:23.021 [main] INFO  org.leo.traceroute.Main - Open Visual Traceroute 1.6.3 
19:17:23.029 [main] INFO  org.leo.traceroute.install.Env - Java run-time version: 1.8.0_91
19:17:23.030 [main] INFO  org.leo.traceroute.install.Env - NASA World Wind Java 2.0 2.0.0
19:17:23.030 [main] INFO  org.leo.traceroute.install.Env - /usr/java/packages/lib/amd64:/usr/lib64:/lib64:/lib:/usr/lib
19:17:23.030 [main] INFO  org.leo.traceroute.install.Env - OS:Linux / arch:amd64
Locale en_GB
19:17:25.377 [SwingWorker-pool-1-thread-1] INFO  o.leo.traceroute.core.geo.GeoService - Use geoip db /root/ovtr/GeoLiteCity.dat which is 0 day(s) old
19:17:27.110 [pool-2-thread-1] INFO  o.leo.traceroute.core.ServiceFactory - Try using device eth0 null
19:17:27.275 [pool-2-thread-2] INFO  o.leo.traceroute.core.ServiceFactory - Try using device br0 null
19:17:27.446 [pool-2-thread-3] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr_vmw null
19:17:28.509 [pool-2-thread-8] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr1 null
19:17:29.659 [pool-2-thread-9] INFO  o.leo.traceroute.core.ServiceFactory - Try using device vmnet1 null            
19:17:30.682 [pool-2-thread-10] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr2 null           
19:17:31.665 [pool-2-thread-11] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr3 null           
19:17:32.666 [pool-2-thread-12] INFO  o.leo.traceroute.core.ServiceFactory - Try using device vmnet3 null           
19:17:33.683 [pool-2-thread-13] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr4 null           
19:17:34.703 [pool-2-thread-14] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr5 null
19:17:35.688 [pool-2-thread-15] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr6 null
19:17:36.718 [pool-2-thread-17] INFO  o.leo.traceroute.core.ServiceFactory - Try using device lo null
19:17:38.806 [AWT-EventQueue-0] INFO  org.leo.traceroute.Main - Startup completed in 15790ms 

 
Obviously, OVT now detects some network devices on the system. Do not worry, you probably will not see as many network devices as I do. I have a lot of virtual machines installed and attached to a virtual Linux bridge of my test system... 🙂

And, hey, we now see the button for sniffing - and therefore know that packet capturing is active.

ovt4

And: By some clever intelligence of OVT's packet analysis we now arrive at the real hosting place of the web address we tracerouted (compare the above image with the first picture where the Linux traceroute was used). Good!

I have experienced the better and more thorough analysis of OVT's own traceroute routine in other examples, too.

Solution 2: Run OVT with "sudo"-command - but modify it!

The next challenge for me was to make the sudo-command in the standard script work. The required trick is pretty simple: as sudo itself does not invoke a shell by itself, we must explicitly force sudo to do it:

#!/bin/bash
sudo /bin/bash -c 'export DISPLAY=:0; java -Djava.awt.headless=false -Xmx512m -jar org.leo.traceroute.jar'

When we now start "ovtr.sh" as an unprivileged user, we are prepared to enter the root password. And then - what do you guess?

me@mysystem:/opt/ovt> ./ovtr.sh 
14:55:43.581 [main] INFO  org.leo.traceroute.Main - Open Visual Traceroute 1.6.3 
14:55:43.589 [main] INFO  org.leo.traceroute.install.Env - Java run-time version: 1.8.0_91
14:55:43.590 [main] INFO  org.leo.traceroute.install.Env - NASA World Wind Java 2.0 2.0.0
14:55:43.590 [main] INFO  org.leo.traceroute.install.Env - /usr/java/packages/lib/amd64:/usr/lib64:/lib64:/lib:/usr/lib
14:55:43.590 [main] INFO  org.leo.traceroute.install.Env - OS:Linux / arch:amd64
Locale en_GB
Invalid MIT-MAGIC-COOKIE-1 key14:55:43.875 [main] ERROR org.leo.traceroute.Main - Uncaught error
java.lang.NoClassDefFoundError: Could not initialize class sun.awt.X11.XToolkit
        at java.lang.Class.forName0(Native Method) ~[na:1.8.0_91]
        at java.lang.Class.forName(Class.java:264) ~[na:1.8.0_91]
        at java.awt.Toolkit$2.run(Toolkit.java:860) ~[na:1.8.0_91]
        at java.awt.Toolkit$2.run(Toolkit.java:855) ~[na:1.8.0_91]
        at java.security.AccessController.doPrivileged(Native Method) ~[na:1.8.0_91]
        at java.awt.Toolkit.getDefaultToolkit(Toolkit.java:854) ~[na:1.8.0_91]
        at java.awt.KeyboardFocusManager.initPeer(KeyboardFocusManager.java:446) ~[na:1.8.0_91]
        at java.awt.KeyboardFocusManager.<init>(KeyboardFocusManager.java:442) ~[na:1.8.0_91]
        at java.awt.DefaultKeyboardFocusManager.<init>(DefaultKeyboardFocusManager.java:65) ~[na:1.8.0_91]
        at java.awt.KeyboardFocusManager.getCurrentKeyboardFocusManager(KeyboardFocusManager.java:225) ~[na:1.8.0_91]
        at java.awt.KeyboardFocusManager.getCurrentKeyboardFocusManager(KeyboardFocusManager.java:216) ~[na:1.8.0_91]
        at javax.swing.plaf.synth.SynthLookAndFeel.initialize(SynthLookAndFeel.java:616) ~[na:1.8.0_91]
        at javax.swing.plaf.nimbus.NimbusLookAndFeel.initialize(NimbusLookAndFeel.java:105) ~[na:1.8.0_91]
        at javax.swing.UIManager.setLookAndFeel(UIManager.java:538) ~[na:1.8.0_91]
        at javax.swing.UIManager.setLookAndFeel(UIManager.java:583) ~[na:1.8.0_91]
        at org.leo.traceroute.Main.main(Main.java:75) ~[org.leo.traceroute.jar:na]
me@mysystem:/opt/ovt> 

 
What the hack!? However, this problem is simple one. We started Java from a subshell as root. But why should root get access to our X11 display from such an environment? Actually, even root cannot do this without special precautions!

In the case of starting a KDE root terminal as in "solution 1" KDE and pam take care of the X11 access. Also, in case of a "su -" command at the prompt this is handled by pam. See the "/etc/pam.d/su" file and look for a line

session optional pam_xauth.so

Let us quickly test our theory about a missing X11 access right by using the "xhost"-command in addition:

me@mysystem:/opt/ovt> xhost +SI:localuser:root
localuser:root being added to access control list
me@mysystem:/opt/ovt> ./ovtr.sh
root's password:
15:40:20.920 [main] INFO  org.leo.traceroute.Main - Open Visual Traceroute 1.6.3 
15:40:20.928 [main] INFO  org.leo.traceroute.install.Env - Java run-time version: 1.8.0_91
15:40:20.929 [main] INFO  org.leo.traceroute.install.Env - NASA World Wind Java 2.0 2.0.0
15:40:20.929 [main] INFO  org.leo.traceroute.install.Env - /usr/java/packages/lib/amd64:/usr/lib64:/lib64:/lib:/usr/lib
15:40:20.929 [main] INFO  org.leo.traceroute.install.Env - OS:Linux / arch:amd64
Locale en_GB
15:40:23.313 [SwingWorker-pool-1-thread-1] INFO  o.leo.traceroute.core.geo.GeoService - Use geoip db /root/ovtr/GeoLiteCity.dat which is 0 day(s) old
15:40:24.913 [pool-2-thread-1] INFO  o.leo.traceroute.core.ServiceFactory - Try using device eth0 null
15:40:25.090 [pool-2-thread-2] INFO  o.leo.traceroute.core.ServiceFactory - Try using device br0 null
15:40:25.279 [pool-2-thread-3] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr_vmw null
15:40:26.317 [pool-2-thread-8] INFO  o.leo.traceroute.core.ServiceFactory - Try using device vmnet1 null
15:40:27.339 [pool-2-thread-9] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr1 null
15:40:28.394 [pool-2-thread-10] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr2 null
15:40:29.417 [pool-2-thread-11] INFO  o.leo.traceroute.core.ServiceFactory - Try using device vmnet3 null
15:40:30.445 [pool-2-thread-12] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr3 null
15:40:31.469 [pool-2-thread-13] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr4 null
15:40:32.492 [pool-2-thread-14] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr5 null
15:40:33.512 [pool-2-thread-15] INFO  o.leo.traceroute.core.ServiceFactory - Try using device virbr6 null
15:40:34.534 [pool-2-thread-17] INFO  o.leo.traceroute.core.ServiceFactory - Try using device lo null
15:40:36.711 [AWT-EventQueue-0] INFO  org.leo.traceroute.Main - Startup completed in 15796ms
...
15:40:47.831 [Shutdown] INFO  org.leo.traceroute.install.Env - Preferences saved.
15:40:47.833 [Shutdown] INFO  o.leo.traceroute.ui.TraceRouteFrame - Application exited.
me@mysystem:/opt/ovt> xhost -SI:localuser:root
localuser:root being removed from access control list
me@mysystem:/opt/ovt>

 
Success! But, do not forget to retrieve the access right after having used OVT.

Solution 3: Run OVT with sudo and preserved XAUTHORITY variable

How can we automatize the X11 access a bit? A look at the simple standard "/etc/sudoers"-file reveals that it should be possible to run the "sudo"-command in our changed "ovtr.sh" also with the option "-E" (= --preserve-env):

#!/bin/bash
export PATH="$PATH:/usr/sbin/"
env | grep XAUTHO
sudo -E /bin/bash -c 'export DISPLAY=:0; env | grep XAUTHO;  java -Djava.awt.headless=false  -Xmx512m -jar org.leo.traceroute.jar'

 
And:

 
me@mysystem:/opt/ovt> ./ovtr.sh 
XAUTHORITY=/tmp/xauth-1011-_0
root's password:
SUDO_COMMAND=/bin/bash -c export DISPLAY=:0; env | grep XAUTHO;  java -Djava.awt.headless=false  -Xmx512m -jar org.leo.traceroute.jar
XAUTHORITY=/tmp/xauth-1011-_0
19:31:41.610 [main] INFO  org.leo.traceroute.Main - Open Visual Traceroute 1.6.3 
...
....
19:31:45.617 [pool-2-thread-1] INFO  o.leo.traceroute.core.ServiceFactory - Try using device eth0 null
...

 
It works ! I do not recommended the environment of a standard way to root this way, but this test just gave us the right idea. The following has the same effect - and is safer as we only use the contents of one variable (XAUTHORITY) of the unprivileged environment:

#!/bin/bash
sudo XAUTHORITY=$XAUTHORITY /bin/bash -c 'export DISPLAY=:0; java -Djava.awt.headless=false  -Xmx512m -jar org.leo.traceroute.jar'

 

Summary

My objective in this post was to present some elementary steps to get OVT running on an Opensuse Leap 42.1 system. We have seen that we need to make the right "libpcap.so" library available and that the "sudo" command in the start script has to be adapted to guarantee both the recognition of the DISPLAY variable and the access to the X11 display. In a first test we have seen that OVT's internal traceroute is a bit better than the standard Linux traceroute.

On a multi-user system, however, we need to restrict the access to OVT to a group of trustworthy users. This will be the topic of the next article.

Have some fun with testing out OVT in the meantime!

Bumblebee auf Optimus-Notebooks und Laptops mit Opensuse 13.1 / 13.2 / Leap 42.1

Da mein früherer Artikel zu diesem Thema
http://linux-blog.anracom.com/2015/06/11/bumblebee-auf-laptops-mit-opensuse-13-1-13-2/
etwas veraltet ist und zu einer nicht mehr funktionierenden Schalterfunktion für die Nvidia-Karte führt, hier eine neue Zusammenstellung der Schritte, mit denen ich Bumblebee auf Optimus-Systemen zum Laufen gebracht habe. Getestet habe ich das auf einem Laptop mit "i7-3632QM" Prozessor mit integrierter Intel HD 4000 Grafikkarte; ferner existiert eine zusätzliche Nvidia GT 645M. Ich erläutere die Installation für Opensuse 13.1; für 13.2 und Leap 42.1 geht alles ganz analog (und funktioniert auch) - nur sind andere Repositories zu wählen.

Installation und Einrichtung von Bumblebee

Folgende RPM-Repositories sollte man der SW-Verwaltung unter YaST für Opensuse 13.1 Installationen verfügbar machen:
http://download.opensuse.org/repositories/X11:/Bumblebee/openSUSE_13.1/
http://download.opensuse.org/repositories/home:/Bumblebee-Project:/Bumblebee/openSUSE_13.1/
http://download.opensuse.org/repositories/home:/Bumblebee-Project:/Bumblebee3/openSUSE_13.1/
http://download.opensuse.org/repositories/home:/Bumblebee-Project:/nVidia:/364.19/openSUSE_13.1/

Unter Opensuse 13.2 bzw. Leap 42.1 ist "13.1" in den Adressen natürlich jeweils durch "13.2" bzw. "LEAP_42.1" zu ersetzen.

Will man eine neueren Nvidia-Treiber nutzen, findet man eine Auswahl an entsprechenden Repositories unter
http://download.opensuse.org/repositories/home:/Bumblebee-Project:/nVidia:/

In meinem Fall habe ich einen Mix aus verschiedenen Versionen der notwendigen Pakete aus den oben genannten Repositories verwendet. Vorsichtigere Menschen sollten sich aber konsequent für konsistente Pakete aus dem Bumblebee3 Repository entscheiden. Das dkms-Paket nehme ich allerdings immer aus dem Repository für den Nvidia-Treiber.

Man muss halt schauen, ob es Probleme gibt. U.U. funktioniert z.B. das Schließen von Anwendungen nicht ganz richtig, etc..
Interessant ist es nach einer Installation auch immer, komplexere 3D- Anwendungen wie z.B. Alienarena mal mit

primusrun alienarena

und danach mit dem Aufruf

optirun alienarena

zu starten. Beides sollte funktionieren; tut es aber bei manchen Kombinationen der Paketversionen nicht. (Off Topic: Alienarena nimmt ungefragt Verbindungen zu einem Server im Internet auf, auch beim Single Player Modus; wer das nicht mag, kann es über Nutzen von "Host Server" im Menü unterbinden, eine Firewall nutzen oder schlicht das Netzwerk deaktivieren.)
Es bleibt einem halt leider nur auszuprobieren, welche Zusammenstellung von Paketversionen auf seinem eigenen System korrekt läuft.

Bei mir ist mit den Paketen aus dem Bumblebee3-Repo alles in Ordnung; folgende Versionen habe ich konkret installiert:
bumblebee1
bumblebee2

Entscheidend ist die Installation des Nvidia-Kernel-Modules "nvidia" über das Paket "X11-video-nvidia" aus dem 4-ten Repository. Während der Installation wird die erforderliche Kompilation für den aktuellen Kernel vorgenommen.

Blacklisten des nouveau-Treibers

Nach einer Neuinstallation von Opensuse auf einem Laptop mag es sein, dass der Nouveau-Treiber installiert ist. Ich habe Bumblebee bislang nicht mit dem Nouveau-Treiber ausprobiert, sondern aus verschiedenen Gründen immer den proprietären Nvidia-Treiber (erfolgreich) genutzt. Damit dies möglich wird, muss der Nouveau-Treiber, soweit auf dem Laptop-System vorhanden, deaktiviert werden. Ich habe ihn deshalb zur Sicherheit in eine Blacklist für Kernelmodule aufgenommen. Die Dateien "/etc/modprobe.d/50-blacklist-nouveau.conf" und auch "/etc/modprobe.d/50-blacklist.conf" enthalten somit folgende Statements:

blacklist nouveau
options nouveau modeset=0

Keine Installation des proprietären Nvidia-Treibers über das Linux-Installationsprogramm von der Nvidia-Web-Seite!

Versucht weder während der Opensuse-Erstinstallation noch später den proprietären Treiber von Nvidia mit den Installationsroutinen von der Nvidia Webseite zu installieren! Das wird in einer Optimus-Konfiguration zu keinem Erfolg führen! Der proprietäre Nvidia-Treiber sollte auf Optimus Laptops statt dessen immer aus dem oben genannten "Bumblebee-Project:/nVidia"-Repository geladen werden.

KMS nicht abschalten!

KMS (Kernel mode setting; s. http://de.wikipedia.org/wiki/Mode-Setting) wird für das "i915"-Modul (also den Treiber für die Intel Graka) zwingend benötigt - DKMS sollte man also (im Gegensatz zu früheren Desktop-Installationen mit Nvidia-Karten) nicht über Kernelparameter wie "nomodeset" beim Starten des Systems deaktivieren.

Ferner gilt: Das Opensuse Community Repository für Nvidia Treiber sollte deaktiviert sein; keines der Nvidia RPMs - sprich kein Nvidia-Treiber-Paket - aus diesem Community Repository sollte installiert sein oder werden.

Welche Services sind zu aktivieren?

Folgende Services müssen für den Systemstart unter systemd ggf. noch explizit aktiviert werden:

systemctl enable bumblebeed.service
systemctl enable dkms.service

Desktop-User als Mitglied der Gruppe "bumblebee" einrichten

Zudem muss der User, unter dem man den Desktop nutzt, Mitglied der ggf. neu anzulegenden Gruppe "bumblebee" werden.

xorg.conf-Datei

Es lohnt sich ein Blick in das Verzeichnis /etc/bumblebee": Dort befindet sich eine spezielle Datei "xorg.conf.nvidia", die von Bumblebee genutzt wird. Eine evtl. vorhandene "/etc/X11/xorg.conf" im Verzeichnis "/etc/X11" sollte man entfernen !

Neustart des Systems - KDE Desktop mit 3D Effekten über den Intel Grafikkartentreiber

Sind alle Voraussetzungen geschaffen, startet man das System am besten neu. Man sollte dann auf der gewohnten Desktop-Oberfläche landen. Diese wird vom Intel Treiber - in meinem Fall vom kernel-Modul "i915" - gesteuert. KDE-3D-Effekte lassen sich auch über die Intel-Grafikkarte nutzen; dafür reicht deren Performance allemal.

"lsmod" sollte folgende Module anzeigen, die im Zusammenhang mit Grafik von Bedeutung sind:

i915 710403 8
bbswitch 13943 0
drm 313440 11 i915,drm_kms_helper
drm_kms_helper 56806 1 i915
video 19507 1 i915
button 13952 1 i915
thermal_sys 36646 5 x86_pkg_temp_thermal,intel_powerclamp,thermal,video,processor
 
videobuf2_core 44595 1 uvcvideo
videodev 141701 2 uvcvideo,videobuf2_core
videobuf2_vmalloc 13216 1 uvcvideo
videobuf2_memops 13362 1 videobuf2_vmalloc

Start von 3D-Anwendungen

Spezielle 3d-Applikationen, wie etwa Spiele (z.B. "alienarena")oder OpenGL-Anwendungsprogramme, die mehr Rechenpower erfordern, kann man dann als Desktop-User über

optirun alienarena

oder

primusrun alienarena

starten.

bumblebee4

"primusrun" reduziert die Framerate der Graka auf die Schirmrate, wenn keine weiteren Parameter angegeben werden. Frame- und vertikale Bildschirmfrequenz werden also synchronisiert. Für volle Performance muss man

vblank_mode=0 primusrun alienarena

benutzen.

Folgender Artikel liefert ein paar Hinweise zum ursprünglichen Unterschied zwischen "primusrun" und "optirun": http://www.webupd8.org/2012/11/primus-better-performance-and-less.html
Faktisch sind die aktuellen Performance-Unterschiede zwischen "primusrun" und "optirun" auf meinem System jedoch minimal.

"lsmod" zeigt nach dem Starten einer 3d-Applikation dann folgende zusätzlichen Module an:

nvidia 8370147 37
drm 313440 11 nvidia,i915,drm_kms_helper

Aufruf des Tools nvidia-settings

Übrigens: Das Tool "nvidia-settings" ruft man wie folgt auf:

optirun -b none nvidia-settings -c :8

Ab- und An-Schalten der Nvidia-Karte im normalen Desktop-Betrieb

"optirun" und "primusrun" aktivieren die Nividia-Karte bei Bedarf selbständig und deaktivieren sie auch wieder - wenn die geladenen Pakete richtig zusammenarbeiten. Zum gezielten Abschalten der Nvidia-Karte im normalen Desktop-Betrieb benutzt man als root-User dagegen folgendes Kommando:

tee /proc/acpi/bbswitch <<< OFF

Anschalten geht über

tee /proc/acpi/bbswitch <<< ON

Der Laptop zeigt die Aktivität der Nvidia-Graka i.d.R. über eine LED an. Ein gezieltes Anschalten im normalen Desktop-Betrieb kann dann nützlich sein, wenn man - wie in meinem Fall - durch ein erratisches Anspringen des Laptop-Lüfters gerade bei zu kühlem Zustand des Laptops genervt wird. Manchmal hilft die zusätzliche Abwärme der Nvidia-Graka im Leerlauf, einen kontinuierlichen Lüfterbetrieb zu erzwingen.

Startbedingungen zum An- und Abschalten der Nvidia Graka legt man über entsprechende Parameter in der Datei "/etc/modprobe.de/50-bbswitch.conf" fest:

options bbswitch load_state=0 unload_state=1

Was ist nach einem Kernelupdate und Neustart des Systems zu tun ?

Nach einem Kernelupdate läuft der "intel915" Treiber ja typischerweise noch. Man gelangt dadurch auf die grafische Oberfläche. Am einfachsten ist dann eine Deinstallation und anschließende Neuinstallation der oben dargestellten Pakete aus den Bumblebee Repositories - u.a. dkms, dkms-nvidia, bbswitch, vor allem aber "X11-video-nvidia". Der Nvidia-Treiber wird dabei neu kompliert.