Pointy's questions

I've been having a problem that I can't explain with my limited understanding of "how things work". I connect to various servers (I'll use svn.mycompany.com as the example) for work. We use openvpn to create a secure connection generally to those servers. The company has OpenDNS set up to provide ordinary resolution for the servers, and the resulting addresses of course only work via an established VPN tunnel.

Well after some innocuous reconfiguration of some of the servers (not done by me, and apparently working for everybody else) I notice the following pattern. I have a VPN tunnel, and I try a svn command:

svn update

I get an error back immediately that the host svn.mycompany.com can't be resolved. If I then do a host lookup:

host -a svn.mycompany.com

that responds with the correct IP address. If I then re-try the svn command, it works, and svn keeps working, for a while. After some unmeasured period of time however, it stops working again and the cycle repeats.

The same pattern holds for other servers on the other side of the tunnel. I've seen this happen from different networks (i.e., at my house, out at a coffee shop, etc).

I'm not looking for an overall solution. My real question is, how is it that simply running host -a can at least temporarily "fix" the situation of a domain not resolving? Does host do something special to bypass a local cache? (If so, I'm still confused, because the address of the servers don't change, or change rarely.)

edit — OK more information. By turning up logging for systemd-resolved, I was able to use journalctl to track what my local machine is doing with DNS lookups. What I saw seems interesting but I still don't know enough to understand what it means: DNS lookup requests of query type ANY seem to overflow the UDP packet size, so systemd-resolved falls back to making a TCP query. For a normal non-ANY lookup on these foo.mycompany.com names, I don't get the packet overflow but it goes on to make a NODATA local cache entry.

When the ANY queries force the TCP fallback, systemd-resolved gets a useful result and makes a positive cache entry.

To me this means that something weird is going on with the UDP responses, but I don't know what that implies about the root cause.

I've got an external USB drive with an encrypted LUKS partition. I generally have no problems mounting the device; I click it from the xfce4 desktop and get a dialog where I enter the password.

When I try to mount the drive now, I type in the password and get the error:

Error unlocking /dev/sdc1: Command-line `cryptsetup luksOpen "/dev/sdc1" "luks-..."' exited with non-zero exit status 5: Device luks-... already exists

What do I need to do in order to mount the drive? Is it somehow corrupted? How can the device already exist if it's not mounted (and wasn't plugged in until a short time ago)?

I've just upgraded to 15.10 (and yes I plan to go to 16.04 next time I have a little free time). I've been using openvpn to connect to a work VPN for years and years via a .ovpn config file that's always worked.

Now, however, something having to do with systemd seems to have changed the way things work. When I try to start it, I get a message broadcast out via wall:

Broadcast message from root@turandot (Sun 2016-05-01 10:25:50 CDT):

Password entry required for 'Enter Auth Username:' (PID 17284). Please enter password with the systemd-tty-ask-password-agent tool!

I've googled around for it and found nothing that seems like a solution to the problem. What's the new way to open a VPN connection?

There is information on the Internet (here for example) about using the PPA ppa:moorkai/cinnamon to get Cinnamon 2.6 installed on Ubuntu 15.04. I'm noticing that that the install fails, as it requires libcjs0 at least 2.0.0.

If I install that package manually, attempting to install Cinnamon then results in a failure about the cjs package. The setup of those packages won't let me have libcjs0 and cjs packages installed at the same time; each insists on the removal of the other.

I could force the install of one of them, but that seems undesirable given that I really don't know what the packages do or what they affect.

The install from ppa:gwendal-lebihan-dev/cinnamon-nightly does work. (Well, the installation works, but the desktop environment is completely unusable; something's obviously wrong because it's just broken.

I'm running an up-to-date 14.04 installation. In a Java project built with Ant, I'm using the <svn> task to grab the current revision information. I've got (as far as I can tell) up-to-date versions of the "svnClientAdapter.jar" and "svnant.jar" libraries, and I've got the libsvn-java library installed for the JNI version of the adapter.

After doing the 13.10 -> 14.04 upgrade, I was told by svn that I needed to run svn update because the version of the runtime was ahead of the stuff in my checked-out tree. After doing that, I get errors from the svn ant task:

  [svn] <WcVersion> started ...
  [svn] svn: E155021: This client is too old to work with the working copy at

So what's up? Judging from various old questions it seems like this isn't a new phenomenon.

I'm running 13.10 (all up-to-date) on a Lenovo laptop with a Griffin iMic USB audio device. Though I occasionally have to use the Pulse Audio volume control tool to re-select the iMic, it works consistently from everything on the system except Chrome, which basically has no audio at all.

With most applications that want to use the audio stuff, you see something in the first tab of pavucontrol. Not Chrome; there's just nothing.

Some ancient forum posts here and there suggested symlinking the Firefox "plugins" directory over to Chrome's installation directory, which seems pretty goofy and which doesn't work now anyway.

Chrome version is 34.0.1847.132.

Is there some trick to making Chrome work with a USB audio device? (As far as I can tell it doesn't work with built-in audio either ...)

edit — Still not working, now on 14.04 and Chrome 37.0.2062.120

More info:

Chrome's been reinstalled more than once, with no effect. I've also tried the beta (currently Chrome 38.0.2125.77 beta). The PulseAudio manager tool, in its list of clients, shows Firefox and various other things, and also "Chrome input" but no "Chrome output".

Chromium behaves exactly the same way.

edit — now on an (old and tired) 15.04 installation. Chrome (Version 49.0.2623.112 (64-bit)) still does not work, though on full moon nights or something else random it'll send sound through the built-in analog audio on the laptop. However, Chromium (Version 48.0.2564.82 Ubuntu 15.04 (64-bit)) does work now, and it works through the USB audio device. I don't know of any particular thing I've done lately to make that true, but

I've got a new laptop on which I've installed 13.10 and I use XFCE as my desktop. I want to use the laptop with an external monitor, and that works fine, mostly. However, when I use the "Display" settings control to select the screen to use, the blue box that shows up to tell me that my external monitor is "HDMI1" does not go away after I've set things up. Here's what it looks like:

It won't go away. I've dismissed the "Display" settings application, but the thing is still there. What is going on?

Overall the XFCE display chooser seems to work much, much worse in 13.10 than in 13.04. It's a crapshoot whether it can re-enable the laptop display or not. I can do it reliably with xrandr, so I don't think it's a low-level problem.

edit If I turn the external monitor off with xrandr and then back on, the stupid thing comes back.

Just upgraded a laptop to 13.04. Via Brasero, I can burn CDs and DVDs, but when I try cdrecord from the command line I get:

$ cdrecord --devices
wodim: No such file or directory. 
Cannot open SCSI driver!
For possible targets try 'wodim --devices' or 'wodim -scanbus'.
For possible transport specifiers try 'wodim dev=help'.
For IDE/ATAPI devices configuration, see the file README.ATAPI.setup from
the wodim documentation.

I get that same error regardless of the parameters with which I invoke the command. Is there some sort of package dependency issue?

edit — running wodim with strace shows that it tries to open all the drives /dev/hda through /dev/hdz, then everything from /dev/scd0 through /dev/scd255, and then it quits. It's probably a configuration thing.

(12.10, both KDE and xfce4)

For some reason, Firefox has stopped working when launched from the window manager's launcher. The program runs, but nothing happens. If I try again, I get an error popup that tells me another instance of Firefox is already running, which is true according to ps.

If I strace the process, it's stuck on a FUTEX_WAIT_PRIVATE system call. Does anybody know how to make that stop happening? (I found an old blog post that claimed it has something to do with assistive technologies being enabled, but on my system they're not as far as I can tell.)

Oh, and I can start Firefox successfully by launching it from an xterm window.

I'm trying to make a bootable installation USB drive. I've tried usb-creator-gtk, and it starts working only to fail with a segmentation fault, leaving a non-bootable drive. I've tried "unetbootin", and it makes a bootable drive that cannot even begin an installation successfully. (I can't tell what it's problem is, but the result is utterly unworkable; the installation process just hangs.)

This seems like something that should be pretty simple, and pretty easy. What am I doing wrong?

My 12.10 installation from which I'm trying to run usb-creator-gtk is up-to-date, and other than this it works fine. The USB stick is fine (and I've tried this with several different USB sticks with the same result).

% uname -a
Linux phoebe 3.2.0-33-generic #52-Ubuntu SMP Thu Oct 18 16:29:15 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

I'm running 12.04 on the machine I'm typing this on, and I've been running it for quite some time. I just did a dist-upgrade, and it included some update for Firefox (to 17.0.1 now). After restarting the browser, I noted with some surprise that I can't type back-quotes or tildes. They worked just a few minutes ago before the update, but not now.

I use a Deck keyboad, and I have the back-quote/tilde key remapped. I've had the keyboard for many years now and the remapping has always worked:

% xmodmap -pk | egrep tilde
133         0x0060 (grave)  0x007e (asciitilde)     0x0060 (grave)  0x007e (asciitilde)

I can type those characters into xterm windows, into Chrome, gvim, gedit, etc. What in the world is going on with Firefox? I can't find any relevant configuration options, and I certainly haven't been tinkering with them. I've tried restarting it a couple times too.

I'll try running a new clean profile and see if that makes a difference; I do have some add-ons but nothing new and nothing particularly crazy. edit Nope, can't type the characters in a plain-jane profile either.

edit — reinstalled Firefox, same thing.

edit again — I've also found that VirtualBox seems to ignore the xmodmap remapping. I've basically got the grave/tilde key swapped with another key that's remapped to "Alt". In VirtualBox now (I don't think it was always this way, but I'm not 100% sure), I can type back-quote or tilde with the key that's swapped to "Alt". However, that does not work in Firefox — neither key works as the grave/tilde key, though I've just verified that the remapped "Alt" key in Firefox does work (as "Alt")!!

edit yet again — upgraded to 12.10, same problem. Something changed somewhere, as I am 100% positive that that key was working today and over the past many months. (Why? How do you quote code snippets inline in Markdown? I'm on Stackoverflow a lot and that's how I noticed it was broken.)

I'm trying to implement the fixes from Adobe (as mentioned in this other question) but I'm having no success.

I'm running 64-bit 12.04. When I try the Flash test as described on this Adobe guide:

1. First, launch the Adobe demo video player container: http://drmtest2.adobe.com:8080/SVP/SampleVideoPlayer_FP.html
   
   Note the Adobe Flash Player version in the lower left of the demo video shell. Make sure that indicate 11.2 or higher

2. Play sample protected content.
   
   Enter the following URL in the field "Input the video URL: (case-sensitive) http://drmtest2.adobe.com:8080/Content/anonymous.f4v

3. Press the [Play] button at the bottom of the video container.

Well that doesn't work. From that test page, when I select the "Show DRM Events" checkbox I get "Error #3344 [MissingAdobeCPModule]", and then it just sits there and nothing plays.

I've installed the HAL (several times) and cleared out the directories from my home .adobe directory as those instructions say, restarted Firefox, etc.

The Flash info page says "You have version 11,2,202,160 installed" and that 236 is the more up-to-date version, but I can't figure out how to install that. (I've tried, but to no avail; I just end up with 160).

So does anybody have ideas on what to do to make this work?

I've got a Firefox process with nothing in it but the Google home page (the "search" page, not the "homey" home page with news etc.) and this site. I've killed it and restarted it multiple times. It's burning massive amounts of CPU cycles doing ... what?

I notice that Firefox, over time, leaks memory terribly and uses more and more CPU, but not anything close to this much. What is it doing? At present it's the primary process loading down the system, and that's new behavior. Eating memory and getting bloated is usually something that makes it sluggish. Now it's just hogging resources and spinning my fan.

I'm running 12.04 on an HP laptop but that shouldn't matter of course.

Are there some Firefox tools for looking at its internal thread activity or something?

edit newly-created profiles show exactly the same behavior ...

I've got a new installation and I'm trying to get Postgresql working.

The database was installed and I started a restoration from a dump on another machine, but that ran into problems because I had forgotten to install the "contrib" package. I used "pg_dropcluster" to drop the old cluster so I could start from scratch, and that's when things went weird.

The first manifestation of this was that /etc/postgresql was just empty. I uninstalled the postgresql package and reinstalled, several times, to no avail.

Is there something I can do to figure out why apt is confused here? I've done this many times on many systems and never seen anything like this happen.

Back in the day, there was grip. It had a complicated, somewhat ugly UI, but everything was configurable and, generally, it worked. Then it was abandoned.

Now, there are all sorts of tools, but finding one that has reasonable configurability (please let me provide the command line for lame ... that's all I ask!) has been a multi-year frustration. I guess I understand the desire to keep things simple, but surely somebody else out there wants to use "--preset extreme". Grip also had a neat system for handling special characters in track names (ripped any Sigur Rós or múm albums?) that I've not seen in other tools.

Here are the ones I've tried:

• abcde — painful "ui", irritating and tricky to deal with when CDDB has no data, but it is configurable and overall it's the best of the lot (amazingly)
• asunder — cannot configure lame options, doesn't act unix-y w.r.t. configuration options (eg paths with leading slashes aren't considered absolute)
• ripperx — more configurable, in that it allows "extra" options to lame, confusing UI, documentation way out of date
• sound-juicer — Ubuntu docs warn people away due to some gstreamer problems
• rubyripper — tried this and hated it for reasons I can't recall now, and it's not in the repositories anyway
• banshee — inadequate configuation options, wretched iTunes-like UI (I'm probably a minority in that view), sluggish and bloated

I haven't tried the KDE tools.

Is my basic problem that I'm living in the past here? Is there some awesome ripper/encoder in the repository that cleverly evades searches via synaptic?

I'd like to set up a bridge between a wireless subnet and an Ethernet subnet (actually just one client, a DVD player with absurdly poor wireless). That is, the wireless network is the "real" network, to which clients get to the router and gateway, and I want the wired subnet to transparently get through the bridge machine to the wireless subnet. (I'm not a networking person so I don't know whether I'm saying that right.) There are two packages I can find that I think are relevant:

• bridge-utils
• parprouted

I'm not sure which of those I need, or if I need both. (I guess I can just install them and see if it's obvious.)

Is this an easy thing to set up? It just occurred to me today that it must be possible, but I haven't found any (recent, at least) how-to's on the subject.

edit — oh durr I just found this reference which tells me a lot about the bridge-utils stuff (and the deal with the AP not wanting to handle alien-looking packets makes sense). So maybe that second thing is what I want.

I'm getting a new laptop running with 11.04 Ubuntu. I've got the xfce4 packages all installed, which is something I've had no problems with on any of my other machines.

On this new laptop, however, though I can log in and use an xfce session without any problems, logging out of a session is problematic:

• I click the "Log out" widget from the panel and then "Log out" from its option dialog. Then the thing just sits there, not logging out. Subsequent attempts to open the "Log out" widget fail with an error about the session manager being busy. After maybe a minute or so, it logs out.

• Though I've got the "Save session" option checked in the log out dialog, xfce just makes a complete hash of the business. It does remember the applications that I had running, but it seems to forget about the window manager (!!) and the workspace configuration.

I don't log in/out that often, and generally I don't care much about restarting applications, but the window manager being missing is of course pretty annoying.

I like xfce because it's simple and unobtrusive and usually works pretty well. I've never experienced this, and I've got two other machines also running 11.04 with pretty much the same setup (straight Ubuntu install with xfce4 packages added). Is there some good way to diagnose stuff like that?

Linux sound is generally very confusing to me, especially when two different machines running the same version of Ubuntu don't behave at all the same way.

I'm trying to use mencoder to record a video stream to an AVI file. From one of my laptops (an HP dv7) it works exactly as some random Ubuntu wiki post says it should with this command:

mencoder tv:// -tv driver=v4l2:width=640:height=480:device=/dev/video0:forceaudio:adevice=/dev/dsp1 -ovc lavc -oac mp3lame -lameopts cbr:br=64:mode=3 -o webcam.avi

On another laptop (a Lenovo Ideapad U460), that fails because there's no /dev/dsp1. There's no /dev/dsp or /dev/audio anything. There is a /dev/snd directory with various cryptically-named files in it, but I have no idea which is which.

The webcam on the lenovo does work; I've tried it with Cheese. I would however like to at least understand why the systems would be different. (I have not done a side-by-side installed package comparison, I confess.)

Any suggestions on packages to install or diagnostics to perform are welcome. Thanks.

I've looked around for answers to this, but nothing I've found has been relevant, which seems a little weird but there you go. All I'd like to do is drive a second monitor via HDMI from a laptop. From Windows 7 on this laptop that works fine, of course; it's actually surprisingly nice. Thus I know the hardware is OK.

From another laptop I've got, one with an NVidia card, the NVidia settings app seems to do the trick (though it's clunky). However, this laptop I'm using has an Intel GPU, and I can't find any information about how to get it to recognize that a second display is available. Is it even possible to do this dynamically, or do I need to statically alter my "xorg.conf" to make the second screen available?

edit — sorry I should have noted that I run xfce.

I'm running Lucid on a few laptops, with xfce as the desktop environment.

Before Lucid, the little NetworkManager tray widget would automatically scan for access points for me. Now, however, I find that I have to go run iwlist myself to look for SSIDs (if I don't know one already) and manually add that to the list. Once I've done that, then it'll automatically connect (if the AP is configured that way).

Part of what makes this a real pain is that the tray widget (sometimes) seems to not show itself at all when there's no current connection and no known access points. I have to kill NetworkManager (which is a pain, because init really wants to re-launch it), run iwlist to find an SSID, go to the (now showing) tray widget and add the SSID, and then re-start NetworkManager.

I realize that the widget is just messing with a config file or two, and I'm happy to fool with those manually, but what I'd really like is for it to work the way it used to and just do the scan for me. Note that once I've got a connection up, then it shows me all the APs it can see, as before.

I acknowledge that it's possible I screwed up some setup file, but I've really tried scrubbing everything back to "as new" state and that hasn't helped. Am I crazy or did that stuff really change?