chromium failed to launch "Check failed: NamespaceUtils::WriteToIdMapFile("/proc/self/gid_map", gid_

Hi,
just installed latest kernel (3.16.7-32-desktop) and now cannot launch (either direct from desktop or from within a terminal) the Chromium web browser (version 48.0.2564.82-67.1-x86_64). I now get error message (when trying to launch from with a terminal) -

Check failed: NamespaceUtils::WriteToIdMapFile("/proc/self/gid_map", gid_)


Have checked that Chromium still works with previous kernel (3.16.7-29-desktop) and it does.

Has anybody else experienced this? And could it be a recurrence of this :

https://code.google.com/p/chromium/issues/detail?id=480017

cheers

I now see that a bug has been raised for this - Bug 965308

I encountered the same problem. And the same error message in the terminal.

I have the same problem on both chromium and chrome. And what is the link to the above mentioned bug report?

A little searching on google provided a fix for chrome and chromium not loading after kernel update.

Start chrome as:
/usr/bin/google-chrome-stable %U --disable-namespace-sandbox

and start chromium as:
chromium --disable-namespace-sandbox

This has worked for me to allow loading for these programs. I did not follow logic of what the reasoning behind this and whether this will be fixed later on but above works for me.
32bit, opensuse 13.2, mate desktop.

Tom Kosvic

Excellent solution. Worked for me.

Worked here too - thanks

Yes, it works for me. Here is some additional info that was generated on my terminal:

[2362:2392:0206/104203:ERROR:logging.h(808)] Failed to call method: org.kde.KWallet.isEnabled: object_path= /modules/kwalletd: org.freedesktop.DBus.Error.ServiceUnknown: The name org.kde.kwalletd was not provided by any .service files[2362:2392:0206/104203:ERROR:native_backend_kwallet_x.cc(411)] Error contacting kwalletd (isEnabled)
[2362:2392:0206/104204:ERROR:logging.h(808)] Failed to call method: org.kde.KWallet.isEnabled: object_path= /modules/kwalletd: org.freedesktop.DBus.Error.NoReply: Message did not receive a reply (timeout by message bus)
[2362:2392:0206/104204:ERROR:native_backend_kwallet_x.cc(411)] Error contacting kwalletd (isEnabled)
[WARNING:flash/platform/pepper/pep_module.cpp(63)] SANDBOXED

Does this help to clarify the origin of the problem?

Same here with “google-chrome-stable” and openSUSE 13.2 x86_64 using “kernel-desktop-3.16.7-32.1.x86_64”.

With the previous kernel it worked fine.

Running “google-chrome-stable --disable-namespace-sandbox” works for me as well.

I’m not on a multi-user system so this is fine in the interim to a fix.

Cheers all,

Tele2win

So what are the real security implications of using

--disable-namespace-sandbox

Is it better to revert the kernel update and wait for a replacement? (I myself tend to be more concerned about web/network born exploits than potential local kernel exploits for example)

Bugreport here:
https://bugzilla.opensuse.org/show_bug.cgi?id=965356

Maybe someone can use the new Kernel and leave a message?

Several people had already added to the bug report and a fix is on the way. I added the additional output as reported below.

Virtual Box also refuses to load or update (“Operation aborted by user”). Haven’t seen that for a while. I rebuilt the db but it didn’t help. Frantic (I believe the superbowl is available on W10) (vbox guest in my case).

pacolaser

Chromium busted for me too, however Virtual Box is working fine for me but I get the RPMs from the VB website and compile the modules myself and not the Opensuse Repos as I’ve too many problems with them.

I noticed some update activity last night but no change to Chrome browser situation. Did some work on vbox but it can’t read my W10 vdi. I also use the VB downloads. Will revert to 5.0.10.

(temp fix, I hope) Works for me too, running Google-Chrome-beta on opensuse 13.2 after 2015.02.08 kernel update.

terminal: # /usr/bin/google-chrome-beta %U --disable-namespace-sandbox

thanks to poster: **tckosvic
**
p.s I also downgraded google-chrome, google-chrome-beta to earlier versions and still had same ‘non-start’ problem. so, not a google issue.
p.p.s - google-chrome process actually starts, just doesn’t launch UI and process has to be killed

Thanks Landis. I am already using that work around (see earlier posts). Very grateful to tckosvic . . . . Hope they produce a patch soon.

taking too long to patch a known major flaw in kernel update…

I down graded my kernel to 3.16.7-29.1 (previous version, before 2016.02.07 update).
having to change all my kickoff links, chrome-apps, etc to reflect --disable-namespace option or having a terminal window open for each chrome instance is a bigger pain than downgrade.

google-chrome-beta works, again.

the only issue I saw that the 02.07 update helped with, was, on exit (shutdown) my wicked network deamon would hang (waiting 1 min 39 sec), but with updates this morning, wicked was patched too.

I had a similar experience. What happened was that KWin crashed upon rebooting after installing the kernel update. This apparently shuts down OpenGL acceleration as a precaution. With that off, VirtualBox didn’t want to start any VMs. I didn’t realize this had happened at first, and my solution involved rolling back the update (thanks BtrFS!!!), removing FGLRX graphics drivers, installing the update, and then reinstalling FGLRX. You can go into the Desktop Effects section of the System Settings and see if OpenGL is disabled. If it is, you might just need to check the box to turn it back on to solve the problem. It’s also possible your situation is caused by something else though. Did you try running VirtualBox from the command line and see if it gives you any helpful messages?

Tip…
I found that by default the system seems to keep 'multiple kernel’s
Using the /etc/zypp/zypp.conf file you can ‘order’ the use of these kernels.

Example (on or about line 435):

## Default: Do not delete any kernels if multiversion = provides:multiversion(kernel) is set
## 2016.02.08 - added 3.16.7-29.1 because an issue with 'latest' (3.16.7-32.1) was keeping google-chrome from starting it's UI - Landis
multiversion.kernels = 3.16.7-29.1,latest,latest-1,running

I just added the previous kernel version to the beginning of the ‘multiversion.kernels =’ list

later, instead of reinstalling, hoping the kernel update doesn’t fail, I can just delete or comment the line / version or swap ‘latest,’ with version number entries.

thought this might help someone else.

**NOPE!!! **
Does NOT work.
I wondered why kernel use order would be in a Zypper config file.
I seems it Only tells zypper (yast2) what versions to keep.

So, I, instead used GRUB2 at boot to choose my next older kernel…
(grub boot screen > options > choose a different kernel > enter)

Works just fine. For now.

Landis.