Dear All, I have several CentOS 7 machines (but not all of them) on which I have noticed that something gets written to /root/.cache/mesa every so often (once every several days). It is my understanding that mesa is related to GUI (X11), but on these machines (on neither of my machines, actually) root never logs in to GUI X11. I may have remote root logins with Xforwarding though. Hence my puzzle: what is that that writes to /root/.cache/mesa? Some, but not all of occasions seem to happen upon machine [re]boots. Thanks. Valeri PS This gives me dejavu. A while ago when people started demanding to have google chrome browser installed on their workstations I had hard time to get rid of google's cron jobs that were writing where only root should - without explicit permission to do so. Dough. Somebody's software thinks it is smarter than everyone who uses it... "Machine learning" all the way ;-) -- ++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++
On Wed, 20 Feb 2019 at 09:23, Valeri Galtsev <galtsev at kicp.uchicago.edu> wrote:> > Dear All, > > I have several CentOS 7 machines (but not all of them) on which I have > noticed that something gets written to > > /root/.cache/mesaI found that I had similar files but none of them had been touched since 2016. The data looks to be with 3d shaders and the only regular X item I can think of is gdm but there may be others depending on the system. What kind of video cards does the system have?> every so often (once every several days). It is my understanding that > mesa is related to GUI (X11), but on these machines (on neither of my > machines, actually) root never logs in to GUI X11. I may have remote > root logins with Xforwarding though. Hence my puzzle: what is that that > writes to /root/.cache/mesa? Some, but not all of occasions seem to > happen upon machine [re]boots. > > Thanks. > Valeri > > PS This gives me dejavu. A while ago when people started demanding to > have google chrome browser installed on their workstations I had hard > time to get rid of google's cron jobs that were writing where only root > should - without explicit permission to do so. Dough. Somebody's > software thinks it is smarter than everyone who uses it... "Machine > learning" all the way ;-) > > -- > ++++++++++++++++++++++++++++++++++++++++ > Valeri Galtsev > Sr System Administrator > Department of Astronomy and Astrophysics > Kavli Institute for Cosmological Physics > University of Chicago > Phone: 773-702-4247 > ++++++++++++++++++++++++++++++++++++++++ > _______________________________________________ > CentOS mailing list > CentOS at centos.org > https://lists.centos.org/mailman/listinfo/centos-- Stephen J Smoogen.
On Wed, Feb 20, 2019 at 08:22:39AM -0600, Valeri Galtsev wrote:> PS This gives me dejavu. A while ago when people started demanding to have > google chrome browser installed on their workstations I had hard time to get > rid of google's cron jobs that were writing where only root should - without > explicit permission to do so. Dough. Somebody's software thinks it is > smarter than everyone who uses it... "Machine learning" all the way ;-)Since I work at a university that uses Google Apps, we're asked to provide Google Chrome for all our users, but since the package is not very enterprise ready, I have to make several adjustments. One thing I do is put an 'exit 0' in /etc/default/google-chrome. This effectively stops most of the evil that the RPM does in its postscripts. (As a reminder, this is what the latest RPM does, https://paste.fedoraproject.org/paste/MS~7Fkr5AWYo7SAWAl8t6A ) I also manage private repos (with pulp) of the Google Chrome repos, in case I need to go back to a previous version, so having it overwrite my repositories is actually damaging. I also disable 'at' and the atd service on our workstations, and this RPM turns atd back on and schedules an at job to run the /etc/cron.daily/google-chrome script. It's an absurd RPM and Google should be ashamed of it. -- Jonathan Billings <billings at negate.org>
On 2/21/19 9:30 AM, Jonathan Billings wrote:> On Wed, Feb 20, 2019 at 08:22:39AM -0600, Valeri Galtsev wrote: >> PS This gives me dejavu. A while ago when people started demanding to have >> google chrome browser installed on their workstations I had hard time to get >> rid of google's cron jobs that were writing where only root should - without >> explicit permission to do so. Dough. Somebody's software thinks it is >> smarter than everyone who uses it... "Machine learning" all the way ;-) > > Since I work at a university that uses Google Apps, we're asked to > provide Google Chrome for all our users, but since the package is not > very enterprise ready, I have to make several adjustments. > > One thing I do is put an 'exit 0' in /etc/default/google-chrome. This > effectively stops most of the evil that the RPM does in its > postscripts. (As a reminder, this is what the latest RPM does, > https://paste.fedoraproject.org/paste/MS~7Fkr5AWYo7SAWAl8t6A ) > > I also manage private repos (with pulp) of the Google Chrome repos, in > case I need to go back to a previous version, so having it overwrite > my repositories is actually damaging. > > I also disable 'at' and the atd service on our workstations, and this > RPM turns atd back on and schedules an at job to run the > /etc/cron.daily/google-chrome script. > > It's an absurd RPM and Google should be ashamed of it. >Thanks for your input, very instructive! Google is not ashamed of what it is doing. As one clever man said over decade ago: you don't need to recruit spies anymore, just roll out "free" services. And we all know, these "intelligence" agencies were never ashamed of whatever they have done. But as another clever man said: the people do deserve the government they have. Valeri -- ++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++