[CentOS] Understanding yum automatic upgrades

Wed Apr 6 18:26:00 UTC 2011
email builder <emailbuilder88 at yahoo.com>

> > It does  look like updates are happening, but it's not clear to me by

> >  whom.
> > do_update is set to "no", but notification is by "dbus", so I  assumed
> > that
> > "dbus" is notifying another process to do the  actual updates.  Is
> there
> > a way I
> > can track that  down?
> > 
> > > Are you sure the updates are actually getting  installed,  and it's
> > not
> > > just noise in the log from  yum-updatesd?
> > 
> > Well, if I can take it at its word, updates *are*  happening.  Here is
> a
> > snippet
> > I clipped out of a  logwatch a few months ago:
> > 
> >  --------------------- yum  Begin ------------------------
> > 
> > 
> >  Packages  Updated:
> >     php-dba - 5.1.6-27.el5_5.3.i386
> >      php - 5.1.6-27.el5_5.3.i386
> >     php-devel -  5.1.6-27.el5_5.3.i386
> >     php-cli -  5.1.6-27.el5_5.3.i386
> >     php-common -  5.1.6-27.el5_5.3.i386
> >     php-gd -  5.1.6-27.el5_5.3.i386
> >     php-pdo -  5.1.6-27.el5_5.3.i386
> >     php-mysql -  5.1.6-27.el5_5.3.i386
> > 
> >  ---------------------- yum End  -------------------------
> 
> A much more reliable way to check is
> rpm -qa  --last |less

Thanks for that.  I think this clears things up -- it looks like the updates are 
all manual ones, especially judging from how dates are grouped.

So I must apologize to you and the other responders on this thread.  The 
notifications I have seen in logwatch must have been from manual updates that I 
was unaware of or did not remember. 


> or simply run
> yum update
> and see what it thinks needs  updated yet.
> 
> If things are reasonably up-to-date I would expect the  --last list to
> have a tzdata-2011b package listed near the top.

Indeed, it's not there yet I see an update available.

> One other  thing the --last list will revel is WHEN the updates were
> applied, if they  consistently are at a particular time of the morning
> then it may be based on  a cron job.

The date/times appear to indicate manual interaction from a human.  In any case, 
my question seems to be answered.  I will watch logwatch a little longer and 
make sure this is the case.

Thanks to everyone and sorry for the confusion.