Recovering from X "upgrade" disaster - now what?

Warren Block wblock at
Thu Apr 7 15:21:55 UTC 2016

On Thu, 7 Apr 2016, William A. Mahaffey III wrote:

> On 04/06/16 22:58, Warren Block wrote:
>> On Wed, 6 Apr 2016, Will Parsons wrote:
>>> Thank you for the suggestions, even if they failed to solve the
>>> problem.
>>> Happily, I have now got X working again, but to do so I had to restore
>>> the entire contents of /usr/local from a backup I had made about 6
>>> months ago.  I think what this means is that my pkg database is out of
>>> sync with the actual contents of /usr/local.  What does this mean as
>>> far as further updating?  I'd just as soon try to resolve the real
>>> problem, but am understandably reluctant to leave my system crippled
>>> again.
>> First, make a full backup of the system as it is.
>> The easy way to fix the package database is the same thing that might 
>> resolve the problem.  Reinstall all the packages, either as binaries or 
>> from ports.  There is a procedure for ports at the end of the portmaster 
>> man page.  For packages, I want to say it would be 'pkg install -af', but I 
>> have not tested it.

> If he blindly reinstalls *all* pkg's, he will replicate the problem he just 
> got done fixing. Might want to lock X11, maybe other stuff as well, before 
> restoring pkg's .... $0.02, no more, no less ....

No, locking packages is more likely a cause than a cure.  The problem 
now is that there are files and applications that do not match what the 
package database thinks there are.  Rebuilding everything makes sure 
that it is all consistent, and rebuilds the package database at the same 

The backup is there in case this is a problem with X rather than a 
problem with inconsistent packages.

More information about the freebsd-questions mailing list