[mythtv-users] Running MythTV on CentOS

[Gary Buhrmaster]

On Mon, Oct 6, 2014 at 3:32 PM, Ben Kamen <bkamen at benjammin.net> wrote:
...
> I'm also one that doesn't like having to constantly update the OS install
> because of EOLs on a previous version.

<background>
RHEL (most the rest are just derivative works) has
the advantage of long term stability and support.
But, few newer features that are not demanded
in the enterprise server space, and few new versions
of applications libraries (fixes are backported only
when necessary).

That is fine as long as you do not need or
intend to upgrade the applications.  However,
applications such as MythTV move too quickly
to defer upgrades for long periods.  There is
no staff to backport important fixes to old
MythTV releases.  And many want new the
features.  Sometimes, that long term stability of
the OS and desire to upgrade apps conflicts.

The case in point was that MythTV 0.27 changed to
require Qt 4.8. RHEL 6 was fixed on Qt 4.6.  Some
complexity was encountered (yes, one can run
parallel Qt's, but it is not at all easy for most,
although if one can borrow someone else's
competence it is easier).  And the "preview" Qt 5
available in RHEL was not adequate in all cases
due to missing Qt 5 functionality in MythTV.

Rock, meet hard place.
</background>

RHEL 7 "fixes" this (Qt4.8) for this round.  And
it is the minimum release I would consider.  But
over the 10 year lifetime of RHEL 7, some next
new feature/requirement will likely happen.  I
would expect that while one may be able to
defer upgrades a bit, sooner or later RHEL 8
will be the minimum one will need.

As with much else, one has to make trade-offs
that one is happy with.

<opinion>
I certainly understand wanting to setup and
ignore an appliance.  But I do not consider
MythTV an appliance.  I consider it a hobby.
If I really insisted on an appliance I would
probably rent an X2 STB.

Personally, with the recent improvements in
Fedora upgrade process (and, yes, previously
things were much worse), I can usually upgrade
a Fedora frontend in less than 15 minutes of effort
(mostly prep to deal with results from rpmconf
(that I should have dealt with previously) and to
insure I have adequate backups "just in case"),
although the elapsed time is quite a bit longer
due to the time to download and actually apply
the updates.  I personally find spending 15
minutes every six months or so to not being
onerous (well, I have to do it across 8 real systems,
and another half dozen virtual ones, but after the
first, it is more like 5 minutes per system)
especially since I also run my desktop on
Fedora since I do do development, where I want
(sometimes need) the latest development
libraries and functionality.
</opinion>

As with much else, one has to make trade-offs
that one is happy with.