[mythtv-users] Samba share of mythfs.py filesystem not working

[E. Westbrook]

On Fri, Jan 28, 2011 at 20:24, Raymond Wagner <raymond at wagnerrp.com> wrote:

> To be honest, I don't know how that would work.  When the system event is
> sent to the connected machines, only the chanid and starttime is defined.
> Normally, the machine can then look up the recording to pass on to the
> registered command, but if the recording has been deleted, there will be
> nothing to look up.
>

Ah.  I didn't realize the db record was actually deleted.  Maybe it
shouldn't be.

This could be easily accomplished if, instead, the record was retained, but
simply marked as deleted (or even moved to another table, if you like, and
perhaps with fewer fields).

This would permit mythlink to use recording creation and recording deletion
events to handle everything "as it happens" on a specific file basis every
time, with surgical precision and efficiency.

If you like, of course, maybe add a full cleanup run every week or so to
catch leftovers from unexpected exceptions or unstable systems.  Maybe also
go ahead and delete old records after a specified time, if limits to
retention (i.e. table size) are desired.  (Although, myself, I might want to
keep them a good long time, if it's not too costly, for the statistics they
might yield.)

The easiest option would probably be to just have a 'cleanup' mode, where it
> scans the directory, and removes any dead symlinks.
>

Easiest, true, but maybe not the most resource-efficient nor timely.

On Sat, Jan 29, 2011 at 12:46, Michael T. Dean <mtdean at thirdcontact.com>wrote:

> When run using MythTV System Events or User Jobs, it's very likely that
> multiple mythlink.pl processes will be running concurrently, since many
> recordings start or end at the same time.


Indeed, and why anything run as such a process must only do the minimum
necessary and get out of the way fast.  Under this plan, even in total, that
should still cost much less than a full run (due to the specificity), and
even more so over time since no unnecessary full runs at all need occur.

Besides, we have a working solution.  IMHO, we should concentrate efforts on
> all the things that don't work or don't have solutions.
>

No intention to distract anyone!  As for myself, though, I don't really have
anything else "pressing" on my plate right now, and like I said, it's been
rolling around in the back of my head for a while now.  So maybe I'll play
with it locally and report back if anything comes of it.

 $0.02,
Eric
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mythtv.org/pipermail/mythtv-users/attachments/20110129/cd591e98/attachment.htm>