Re: [mh] Trying out the updated insteon branch -> scan all and orphan deletes still not quite there.

[Gregg L. <gr...@li...>]

On 6/22/2011 3:43 PM, Marc MERLIN wrote:

>>> Summary of showstoppers:
>>> A) scan link table with PLM is not working but it works for other devices
>
> Note, by that I did write "scan link table" (not tables). In other words, I
> can scan devices, but I can't scan the PLM's own table.

Thanks for clarifying.  Now, I do see it happily returning 
026915--meaning the PLM's database is empty.  I guess that makes sense 
if this is a brand new PLM--right?  So, what would be better is 
appropriate feedback telling you that.  This should be a relatively 
quick fix.  But, shouldn't really impact any of the other issues.

>> There are really two significant issues here: 1) I need to better
>> understand the specific cause of why your scans are stopping and 2) I
>> need to implement date-time and health codes for the ALDB objects.  I
>> was really surprised by your logs as I was expecting to see other info
>> like timeouts.  Instead, they "just stopped".  Not to diminish the
>> importance of having full scanning working (which of course it does for
>> me), I think it's more important to focus on #2.  The reason is that I
>> would also provide a diagnostic utility that would list all the devices
>> and the status and time of last successful scan.  That way, you would
>> know which devices might require manual scans.  It also wouldn't leave
>> you guessing.  I'm hoping to get to this in the next couple of days.
>
> Sounds good.

Now that I understand, all of the above relates to "D".

>>> B) delete orphan links deletes too much
>>
>> Yes.  As you noted, this is specific to the "deaf" devices which get
>> ignored by the full scans and the possibility of the full scans not
>> getting to a device.  I really think that focusing on A above will
>> address the issue.  Also, I could address the "deaf" devices more
>> properly by using the date-time health code info rather than a brute
>> force "isa" against their type.  So, doing "B" relies on my interim
>> solution for "A".
>
> Good to know.
>
>>> D) scan all link tables never finishes for me (lots of details below).
>>
>> Isn't this the same as "A"?  Or, am I not understanding something?
>
> No, it's not the same. A) was indeed that I couldn't scan the PLM's own
> table and D is a full scan of my network.

Got it.