[mythtv-users] My mythweb shows me only new, mythfrontend only old recordings :-(

[Kenni Lund]

2011/4/12 Christoph Paul <chx at cvmx.net>:
> Am 10.04.2011 um 18:12 schrieb Christoph Paul:
>> Am 10.04.2011 um 17:16 schrieb Kenni Lund:
>>> 2011/4/10 Christoph Paul <chx at cvmx.net>:
>>>> Am 10.04.2011 um 14:08 schrieb Michael T. Dean:
>>>>> On 04/10/2011 05:46 AM, Christoph Paul wrote:
>>>>>> Hi all,
>>>>>>
>>>>>> since 7th of April, I see new recordings only on my mythweb and recordings up to this date only on the frontend.
>>>>>> I see  +/- all files in the filesystem storage and +/- all entries in the mythconverg recorded table.
>>>>>> Any ideas how to get things working again would be highly appreciated.
>>>
>>> Using EIT? On around the same day - the 7th of April (depending on
>>> timezone), a Danish DVB-C provider started sending invalid characters
>>> ...
>>> Try to check for NULL characters in your recorded table, something
>>> like the following should do it:
>>> SELECT chanid,starttime,title FROM recorded WHERE title LIKE '%\0%' OR
>>> subtitle LIKE '%\0%' OR description LIKE '%\0%' LIMIT 10;
>> Kenni, the select returned an empty set. So I also cleaned up all channels from the channel
>> or program table, that I set invisible - but no luck :-(
> Yesterday I restored my database to a date before the 7th and voila - everything fine again!
> So I re-inserted the missing recorded lines: Strange behavior again!
> Did that two times, as I first did not trust my eyes. Then I remembered your post ...
>
> There are about 50 lines. At least one seems to trigger this behavior.
> So maybe it is not the \0 but something similar!

In that case, please see if you are able to identify the character(s)
causing the problem. I'm currently looking into how to solve the \0
issue and if several characters are causing issues, we should of
course take care of all of them at the same time. Are you using EIT or
something else?

I did a little testing, and it seems like mysqldump is a good way to
identify the special characters. If you run mysqldump [database]
[table] > dump.txt, mysqldump will escape the special characters, so
inline control characters will be visible in a text editor. For
example; a string seperated with NUL-chars will get printed like 'This
is a string' if you try to select it in mysql...but if you dump it
with mysqldump, it will be shown as 'This\0is\0a\0string' in the text
file. I assume it will use the same approach for other control/special
characters, making it possible to identify the invalid characters.

There's a ticket about the null character issue:
http://code.mythtv.org/trac/ticket/8707

Best regards
Kenni