-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Looks like there's plenty of RAM spare on your box, Bill - try free -m
for a slightly clearer breakdown
total used free shared buffers cached
Mem: 2002 1953 48 0 76 772
- -/+ buffers/cache: 1104 898
Swap: 2999 227 2772
On this box, supposedly I only have 48 MB of RAM free- but in Linux
speak you have to bear in mind the OS uses RAM for caching filesystems,
and this cached content is counted towards RAM usage. Look on the middle
row, and you can see 1.1 gigs of RAM are being used to cache filesystem
contents (which sounds about right for this machine, which is a general
purpose box I use for development).
More seriously, you should look to MySQL for performance gains. Over
time, databases will get cluttered- look into the MySQL OPTIMIZE TABLE
command and friends, and tools like myisamchk to do table optimization
and analysis which will help remove cruft and tidy up indexes. Running
something like OPTIMIZE TABLE tablename; ANALYZE TABLE tablename; for
each table on a weekly basis sometime when the DB isn't likely to be
busy is a good idea.
http://dev.mysql.com/doc/refman/5.1/en/optimization.html has lots of
further reading-
http://dev.mysql.com/doc/refman/5.1/en/buffering-caching.html and
http://dev.mysql.com/doc/refman/5.1/en/optimizing-the-server.html are
good general purpose documents that you should have a good look at if
you're running MySQL. Since Rivendell relies heavily on MySQL for things
like scheduling and logging, it's worth keeping it all in check and
running smoothly.
https://launchpad.net/mysql-tuning-primer is a good little script that
will advise you on good options to tweak based on your MySQL server
performance to date and settings relative to hardware configuration.
Running that and tweaking configurations a little to be in line with the
recommendations is a fairly safe bet for an easy set of improvements.
It may also be worth enabling the slow query log and having a check for
any really slow running queries. If those queries are related to carts
or playout events, that might be something to look into.
Cheers,
James Harrison
On 29/08/2011 02:47, Robert Jeffares wrote:
> Bill,
> my observation is that over time rivendell computers get a bit slow and
> that seems to be a function of the size of the library. When more than
> 80% of the HD is used things get a bit pear shaped.
>
> My server is 4gb and it's all being used : I do have problems with hour
> long programmes being loaded in which is causing me some annoyance. The
> problem I have is at dropbox not at air time.
>
> I will be looking closer this week at the problem and see if there are
> any clues.
>
> regards
>
> Robert
>
> On 29/08/2011 09:37, Bill Putney wrote:
>> Robert,
>>
>> I just noticed that this machine is very light on RAM. It was a machine
>> I pressed into service because another one was crashing all the time.
>> Looks like I should add a few more GB.
>>
>> Still, I wonder if it was running out of memory why it didn't complain.
>> I can see it getting slow because it has to swap out but from the top
>> below you can see that there probably isn't much swapping going on (not
>> at least during the snapshot).
>>
>> Bill Putney - KPTZ Port Townsend, WA
>>
>> top - 14:20:20 up 2 days, 18:26, 2 users, load average: 0.63, 0.98, 1.01
>> Tasks: 161 total, 1 running, 160 sleeping, 0 stopped, 0 zombie
>> Cpu(s): 22.9%us, 6.9%sy, 0.0%ni, 70.3%id, 0.0%wa, 0.0%hi, 0.0%si,
>> 0.0%st
>> Mem: 1026380k total, 1011048k used, 15332k free, 61072k buffers
>> Swap: 3004408k total, 0k used, 3004408k free, 729828k cached
>>
>> df -m
>> Filesystem 1M-blocks Used Available Use%
>> Mounted on
>> /dev/sda1 131418 2244 122499 2% /
>> none 498 1
>> 497 1% /dev
>> none 502 1
>> 502 1% /dev/shm
>> none 502 1
>> 502 1% /var/run
>> none 502 0
>> 502 0% /var/lock
>> none 502 0
>> 502 0% /lib/init/rw
>> /dev/sda6 15923 748 14367
>> 5% /var
>> 192.168.192.201:/var/snd 938878 419092 472094 48% /var/snd
>>
>>
>> On 8/28/11 1:35 PM, Robert Jeffares wrote:
>>> next option is a ram upgrade or look to see how much space you have on
>>> the hard drive of the playout machine
>>>
>>> I think RD loads the long programme in somewhere close for convenience
>>> if the playout card is on the playout machine not the server .
>>>
>>> Dontch love this program? I am trying to get rd 2.x to compile on centos
>>> 5.6 and i am rapidly looking for an appliance install becaus there are
>>> so many ' missing ' bits.
>>>
>>> Robert
>>>
>>> On 28/08/2011 07:28, Bill Putney wrote:
>>>> Robert,
>>>>
>>>> The cart will play. All I have to do is add it into the running log and
>>>> start it. Which is what happens if someone happens to be there to catch
>>>> it when this happens. If I open the cut in RDLibrary "Edit Markers"
they
>>>> all look fine. I would think if the markers were screwed up the cut
>>>> wouldn't play after it was added back into the log.
>>>>
>>>> The audio is on a common server and is NFS mounted on the playout
>>>> machine. The server is connected to a switch and the server and the
>>>> playout client are on the same layer 3 network. Everything is 100 Mbps
>>>> and the server isn't doing anything but audio and database for
Rivendell.
>>>>
>>>> It is various programs. They seem to be the longer ones not spots, id's
>>>> or music cuts.
>>>>
>>>> We're using USB -> S/PDIF as output devices. The chipset is in the
Linux
>>>> HCL for audio.
>>>>
>>>> Bill Putney - KPTZ Port Townsend, WA
>>>>
>>>> On 8/27/11 12:08 PM, Robert Jeffares wrote:
>>>>> Bill,
>>>>>
>>>>> Does the cart play? Is all the audio in there? Are all the markers in
>>>>> the right place?
>>>>>
>>>>> What was the source of the Audio?
>>>>>
>>>>> Is it always the same programme?
>>>>>
>>>>> Have had probs with incompatible audio here.
>>>>>
>>>>> Robert
>>>>>
>>>>> On 26/08/2011 17:56, Bill Putney wrote:
>>>>>> At 17:03:50 in the following /var/log/syslog excerpt, Cart 80145
Cut 2
>>>>>> should have played. It is an hour long program. It skipped and did
not
>>>>>> play. . I don't know what "position out of bounds" means but it seems
>>>>>> that everything is playing out of bounds. The items before and
after the
>>>>>> cart that failed played normally. I see that the cart that started
after
>>>>>> (cart 80133) did not have a position out of bounds associated with
it.
>>>>>> Has anyone got an idea what that is or what I can do to get rid of
it?
>>>>>>
>>>>>> Any ideas? Anybody? Please!
>>>>>>
>>>>>> Thanks, Bill Putney - KPTZ Port Townsend, WA
>>>>>>
>>>>>> Aug 24 17:00:11 auto-mc rdairplay: *** position out of bounds on
signal:
>>>>>> Line: 24 Cart: 14418 Pos: 250025 ***
>>>>>> Aug 24 17:00:11 auto-mc caed: alsaStopTimerData(16)
>>>>>> Aug 24 17:00:11 auto-mc rdairplay: finished event: Line: 24 Cart:
>>>>>> 14418 Cut: 1 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:00:11 auto-mc caed: UnloadPlayback - Card: 1 Stream: 0
>>>>>> Handle: 166
>>>>>> Aug 24 17:00:11 auto-mc caed: LoadPlayback Card: 1 Stream: 0 Name:
>>>>>> /var/snd/045103_001.wav Handle: 167
>>>>>> Aug 24 17:00:11 auto-mc rdairplay: started audio cart: Line: 25 Cart:
>>>>>> 45103 Cut: 1 Pos: 0 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:00:11 auto-mc caed: PlaybackPosition - Card: 1 Stream: 0
>>>>>> Pos: 78 Handle: 167
>>>>>> Aug 24 17:00:11 auto-mc caed: Play - Card: 1 Stream: 0 Handle: 167
>>>>>> Length: 7863 Speed: 100000 Pitch: 0
>>>>>> Aug 24 17:00:12 auto-mc caed: FadeLevel: 0
>>>>>> Aug 24 17:00:19 auto-mc rdairplay: *** position out of bounds on
signal:
>>>>>> Line: 25 Cart: 45103 Pos: 7801 ***
>>>>>> Aug 24 17:00:19 auto-mc caed: alsaStopTimerData(16)
>>>>>> Aug 24 17:00:19 auto-mc rdairplay: finished event: Line: 25 Cart:
>>>>>> 45103 Cut: 1 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:00:19 auto-mc caed: UnloadPlayback - Card: 1 Stream: 0
>>>>>> Handle: 167
>>>>>> Aug 24 17:00:19 auto-mc caed: LoadPlayback Card: 1 Stream: 0 Name:
>>>>>> /var/snd/070003_001.wav Handle: 168
>>>>>> Aug 24 17:00:19 auto-mc rdairplay: started audio cart: Line: 26 Cart:
>>>>>> 70003 Cut: 1 Pos: 0 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:00:19 auto-mc caed: PlaybackPosition - Card: 1 Stream: 0
>>>>>> Pos: 0 Handle: 168
>>>>>> Aug 24 17:00:19 auto-mc caed: Play - Card: 1 Stream: 0 Handle: 168
>>>>>> Length: 210372 Speed: 100000 Pitch: 0
>>>>>> Aug 24 17:00:20 auto-mc caed: FadeLevel: 0
>>>>>> Aug 24 17:03:50 auto-mc rdairplay: *** position out of bounds on
signal:
>>>>>> Line: 26 Cart: 70003 Pos: 210278 ***
>>>>>> Aug 24 17:03:50 auto-mc caed: UnloadPlayback - Card: 1 Stream: 0
>>>>>> Handle: 168
>>>>>> Aug 24 17:03:50 auto-mc rdairplay: finished event: Line: 26 Cart:
>>>>>> 70003 Cut: 1 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:03:50 auto-mc caed: LoadPlayback Card: 1 Stream: 0 Name:
>>>>>> /var/snd/080145_002.wav Handle: 169
>>>>>> Aug 24 17:03:50 auto-mc caed: alsaStopTimerData(16)
>>>>>> Aug 24 17:03:50 auto-mc rdairplay: started audio cart: Line: 27 Cart:
>>>>>> 80145 Cut: 2 Pos: 0 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:03:50 auto-mc rdairplay: finished event: Line: 27 Cart:
>>>>>> 80145 Cut: 2 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:03:50 auto-mc caed: PlaybackPosition - Card: 1 Stream: 0
>>>>>> Pos: 0 Handle: 169
>>>>>> Aug 24 17:03:50 auto-mc caed: Play - Card: 1 Stream: 0 Handle: 169
>>>>>> Length: 3594892 Speed: 100000 Pitch: 0
>>>>>> Aug 24 17:03:50 auto-mc caed: UnloadPlayback - Card: 1 Stream: 0
>>>>>> Handle: 169
>>>>>> Aug 24 17:03:50 auto-mc caed: LoadPlayback Card: 1 Stream: 0 Name:
>>>>>> /var/snd/080133_001.wav Handle: 170
>>>>>> Aug 24 17:03:50 auto-mc caed: FadeLevel: 0
>>>>>> Aug 24 17:03:50 auto-mc rdairplay: started audio cart: Line: 28 Cart:
>>>>>> 80133 Cut: 1 Pos: 0 Card: 1 Stream: 0 Port: 0
>>>>>> Aug 24 17:03:50 auto-mc caed: PlaybackPosition - Card: 1 Stream: 0
>>>>>> Pos: 0 Handle: 170
>>>>>> Aug 24 17:03:50 auto-mc caed: Play - Card: 1 Stream: 0 Handle: 170
>>>>>> Length: 1740094 Speed: 100000 Pitch: 0
>>>>>> Aug 24 17:03:50 auto-mc caed: FadeLevel: 0
>>>>>> Aug 24 17:04:00 auto-mc caed: ****** ALSA Capture Xrun - Card: 0
******
>>>>>> On 8/24/11 8:04 PM, James Harrison wrote:
> Check rd.conf for your own system specific logging configuration, but
> check both the daemons and syslog logfiles; alternatively, configure a
> specific log path or syslog logfile/service. Not sure if automation
> decisions are logged there, but playout system stuff certainly gets
> logged to those locations from caed...
>
> Cheers,
> James Harrison
>
>
> On 25/08/2011 02:49, Bill Putney wrote:
> >>>>>>>> This is driving us crazy! At least once a day, a cart that is
scheduled
> >>>>>>>> to play (usually a long one like a prerecorded show) is
skipped in the
> >>>>>>>> middle of a log.
> >>>>>>>>
> >>>>>>>> I can't figure out why this happens but I know somewhere there
has got
> >>>>>>>> to be a system log file (like syslog or something) that should
tell me
> >>>>>>>> everything the automation system is doing and I should be able
to see
> >>>>>>>> the skip happen.
> >>>>>>>>
> >>>>>>>> We're using RRAbuntu 1.7.2...
> >>>>>>>>
> >>>>>>>> Help!
> >>>>>>>>
> >>>>>>>> Thanks,
> >>>>>>>>
> >>>>>>>> Bill Putney - KPTZ Port Townsend, WA
> >>>>>>>> _______________________________________________
> >>>>>>>> Rivendell-dev mailing list
> >>>>>>>> [email protected]
> >>>>>>>> http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Rivendell-dev mailing list
>>>>>>> [email protected]
>>>>>>> http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev
>>>>>> _______________________________________________
>>>>>> Rivendell-dev mailing list
>>>>>> [email protected]
>>>>>> http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev
>>>> _______________________________________________
>>>> Rivendell-dev mailing list
>>>> [email protected]
>>>> http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev
>>
>> _______________________________________________
>> Rivendell-dev mailing list
>> [email protected]
>> http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAk5bAEEACgkQmJV2s0zjsDhRJQCfSGiNkk0kDrZU02rIEC3nsqmt
kikAoIWFpYaG+SgXw1XK9WECqSqa/VI9
=wkfY
-----END PGP SIGNATURE-----
_______________________________________________
Rivendell-dev mailing list
[email protected]
http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev
