Re: [RDD] Server for playout?

[Robert Jeffares]

Oh So True!

On 11/10/19 9:11 AM, Cowboy wrote:

  I once saw a RAID-5 fail, where the first disk had failed months prior,
  but because it automatically kept going, no one knew until it failed
  in a non-recoverable catastrophic way.


Same happened here on the one server that held all the spots for the 
biggest [in every respect] station on the network.


Someone [not me] spent many happy hours copying reel to cart which ran 
for a few days.


No one knew the disks had failed one by one, and there was no back up 
other than the reel tapes.


The guy who had to make it happen quit in frustration some weeks later.

R

___
Rivendell-dev mailing list
Rivendell-dev@lists.rivendellaudio.org
http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev

Re: [RDD] Server for playout?

[Cowboy]

On Thu, 10 Oct 2019 14:35:07 -0500
Tom Van Gorkom  wrote:

> I would like to hear about other good setups as well since I plan to
> rebuild the servers in the next few months (still on CentOS 6.9). A lot of
> the decision depends on how big your operation is.

 We run client/server with full redundancy,
 Redundant servers, redundant clients, redundant remote viewers, and a
 "wayback" machine running RDCatch as record, and 168 hours later, playback.
 Redundant sound cards ( ASI livewire ) are in the redundant servers.

 Even wayback is redundant, with one on-site, and one off-site.
 The off site is configured in a way that while it has a full copy of the
 entire audio store, and the database current to within 24 hours, it can not
 "run" that database or audio store. It's strictly off-site backup, current
 to within less than 24 hours.

 None of this except replication is automatic !
 If it fails in a catastrophic way, that's a big enough deal I want
 everyone to know about it, so it gets tended to ASAP.
 Now, we also have redundant Cisco switches, even redundant CAT cable to
 each location, so no one item is permanently fatal.
 Each identified single point of failure has a redundant device, either hot
 or cold backup next to, above, or below it.
 Even in the event the building becomes a smoking hole in the ground, there
 is a full backup available, and 7 days to get it up and running.

 I once saw a RAID-5 fail, where the first disk had failed months prior,
 but because it automatically kept going, no one knew until it failed
 in a non-recoverable catastrophic way.

-- 
Cowboy 

Grief is the Reminder that we Loved.
 And were Loved.
___
Rivendell-dev mailing list
Rivendell-dev@lists.rivendellaudio.org
http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev

Re: [RDD] Server for playout?

[Tom Van Gorkom]

I've had the same question before. 5 years ago when I first set up our
Rivendell network, I decided to go with master and slave servers and put
clients (6 of them) in the production and main airplay studios. I have
never had an issue with this setup but I always make sure that the master
server is backing up automatically to the slave and I have a spare client
to put into the airplay position if needed - and that has been needed once
or twice in 5 yrs. I like having the servers somewhat isolated from the
users under my care and accessible to me at any time to run checks, etc.

For backup on-air audio, I run a stand alone Rivendell box in the rack room
that I update once in a while with new audio files. It runs a continuous
loop of 30 days of logs and if my audio network detects silence for more
than 15 secs, it gets put on the air automatically. I can also put it on
the air when I want to do maintenance on the entire network... maybe once
or twice per year.

The other thing I like about using clients is that you don't have to have
large hard drives everywhere. I use 250GB SSDs in all my clients and now
with Axia, I can use small form boxes (no physical sound cards) and they
run very quietly taking up little space in the studio.

I would like to hear about other good setups as well since I plan to
rebuild the servers in the next few months (still on CentOS 6.9). A lot of
the decision depends on how big your operation is.

>From my limited perspective,

Tom Van Gorkom
Radio Esperanza Engineering, KRIO AM/FM, KOIR FM
Office: 956-380-8150
Cell: 865-803-7427

Rio Grande Bible Institute
4300 S US Hwy 281
Edinburg, TX 78539


On Thu, Oct 10, 2019 at 11:37 AM Rob Landry <41001...@interpring.com> wrote:

>
> Wherever possible, I run standalone Rivendell machines. The "server" is
> the "client".
>
> I do this because with a standalone machine no network failure can make
> the radio station go silent.
>
> With a client/server implementation, the audio is being transmitted in the
> form of data packets from the server to the client across a network. If
> something happens to the network -- say, the power supply of a switch or
> router fails -- the station goes silent. A standalone machine will keep
> running if its network goes away.
>
> The fewer potential points of failure there are in a system, the more
> reliable it will be.
>
>
> Rob
>
> --
> Сквозь грозы сияло нам солнце свободы
> И Linus великий нам путь озарил;
> Нас вырастил Stallman на верность народу,
> На труд и на подвиги нас вдохновил.
>
> On Thu, 10 Oct 2019, Mark Murdock wrote:
>
> >
> > Would it be a good idea, bad idea or indifferent to use a Rivendell
> server
> > as the on-air playout machine? Is it best to use a client for this
> purpose?
> >
> >
> >
> > Thanks,
> >
> >
> >
> > Mark Murdock
> >
> > KAMB
> >
> > 90 E. 16th St.
> >
> > Merced, CA 95340
> >
> > (209) 723-1015
> >
> > m...@celebrationradio.com
> >
> > Website
> >
> >
> >
> >
> >___
> Rivendell-dev mailing list
> Rivendell-dev@lists.rivendellaudio.org
> http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev
>
___
Rivendell-dev mailing list
Rivendell-dev@lists.rivendellaudio.org
http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev

Re: [RDD] Server for playout?

[Rob Landry]

Wherever possible, I run standalone Rivendell machines. The "server" is 
the "client".


I do this because with a standalone machine no network failure can make 
the radio station go silent.


With a client/server implementation, the audio is being transmitted in the 
form of data packets from the server to the client across a network. If 
something happens to the network -- say, the power supply of a switch or 
router fails -- the station goes silent. A standalone machine will keep 
running if its network goes away.


The fewer potential points of failure there are in a system, the more 
reliable it will be.



Rob

--
Сквозь грозы сияло нам солнце свободы
И Linus великий нам путь озарил;
Нас вырастил Stallman на верность народу,
На труд и на подвиги нас вдохновил.

On Thu, 10 Oct 2019, Mark Murdock wrote:



Would it be a good idea, bad idea or indifferent to use a Rivendell server
as the on-air playout machine? Is it best to use a client for this purpose?

 

Thanks,

 

Mark Murdock

KAMB

90 E. 16th St.

Merced, CA 95340

(209) 723-1015

m...@celebrationradio.com

Website

 


___
Rivendell-dev mailing list
Rivendell-dev@lists.rivendellaudio.org
http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev

[RDD] Server for playout?

[Mark Murdock]

Would it be a good idea, bad idea or indifferent to use a Rivendell server as 
the on-air playout machine? Is it best to use a client for this purpose?

Thanks,

Mark Murdock
KAMB
90 E. 16th St.
Merced, CA 95340
(209) 723-1015
m...@celebrationradio.com
Website

___
Rivendell-dev mailing list
Rivendell-dev@lists.rivendellaudio.org
http://caspian.paravelsystems.com/mailman/listinfo/rivendell-dev