Weeks after Red Bee Media's broadcast centre fell over, Channel 4 is still struggling with subtitles
- Reference: 1634650745
- News link: https://www.theregister.co.uk/2021/10/19/c4_subtitles/
- Source link:
Things went horribly wrong at Red Bee Media's broadcast centre back on 25 September. Yes, that was the weekend before we ran [1]an accidentally appropriate episode of Who, Me?
A fire suppression was triggered and severely damaged a lot of critical hardware. The net result was that a number of UK television channels (including the BBC as well as Channel 4) suffered a wobble. While others have recovered, Channel 4 remains unable to provide accessibility services, such as subtitles or audio description.
[2]
A bit of background: an industry insider (who asked to remain anonymous) explained to The Register that TV companies generally provide several items to playout providers, usually file-based. Video and audio tend to be batched together in one MPEG stream. Audio Description (AD) and subtitles arrive in another. And so on.
[3]
[4]
"In the 'olden' days of 2006, we manually got all of the different components and placed them into the various different output servers," explained our insider.
The playout server described in our Who, Me? story would grab the video, audio, AD and subtitles – which explains its sudden shutdown in our story, that caused similar problems.
[5]
Automation took over in the last decade, meaning that skilled intervention was perhaps not so essential. Up until everything went wrong.
Red Bee posted a Twitter thread last week, apologising for the situation, but insisting it was "getting back on track."
As soon as we receive the media (the programmes), our teams proceed to create and add pre-recorded subtitles, audio descriptions and signing. This process has been disrupted since the incident in the Broadcast Centre in London on 25 Sep, but we are getting back on track. (2/3) — Red Bee Media (@RedBeeMedia) [6]October 15, 2021
We don't know how far away those tracks are. Presumably quite some distance if, after three weeks, Red Bee Media has yet to haul itself back onto them.
[7]If your apps or gadgets break down on Sunday, this may be why: Gpsd bug to roll back clocks to 2002
[8]Heart FM's borkfast show – a fine way to start your day
[9]Giant predatory ancient insects pioneered mobile comms 310,000,000 years ago
[10]Women techs fume, offer crowdsourced fixes as Michelle Obama's online keynote crashes
Our source noted that it was not the first time the fire alarm had gone off, but was the first time that nobody had managed to get to the override in time. The thinking behind the system was apparently "kill the flames and sod the consequences." As for those consequences: "Now we know... it kills servers."
As for the hardware, hot-swap spares are likely limited and there is every possibility that parts of the kit could be quite difficult to find nowadays. Red Bee Media got the pictures and audio up and running again quickly, but the ancillary systems (such as the ones dealing with subtitles) are clearly proving problematic.
We have asked Red Bee Media to comment but it has yet to respond. A spokesperson told the [11]BBC :
[12]
"Things are improving every day and we are able to deliver more and more accessible programmes, but we are unfortunately still experiencing issues with receiving the media for which our access teams create pre-recorded subtitles, audio descriptions and signing."
As for Channel 4, it fired up an emergency backup but reported that the system supposed to provide subtitles failed. The building of an entirely new system is now apparently underway, according to a [13]release emitted by the broadcaster .
"Not only will this enable our channels to move back out of disaster recovery," it said, "but it also means we will be able to provide subtitles, audio description and sign language services as well."
It's not great, and there is every chance that well over a month will have passed before full recovery. According to Channel 4, AD and sign language services were "irretrievably lost during the incident," making one wonder exactly what backup strategy was in place. We asked the broadcaster to explain how its backups worked (or if it was something it had outsourced) but have yet to receive a response.
The ongoing woes are a salutary, and highly public, reminder to all administrators of organisations large and small. Check your backups. Rehearse your disaster recovery. Customers have very, very long memories. ®
Keep up the pressure. This is unacceptable. [14]https://t.co/PNpuPK6TiZ — Adam Hills (@adamhillscomedy) [15]October 14, 2021
Get our [16]Tech Resources
[1] https://www.theregister.com/2021/09/27/who_me/
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/servers&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YW7rshCrJqEKXwau26VhrAAAAM8&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/servers&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YW7rshCrJqEKXwau26VhrAAAAM8&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/servers&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YW7rshCrJqEKXwau26VhrAAAAM8&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/servers&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YW7rshCrJqEKXwau26VhrAAAAM8&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://twitter.com/RedBeeMedia/status/1449034190129926145?ref_src=twsrc%5Etfw
[7] https://www.theregister.com/2021/10/19/gpsd_bug_reset/
[8] https://www.theregister.com/2021/10/18/bork/
[9] https://www.theregister.com/2021/07/09/titanoptera_wing_communication/
[10] https://www.theregister.com/2021/06/11/women_tech_obama/
[11] https://www.bbc.co.uk/news/entertainment-arts-58965380
[12] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/servers&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YW7rshCrJqEKXwau26VhrAAAAM8&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[13] https://www.channel4.com/press/news/whats-happened-access-services-channel-4
[14] https://t.co/PNpuPK6TiZ
[15] https://twitter.com/adamhillscomedy/status/1448555004655247362?ref_src=twsrc%5Etfw
[16] https://whitepapers.theregister.com/
Missing a few key points here
1/ Slightly glossed over is the fact the servers are fine but the hard drives in them died. It was generally known that the noise from gas fire suppression systems can damage hard drives, but not sure it was well known. I certainly didn't know how extensively, percentage of HDD's that would be destroyed, this can cause.
A little bit more from C4 themselves.
https://www.channel4.com/press/news/whats-happened-access-services-channel-4
2/ "Red Bee Media got the pictures and audio up and running again quickly," , yes but the key thing is here not from the primary site but from a DR site.
They are all still on DR, see Channel 5 with the white and black square in the right hand corner, indicating they are *still* in DR playout. The issue is Channel 4's DR subtitle playout isn't working. I heard some concern that fixing this is dangerous whilst in DR (as they have no further DR if they break something). So changes to the Red Bee's DR being made are very limited.
Re: Missing a few key points here
"It was generally known that the noise from gas fire suppression systems can damage hard drives, but not sure it was well known."
Fairly well known to El Reg readers me thinks.
My knowledge of fire suppression systems is v outdated. Does anybody know if such a system would be equally effective if it discharged the gas slightly more slowly? Or does it need to discharge so rapidly to get high turbulence = effective mixing?
Oof, this probably hasn't been a fun few weeks for the sys-admins, engineers, and various sundry dogsbodies at C4 and Red Bee.
We've all screwed up in prod before, but having your problems as publicly visible as a TV channel must be terrifying.
I'll save one of these for them when they finally get it fixed>>>>>
What's up with Channel 5 though ? Still has the "backup" black and white square marker nearly all the time.
I've no doubt the Tories will use this as another reason to privatise Channel 4.
Meanwhile, here in the darkest North, thousands still don't have *any* TV or *not much* radio because of the Bilsdale mast fire...
Their site says you should have got a voucher to pay for some sort of OTT hardware as a replacement.
Is this not the case ?
Testing failover.
At great expense, the last system I designed had a proper, offsite, in a different datacentre, DR platform. Whilst it needed a little work to point to a different DB, it was a complete replica of Prod. And it was tested. Yearly.
Getting it approved was a nightmare. "Oh, can't you just rebuild it?" "Of course, me and, ahem, which army?"
But it got built when I decided to turn my phone off one weekend when we had a small outage
So, I ask again, was DR ever tested? I'm guessing that's a big fat NO
People rabbiting on about whether or not a fire suppression system would (or indeed should) damage hardware are missing the point. This is a classic case of a lack of risk management leading to an error in egg location. That is to say that there are a number of catastrophic failures that can occur to a given basket (fire, flooding and overzealous fire suppression being only three). As such it's always best to have some hot or even cold standby eggs in another basket located as far from the first basket as is practical.
Oops
Red Bee: Doesn't really matter, the broadcasters have got a backup anyway.
Ch 4: Doesn't really matter, Red Bee have got a backup anyway.
Insert the adage about the backup hasn't been tested until you restore it here.