Saturday, September 17, 2016

I hadn't thought of sound as destructive to data . . .


. . . but according to this report, it certainly can be.  (A tip o' the hat to reader Snoggeramus for providing the link.)

ING Bank’s main data center in Bucharest, Romania, was severely damaged over the weekend during a fire extinguishing test. In what is a very rare but known phenomenon, it was the loud sound of inert gas being released that destroyed dozens of hard drives. The site is currently offline and the bank relies solely on its backup data center, located within a couple of miles’ proximity.

. . .

The purpose of the drill was to see how the data center's fire suppression system worked. Data centers typically rely on inert gas to protect the equipment in the event of a fire, as the substance does not chemically damage electronics, and the gas only slightly decreases the temperature within the data center.

The gas is stored in cylinders, and is released at high velocity out of nozzles uniformly spread across the data center. According to people familiar with the system, the pressure at ING Bank's data center was higher than expected, and produced a loud sound when rapidly expelled through tiny holes (think about the noise a steam engine releases).

The bank monitored the sound and it was very loud, a source familiar with the system told us. “It was as high as their equipment could monitor, over 130dB”.

Sound means vibration, and this is what damaged the hard drives. The HDD cases started to vibrate, and the vibration was transmitted to the read/write heads, causing them to go off the data tracks.

. . .

Researchers at IBM are also investigating data center sound-related inert gas issues. “[T]he HDD can tolerate less than 1/1,000,000 of an inch offset from the center of the data track—any more than that will halt reads and writes”, experts Brian P. Rawson and Kent C. Green wrote in a paper. “Early disk storage had much greater spacing between data tracks because they held less data, which is a likely reason why this issue was not apparent until recently.”

Siemens also published a white paper a year ago saying that its tests show that “excessive noise can have a negative impact on HDD performance”. Researchers said this negative impact may even begin at levels below 110dB.

There's more at the link, including an interesting video of a test of the effect of sound on hard disk drives conducted in 2008.  Worthwhile viewing, particularly for tech buffs.

I worked in data centers as a computer operator back in the 1970's, when technology was a lot less advanced than it is today.  One contained an IBM System/370 Model 145, identical to the one described in this archived IBM release.  (It was much less powerful in terms of processor speed, etc. than any personal computer today, and probably less powerful than most smartphones!)  It looked very like this:




Those odd-looking plastic dome things on the machines in the foreground are removable hard disks.  Each held only a few megabytes of data.  The technology of the day was much slower and much more primitive than it is now, so the data tracks on the disks were much farther apart and much less vulnerable to vibration caused by sound.  A single laptop computer hard disk drive today will hold many, many times more data than all of those early hard disk drives put together, and fit into a relatively small pocket!

I recall banks of gas cylinders outside the computer room, designed to release fire-suppressing fumes into the data center whenever necessary.  However, none of us ever considered the noise of the gas being released as a potential hazard to disk drives.  The system was more likely to kill us!  One of my not-so-fond memories of that computer room was when we had a fire security inspection.  The inspector turned to the Operations Manager and asked whether he had replacement operators lined up and ready to go after a fire.  Puzzled, the Ops Manager replied that he hadn't - why did he ask?  The inspector then pointed out that the 'gas masks' provided for the operators were to prevent smoke inhalation only.  They had no oxygen cylinder to provide fresh air - but the halon gas that the fire suppression system would inject would absorb all the oxygen in the air.  The operators would be asphyxiated before they could get out.

We got new oxygen masks the very next day.

Peter

12 comments:

STxAR said...

We had a halon dump at a radio station I worked at. The lights would flash and a horn would sound, giving you 20 seconds to leave before the dump. But I didn't realize the mechanism that allowed halon to work. Glad we never needed it!!! I might not be here...

Glen said...

The acoustics can be altered by proper nozzle and piping design. Acoustic loading itself is an interesting subject, and should probably be a more integral part of design engineering.

It is almost 20 years since I've been involved in fire protection. At that time, IBM suggested dry pipe sprinkler systems for data centers. With units properly dispersed, only the affected unit would get wet.

Halon was on its way out, because of perceived destruction to ozone layer. I've been present for Halon dumps, both planned and unplanned. I've seen doors blown off of computer rooms, and terrified techies set new sprint records.

Fun stuff, for young folks.

TRX said...

All the data centers I've seen in the US, built since the 1990 or so, use water flood instead of Freon or other gases.

TRX said...

With a sound level so high it would crash hard disks, I'm guessing anyone still in the room would suffer irreversible hearing loss.

Uncle Lar said...

In my mind I pictured technicians threading all those gas nozzles and affixing sound suppressors.
Worked briefly in a building with a Halon fire system, in the basement. It was made abundantly clear to all of us that if the warning lights went active we had less than three minutes before becoming asphyxiated.
Strikes me that any such system would not be in compliance with the Americans with Disabilities Act. Allowing sufficient time for those with disabilities to evacuate sort of defeats the purpose doesn't it.

Douglas2 said...

I'm suspecting that it was a sudden atmospheric pressure change that did the damage really, all those dust-sealed drives not having adequate time to acclimatize and jets of high pressure air in through any tiny crack.

The SPL was just what they were measuring, so it gets blamed.

Anonymous said...

over 130 dB

Wouldn't that be lethal in and of itself, regardless of oxygen depletion?

Stan_qaz said...

They make special drives for high vibration systems like NAS boxes that are more resistant to vibration and should be better at dealing with high noise levels as well. They also have vibration sensors to help compensate for movement related issues.

http://www.storagereview.com/pick_the_right_drive_for_the_job_24_7_nas_hdds_vs_desktop_hdds

Stu Garfath. Sydney Australia. said...

I was stationed at RAAF Base Darwin, Northern Territory during 1974, as a RAAF Communications Operator/Telegraphist.
The Communications Centre building was divided into three separate sections, with Tech repair at one end, the Communications area at the other, and the computer section in the middle, which housed the data storage mainframe we knew as F.R.E.D. - Fucking Ridiculous Electronic Device, it was Australia!.
We non-techies were warned on pain of death never to touch, lift or even breath on the square transparent lids under which the discs did their work, whilst spinning at hundreds of revolutions a second, so we were told.

During an an inspection by a very senior RAAF Officer and his conga line of lesser mortals, they were proudly ushered into FRED's room, there to gaze in slack jawed wonder at the very expensive modern devices in front of them.

With no warning, the Senior Officer leaned forward and opened the square perspex lid of one of the disc housings, which totally disrupted the precise airflow necessary between reader and disc.
As I was later told by one of the Techs present, the record player styled reader arms, one for each disc, instantly plunged down and destroyed the disc below it, and so-on with all the other 11 discs in the unit, pieces of black disc, arms and other bits exploded with the sound of a grenade, everyone and everything being hit by the flying non-metallic shrapnel. To say the shit hit the fan big time would be a very large understatent!.

As can be imagined, pandemonium ensued, the inspection party was forcefully hustled out of the room, and everyone was forever sworn to secrecy,- in a Communication Centre?, yeah right!.
Sometimes, we Enlisted underlings had good days, this was one of them.



Anonymous said...

Halon does not absorb oxygen. It disrupts the combustion process through decomposition and free radicals combining with hydrogen, etc. Wikipedia had a good article.

BCFD36

Johnny Glendale said...

I worked in an edit facility where we were shooting a "How To" video (late 80's). The lighting guy put a light directly under a heat detector, which set off the Halon on the whole floor. As we were walking (walking!) out, one of the grips was trying to light a cigarette...but the lighter wouldn't even spark! That was an expensive shoot day...

Russ said...

Next year, I will celebrate surviving forty years in the IT business. I was a programmer in mainframe data centers for over ten of those years. Those disk packs are IBM 3333s for use in the 3330 drives. Each spindle was approximately 100 Mb, later expanded to 200 Mb.

Uncle Bill was a long-serving firefighter and he made sure I knew how to don and ditch the Scott SCBAs kept outside most data center doors. Halon does not absorb the oxygen, it displaces it upward, being more dense than air. It is effective at very low concentrations and does not cause "instant death" as was the urban legend at the time. Another mode of operation is that it decomposes on contact with heat and the halogens react with the hydrogen typically present in fires forming non-flammable compounds. By the way, if you were ever in a data center when the halon system went off, the screaming sound as it came out of the pipes would wake up even midnight shift operators and warn them to leave immediately.