Standards-obsessed boss ignored one, and suffered all night for his sin The Register loves standards – we have our very own Standards Bureau. Another standard we observe rigorously is that each Friday morning brings a new instalment of On Call, the weekly reader-contributed tale of tech support incidents that ripped up the rule book. This week, meet a reader we’ll Regomize as "Laslo," who once …
Long ago when I first worked with EMC equipment, I was talking with a longtime EMC tech about various subjects while we were waiting for some longish self check to complete after he'd installed some new drives and I recall asking him why the Symmetrix wasn't designed to fit the same dimensions as normal racks. Its oddball footprint caused it to take up more space since it wouldn't fit in the standard rack/aisle plan. He said that was deliberate, they designed it to spread out across enough floor tiles so that at maximum capacity it would meet standard datacenter per sq ft weight limits (though as I understand it, some datacenters have needed to have their substandard flooring reinforced nonetheless)
Made perfect sense when he told me, but it wasn't what I would have guessed. Apparently it has installation guidelines that specify exactly how it is placed on the tiles it occupies to insure it occupies the correct number in the correct way. They even accounted for the weight of people standing around it when it is being serviced etc.
I haven't been around one for years, I wonder if they have a more compact version now that the world has moved on from heavy hard drives to nice light flash drives.
Did you just drive it onto the floor, or was plywood placed down in the path to spread the load out so it wasn't all on the rather small contact patches of four tires? I've seen some strange things in datacenters before, but driving a vehicle onto the floor would be a new one!
The pair that we had delivered were a bit of a palavar. First the freight elevator didn't quite make it to the 5th floor (I know, who puts a data centre at the top of a building), so we had to unload all the disks, take them up, then take the frame up. EMC then had a type of hover cushion that they used to make movement through the office easier. Mind you getting them up the step that allowed us to have the raised floor was fun.
Ooh! Ooh! Call on me! I know! I know!
Our management moved our data center to the upper floors of a building which has an an open courtyard column extending to and through the topmost floor. Our IT folks caught this about halfway through the planning sessions. Management was hell-bent on using this locaton. So we are renting extra rooms around the perimeter of that open column, and our servers are located no more than three rows inward, from the outermost edge of the rooms.
When I were a lad...
University of Massachusetts, Amherst, 1976-ish. We swapped a CDC 3600 and a CDC 3800 for a slightly used CDC CYBER 74. The entire machine room was emptied and reconfigured for the new machine. A grounding (earthing) grid was installed, and the raised floor was strengthened. All over summer break, so that the machine would be up and operational when the students came back in September. Quite an effort.
It is my considered opinion that those old raised floors could probably handle anything you could throw at them today, supporting,as they did, washing-machine sized hard disks made from Real American Steel. Not to mention, the cruciform CPU.
// no "Atlas with the world on his shoulders" icon...but this is for the folks who did the work --->
I worked for a company that held a company ball in a marquee on the sports field (don't ask). The marquee was also the location for our yearly conference.
One year during the conference there was a storm brewing, and the wind was getting up a bit to much. So much so, in fact, that people were getting nervous sat inside what could be considered a very large kite. The wind got stronger and one of the guy ropes valiantly gave up and came pinging out of the ground. This obviously made the post it was holding in position lean inwards quite a bit. One of the IT managers bravely pushed against the pole to save everyone whilst an evacuation was performed. Once the marquee was empty was out (apart from IT - always expendable), the manager let go of the pole and ran. He needn't have bothered - the post didn't move in the slightest and was held in place by the half mile of canvas and various hardwood posts holding up the rest of the marquee..
He still thought he'd saved the day. We didn't have the heart to tell him
Not to mention if such a large marque DOES blow away, holding onto it is going to do absolutely squat in terms of holding it down. It's just going to launch the person to orbit (or a hospital. Possibly an early grave if the have a particularly solid lithobraking encounter.)
I think this was in the 1960s.
We were on holiday in Keswick, Cumbria at a performance of 'She Stoops to Conquer'* at the then Century Theatre (now the Theatre by the Lake**). In one scene the posh young bachelor has to be kept in a room, and the other actor leapt across the stage to bar the doorway with his arm against the prop doorpost. Which was not secured.
It swayed ominously towards the audience,
paused,
swayed back,
and settled.
Then they got on with the rest of the play (which was actually really good).
Everyone there that night will remember that scene forever. Happy times, I recommend the play (you will never forget the Horse Pond.)
*https://en.wikipedia.org/wiki/She_Stoops_to_Conquer
** https://www.theatrebythelake.com
try also the French language film/movie 'Final Cut'. Ostensibly a zombie movie, with real zombies, the last half hour explains everything. Very funny, and probably much truer to 'real life' than a lot of other movies about making movies.
My SO was in a community production of The Wizard of Oz. Shortly after receiving it (in the same scene), the Cowardly Lion's medal broke in half. The actor was impeccable - not only did he not drop either piece, he (perfectly in character) attempted to reassemble it (intentionally clumsily) and said "does that mean I still have courage?" or some such thing. It wasn't until after the show I found out that wasn't an intentional part of the production!
"Then they got on with the rest of the play (which was actually really good)."
This must be a different play to the one I remember then, a play so boring that I once missed a lighting cue as I hadn't noticed the actors had all gone off stage.
The most interesting thing about it was programming the shiny new computer lighting system to "flicker" some candle bulbs for added realism.
Sometime around 2006, I was having a corridor conversation with a colleague at our Southwest-of-England Very Large Tech Company employer, when certain other colleagues in our datacenter group approached us, carefully pushing a tall 42U rack that they had painstakingly assembled for an upcoming demo.
Well, the corridor in question was the transition point between legacy office space, and what had been until very recently, bonded warehouse space before being converted to offices.
Which meant there was a short ramped section of floor where the floor levels transitioned.
And so my friend and I watched as this very tall rack reached the ramp, and in glorious slow motion like a mighty sequoia, toppled over, with an extremely expensive crashing/shattering/pinging sound.
So we beat a tactful retreat with our coffees, leaving the DC guys to sob over their pile of bits.
We used to supply rotating machinery supported on (usually) very strong floors. These floors would be practical but utilitarian so to make things more aesthetically pleasing some architect decided that 'computer-room' false-flooring be used to improve his poor initial design. Inevitably, this low but suspended floor resonated with the machinery; the vibration was irritating. Of course, as every engineer knows, you can have ten-million pounds worth of rotating equipment running smoothly, under superb automatic hair-trigger control in an unmanned building, but if the flooring rattles it will be a poor day for the manufacturer. We fixed it, but that false-floor would never be removed without welding gear.....
At my previous employer we had a "server room" that was one of the retired professor's office, free as as the university's engineering department was in decline due to no new staff being appointed. A Sun rack in there with a 5kVA UPS, redundant server heads, and around 100TB of storage started to sink in to the wooden floor due to the high pressure around its feet and so we moved it live - literally running - to the other end of the room that was a concrete floor to allow the estates team in to replace the floor with a thicker plywood flooring in place of the modest chipboard originally used.
Then we moved the rack back again, slowly and with it still running (kept network cables in place, unplugged UPS feed), to it intended location once more.
When the facility was decommissioned in 2019 we had to dismantle the rack, as the newer lift was smaller and it did not fit in as it previously did on delivery 10 years previously, and I was astounded by how heavy the rack alone was, it must have been the best part of 150-200kg without equipment as each side needed two-three folks to manhandle it down the stairs.
When we were teen-agers, we constructed what we believed to be a shot-gun-proof door for my parents' garden-shed-appropriated-to-be-our-tech-area.
The original door was ancient and pitiful 1/4" tongue-and-groove construction. We started our new door with a frame of 2"x4"s, added an internal grid of 2"x4", added a layer of 3/4" marine plywood to one side of the frame+grid, added a 1/4" inch sheet of iron to the back side of the frame+grid, added another sheet of 3/4" marine plywood to that, used counter-sunk and washer'd bolts and nuts to hold the sandwich together, lag-bolted industrial hinges to the outside of all that, set it in the doorframe, and lag-bolted the hinges to the shed's internal frame.
It took six of us to manhandle the mega-door into place.
THEN, we nailed the original, wimpy door onto the outside of all that.
Why? Because we were teen-agers, and because we could.
A case of the customer was not right. I've just had to scour some bike manufacturers website because an accessory also needed an extra thingummy that wasn't listed in the 'things you might also like' so it took a while to find it. Oh for a web site where you are told 'this is what you are getting because of what you ordered to start with'.
And if they are ordering 1/4 million of stuff, why not just throw in the right castors for free, given changing a purchase order will likely cost more than several sets of castors and require the asker to be defended from the Paddington-style hard stares of the operations department?
I once worked for a place where EVERY purchase order, no matter the amount, required three signatures. Two of them directors signatures.
One day I spent an hour and a half chasing down signatures to buy a box of blank floppies (this was obviously a while back). One of the signatures needed was the director of purchasing. When I went to said director, I explained that at my pay rate, they just spent 5x the cost of those blank floppies in my time getting the signatures together. He didn't seem to care!
Needless to say, I started looking for another job within the first three months working there, and was out of there a couple of months later.
We have something very similar here, although in our case with travel.
Before there were a couple of part time ladies who handled it all for us, we just told them what we want and they valiantly fought with Concur and various travel sites and got it done.
Both more than worth their weight in gold for what they did, but certainly not paid that much or indeed anywhere near what they were worth.
Then bean-counter central (who never travel anywhere) decided it would be a good idea to get rid of them and make us all do our own travel arrangements etc.
So now we all have to spend 3-5x as long as our former travel gurus did as we don't know the systems as well as they did (and we're really getting to loath them as well), not to mention our hourly pay rate is probably also somewhere between 2-10x what they were originally on I would estimate (given this rule applies all the way up to director and VP level for self-booking).
Now quite where is the economy or the job satisfaction here again, especially given I spent more time travelling and out of the country last year than in it?
@Anonymous Custard - I think you might work for my previous employer - what you describe sounds very familiar!
Our admins, back in the day, knew all the ins and outs of Concur and Amex Corporate Travel. A simple email “Hi Kathy, I need to be in Munich next Monday & Tuesday, could you make the arrangements please?” and it would be done; no fuss, no mess, super efficient.
Then it was decided that to save money (ha!) we should all do our own travel arrangements.
Cue lots of senior engineers spending hours of their time wrestling with Concur and battling the Computer-Says-No nature of the corporate travel rules (“Yes, I have not booked the cheapest flight, because the cheapest flight leaves Heathrow at 3AM, routes me via Ulaanbaatar and has a 57 hour layover before getting me to Munich at midnight the night before my 8AM presentation to the customer…”
Oddly similar to the university I used to work at. At one point the professor and his side-kick went on recruitment trips to France and Singapore, both resulted in about 10-20 MSc or similar students per visit (around £100-200k or more in fees). Then the central office dictated that was inefficient use of resources to pay £2k per trip as they had a recruitment department anyway, so it stopped. Following year has 3 students, year after the related MSc course was cancelled.
I think you can understand why Dundee lost out...
Then it was decided that to save money (ha!) we should all do our own travel arrangements.
My beloved employers - a large UK university - went the other way and decided that rather than making our own travel bookings, everything had to go through an in-house travel agency. Previously I could book train tickets - say - for myself in ten or fifteen minutes, including all the flexibility I needed. That became fifteen minutes to fill in a form, wait two days, get the proposed bookings, point out the errors, wait two days, get the new proposed bookings, approve them, wait two more days, get the tickets, point out the errors, wait a week and get the right ones. Fifteen minutes of my time became at least two hours over a couple of weeks.
Furthermore, the in-house agency would never book advanced purchase tickets, because their notional profit came from commission, so the tickets they bought me were generally two to three times more expensive than I could have bought myself.
Just one of the reasons that the university in question is in deep financial trouble and recently paid me and several hundred others a shit load of money to go away. Win.
We had a similar set up - two very expert (and very patient!!) in-house travel folk (Declan and Fiona - much missed). You could walk into their office and they’d tell you all the options within a couple of minutes and you’d say ‘that sounds like the best option’ and by the time you got back to your desk it’d all be booked.
Then came Concur (or Con-curse) and we had to DIY - simply terrible. I live in U.K and was once in Denver for a few weeks and something in Concur automatically cancelled my return flight (I think it was because it’d been rescheduled by a few hours or something). I found out the day of travel when attempting to check in. I called the airline who said ‘ah yes, your ticket was cancelled yesterday’. The Concur ‘emergency’ support was available between 9am-5pm UK time, which is pretty useless if you’re travelling. In the end I bought a ticket (which was very expensive) at the airport on Corporate credit card. Just that additional spend alone was as much as one month’s salary at the time.
I don’t mind ‘wasting’ company time too much if that’s what they have decided to pay me to do (it’s a bit irritating), but I DO mind the company working against me when I’m literally trying to do the job! Stopping me leaving site and making things difficult getting home is NOT ok!
Not dissimilar to my mileage claims when I was travelling between schools. The precise and detailed claim form required a list of each location I visited, with the reading off the milometer at start and end of each individual journey, the number of miles to the next stop the readings there, and then the next and then.....etc. and the totals. Which didn't vary by more than a mile or two each week and might have equalled about £10 a week or £400/pa
Each stop probably required 2 or three minutes of my time recording the numbers. So about an hour a week ( not including copying and forwarding the bloody thing to finance ) 30 or 40 hours in a year, at a senior teacher's hourly rate must have wasted around £1000 per teacher. If we'd been allowed to just claim what we knew was our mileage and we had all, each and every single one of us fiddled our mileage by adding a mile or two a day it would still have been more than 50% cheaper than the cost of that admin. Except we were already a sunk cost - that 40 hours a year didn't come out of their budget, but our teaching time. Each one of us could have supported an extra kid or two. And of course some staff members didn't bother to claim- which saved the authority a few hundred quid a year.
>And of course some staff members didn't bother to claim- which saved the authority a few hundred quid a year.
...which of course is the point! Make the process sufficiently onerous and/or soul-crushingly tedious, and a decent percentage of people just won't bother claiming money that is rightfully theirs - ka-ching! Trebles all round for the Finance department.
See also: mail-in rebates, HMRC tax refunds, etc.
When I dealt with servers, the standard SUN Rack was rated to host 3 V880/V890, since each of those beast was weighting between 150 and 200Kg ( depending on configuration ), the rack had to be able to hold at least 600Kg. ( and obviously the raised floor had too )
( don't ask me how you service the top V880/V890... just remember one important thing : those flimsy anti-tilt bars on a SUN Rack that can be extended from the bottom are probably not enough, I never tried it, I used Schroff racks with a massive and heavy plate bolted at floor level ( raised or not ) that was doing the job )
I think that's a bit unfair on today's young'uns - even 30-odd years ago when I was choosing which course/uni to go for, engineering degrees were considered hard and therefore something you only did if you were REALLY into the subject, so unless you're sufficiently old to be able to remember a time when engineering genuinely was considered a desirable subject to study, then I think you're just forgetting how little interest there's been for it across several generations-worth of students.
And having only recently completed a fairly lengthy period of recruitment for final year students to see who we'd want to offer graduate-level positions to later this year, we weren't struggling to find a suitable number of potentially good enough candidates to bring in for the onsite evaluations, so as far as UK universities, and electronics/software engineering courses specifically are concerned at least, there is still a decent pool of prospective talent coming through the system. Whether the overall numbers are down compared to decades gone by is another matter, so it may be that there aren't quite as many different unis offering the same courses, but it also doesn't feel like demand has dropped off a cliff and left only a few courses still hanging on by their fingertips to survive.
For my university, engineering of some kind was a very common degree, and a substantial percentage of the university as a whole. Of course, that was only, ah... hm, apparently 25ish years ago.
They were bad about not having enough professors. Well do I remember having to sign up for a 15-minute slot with my "advisor" and being required to bring a copy of my transcript and a list of what courses I proposed to take, as the advisor didn't have time to look up anything or do proper advising. (Each was responsible for hundreds of students.)
They all want to be marketers and social media influencers now. Actual science and engineering is just too hard.
When I read an article a while back about how today's youth don't know how to use basic office equipment, there was a quote from a a recent ENGINEERING graduate about how the office printer was too complicated, and she didn't understand how to use it.
As much as I hate the idea of AI everywhere, maybe we do need some form of "intelligence", since it doesn't seem to exist in the younger generations...
*From another grumpy, almost old man.
Or, and this might be wrong but it's possible, the printer is too complicated?
I hate office printers and try to have as little to do with them as I can. This is especially true when the IT people have tried to set up something to make them easier to use, but it doesn't work. I've had the experience of trying to get a printer to recognize my credentials and print my document and it seems very unwilling to do so. I've had the experience of calling in my colleagues, also programmers, and none of us can. Maybe that's because the printer's broken, though it doesn't say so, but maybe we're all too stupid to understand the obvious method IT had in mind for how to turn this on. I start wondering whether I'm supposed to be pressing more buttons or if the server that's controlling print jobs is not working. Sometimes it eventually works. Sometimes it doesn't.
Most of the time, I'm using the office printer because I don't print much at all and I figure they can spare a sheet, so I don't build up the experience to know exactly what you have to do.
If you have a complex printer system, for example one with multiple printers you can send your print job to, some rule about where the printer you need to use will be which don't include some printers, and several steps when you physically get to a printer, and you write none of it down, don't act surprised when some infrequent users find this less than intuitive. Someone can get confused by that, and it's the configuration, not the printer, that caused the problem.
Ah, been there. One company had the bright idea of a printer system where you print to one huge master queue, no matter where in the building/complex (or, for that matter, state, since there were multiple sites), and when you get to your printer of choice, flash your badge at the badge reader and hit print. In theory, it would simplify printing for everyone, big win! But in practice,
1. If the badge reader was down, that printer was down. Not even copying or faxing was possible without badging in first.
2. Once you submitted a job to the queue, you couldn't see it anymore. I think it was possible to cancel it by going to a printer, looking at your personal queue, and deleting it there. I think.
3. User-supplied printer settings? Yeah, right, like you'd expect them to be properly copied from the local machine to the Big Queue, then transferred correctly to the local printer, which hopefully supported those options.
4. Print jobs from SAP had to go the REAL print queue for the printer. Being a non-US company, SAP always output on A4. Being a US site, every printer was loaded with letter-size paper. There was a line on the work orders that none of us knew existed until the day the print margins were messed up (more than usual) so there was enough space to print it!
It may well be - remember "Xeroxgate" about ten years ago? Most large Xerox scan/print/copy machines habitually swapped characters, due to an over-eager data compression algorithm. So you could never be entirely sure that your scans were accurate or if the scanner had changed some sixes into perfectly aligned, clear eights.
Xerox first claimed that it only happened at some settings, as explained briefly on page 16odd of the manual. Still, it later was proven that it happened at any setting - and, as the issue was a few years old, many archives of scanned data (and hard photocopies, as those were effectively scanned and printed) could not be trusted any more.
Ha! OK, that one made me laugh.
And gives me an excuse - as if needed! - to post the obligatory Far Side cartoon :)
It’s funny you say that, I think most of the young people are just doing a great job on the cost-benefit analysis.
Let me ask you this. What was your salary when you bought your first house, what’s that house worth now, what salary would you need to have the same ratio of salary to house price?
Played this game with my dad at Christmas. The answer was £180k. He was a bank teller, early 20s with no qualifications, 3 kids and the wife didn’t need to work.
Try engaging that oh so powerful brain and consider that people don’t choose engineering, not because it is hard, but because it doesn’t pay anything - and their boss would be out of touch and patronizing
The time that a datacenter at a site I was consulting had to undergo some minor reorganization to account for growth in its SAP environment (I was a storage consultant on that project) The problem was that the servers (across 7 or 8 racks IIRC) that needed to be moved were running some super important homegrown application and among the most important the company had, with almost no tolerance for downtime. Worse, these were not clustered so you couldn't simply shut down a server, move it, then power it up. The business owners of that system were throwing out dates months away when their systems could be relocated. Plus to avoid this situation ever again, the business owners of the systems to be moved wanted them placed in the newly opened section of the datacenter, which was a LONG way (easily over two hundred meters) from the current location. They had a DR site but actually activating it would apparently cost a ton of money, and the CIO vetoed any plan involving use of DR to accomplish the move (probably they just weren't confident it would actually work)
The possibility of waiting months until the new servers could be installed was a big problem for the SAP project (maybe not for me, if I was able to stick around and keep billing while we waited) so the leader of that project asked to send a few of his people to their meetings to figure out a better plan. Before the meeting we went out onto the floor to inspect the systems, and it looked like while they weren't clustered, as such, they were fitted out with redundancy for power, network and fibre channel.
So a plan was hatched to move the racks with these extremely important servers while they were running. It was determined that if network and fibre channel cables were extended from its new home to its current home, one set of network cables and one set of FC cables could be disconnected, and these new extended connectors substituted, then the other set disconnected, and the redundant set of long connectors substituted, so it was running on the extensions with redundancy (in case of a mishap, like a rack wheel running over a fiber cable lol) To handle ethernet distance limitations ethernet switches were laid on top of the rack to be moved, on a table next to the spot where the rack would be moved to, and ethernet cables carefully measured as being 99 meters long were prepared. There were a couple intermediate switches along the path as well (since ethernet is limited to 100 meters) powered by an unplugged UPS. Then a rack could VERY carefully be moved, with people gathering up the slack in those ever so important overly long cables (as well as the intermediate switches and UPSes) until it reached its new home where it would be connected to first one, then another set of normal length cables.
The one hitch was power, the electricians were wholly against the idea of powering everything with very long power cables with multiple extensions because they didn't know what the effect of voltage loss across such a long run would be and it was way out of spec for anything they had ever done before. There were some options discussed about plugging and unplugging it along the way with shorter cables, but the solution ended up being a pair of rather beefy UPSes that would follow along on a cart behind it so the rack would run on battery providing redundant power.
This was rehearsed several times with racks containing test/QA servers from that project being shuttled back and forth multiple times until the business owners were satisfied this crazy plan would work. I actually stayed that weekend even though I wasn't part of the actual move, as did a few of the other SAP consultants, because we just had to watch something we spent a couple long weeks planning actually work. And it did, without a single hitch. Well, other than when one of the racks was moved one of the switches haphazardly laying on top shifted off (possibly because the person gathering the ethernet cable slack was a bit too aggressive) and almost took out one of the guys pushing the rack. But despite crashing to the floor and gaining a rather nasty looking dent in one corner, the switch kept functioning.
Since the guys doing this move were all regular employees and their business owners were nowhere to be seen, us consultants took them all out afterwards (bars open until 4am on Saturday came in handy) boy did they get wasted. They had been under a lot of stress with so much worry about the move, but they pulled it off perfectly!
How else can you keep the uptime record*? A heavy pushcart, a pair of UPSes, several 100 meter cat5 cables, and an expensive, and redundant router. You have to do all of the steps in precisely the right order, and you have to wait until the data center manager has left for the day, but it most certainly can be done.
* - only to be defeated a couple of years later when the Microsoft Windows system administrator powered it off and back on because his MS Exchange server was having problems... Because of him, the router didn't make it to ten years' of uninterrupted uptime.
Many years ago I worked for a security equipment distributor, I used to sell a lot of racks for DVRs, servers, storage and the likes.
We sold 3 different grades of rack castors, 300kg, 600kb, 1200kg.
Everybody bought the 300kg wheels, reasoning they would only be used once to wheel the rack into place, and that would be fine.
One time we got an order for a lot of gear that would completely fill a 42u rack, UPS, lots of storage, 4 servers, switches and power distribution units, about £250k worth of stuff to go in a casino
I totted up the weight from spec sheets and it was 1,100kg, so natural the customer decided to shave £100 off the £250,000 bill by opting for the 300kg castors.
Last I heard, they had to get a forklift to move the rack as the castors collapsed before they could push it into place...
I once did a building move for an outgoing tenant, who insisted on taking the telecoms cabinet with them. Condemning the new tenant to have to trace & re-punch every cable. At least the tight-arse let me gently pull the wires out of the punch-down blocks rather than cut them at the floor so the new tenant had to do a full recabling job!
Speaking as a mechanical engineer, I've encountered the castor problem. For some reason, *some* manufacturers of castors design them to withstand the exact stated load, with no safety factor added to the design. The result of this is that when you use them to support something at their rated load, then push the something over a small bump in the floor, the load on the castors is briefly increased and the castors fail.
Every mechanical design should include a safety factor, the magnitude of which depends on what the design is for, and sometimes, who will be using it!
For this reason, if you're buying castors, either buy some with at least twice the load capacity you think you need, or quiz the manufacturer to find out whether they've included a safety factor in the design and what that safety factor is. (Again, you're aiming for at least twice the static load you think they are going to support, more if you're going to push the object across "rough" ground.)
I'm assuming it was made clear to customers that the castor grades were relating to the weight of the rack, and not to their individual load-bearing capacity? Otherwise it's not hard to understand why a customer might think 4x300kg castors would be good enough to support an 1100kg rack...
This was stated on every quote I issued for an order that included a rack, because one particular customer did try to claim 4x300kg castors could carry a 1,200kg cabinet, I pointed out the text in our catalogue and on our website did advise that 4x300kg castors were for a cabinet with a total weight (rack and equipment) of 300kg...
"it was not a necessary effort. As he set things right, he realized the rack would not have toppled beyond the slight angle at which it rested – because the non-standard steel plate would have held it in place."
That's pretty hilarious!! I would have thought that in such a situation the obvious step, once the initial fall is halted, is to gradually push less and less to see to what extent the heavy object in question can support itself, in which case the standards-obsessed boss would have fairly quickly realised that he could abandon his post at least for the length of time necessary to call for help. I guess he was just in shock-mode and couldn't think it through
《Yes, another example of Chesterton's Fence.》
Or in Terry Pratchett's canine latin of Discworld
"Si non confectus, non reficiat" - family motto of the Vetinari.
《standardized perfection》
Any useful standard ought to be prefaced with the Patrician's motto.
When you think about the essential (and insane) concept of perfection anything, process or system etc once it obtains perfection must necessarily be unique to the particular instance which I would think is the antithesis of standard(ized.)
Standardization is formalizing the art of the possible not aspiration to perfection. Engineering v Theology. :)
Antoine St Exupery probably had the most sensible approach to perfection:
"Il semble que la perfection soit atteinte non quand il n'y a plus rien à ajouter, mais quand il n'y a plus rien à retrancher." -Terre des Hommes, 1939.
Here, in the extreme, perfection would be the complete absence of anything - the (philosophical) void (sans vacuum fluctuations.) He was more practically claiming a minimalist approach to design would be more likely to lead in the direction of perfection.
It would be harder to hold up (or push back into place) the more of a tilt angle there is. If (you think you are) at the limit of your strength merely holding it where it is, and can't get it more upright not matter how hard you try, ... can really you afford to ease off and risk it it tilting a little further as a test?
at the limit of your strength merely holding it where it is, and can't get it more upright not matter how hard you try, ... can really you afford to ease off and risk it it tilting a little further as a test?
Depends whether it's a motorbike and whether you own it..
Had my bike nicked once (lovely VFR750 - probably my favourite ever bike) and got given a Pan European 1300 as a loaner by the insurance company. It's *big* lump of metal and plastic and, even though a lot of the weight is fairly low down, once it starts to go over it's pretty difficult (unless you are Popeye with a handy can of spinach) to hold it up, let alone righting it.
I came close once (stopped on a verge, put the stand down only to find that one side of it sank straight through the tarmac.. Fortunately, I could hold it up long enough for my friends to help me get it upright again (once they'd stopped laughing anyway).
I thought it would handle like a pig but, assuming you use a slightly different technique of moving the bike under you, surprisingly nimble.
Righting a heavy bike that's toppled over is not for the faint-hearted, nor for the weak of back. Simply trying to lift it up is a recipe for a slipped disc or worse. The correct way is to crouch down with your back to the bike, one hand on the bottom handlebar, the other on the bottom passenger grabrail. Then lift using first your legs, and then once its a bit up you can lean your whole weight into pushing it back. Arm strength can help but not usually necessary. Gently gently so it doesn't topple over the opposite way!!
And of course, a lower centre of gravity helps a lot
> The correct way is to crouch down with your back to the bike, one hand on the bottom handlebar, the other on the bottom passenger grabrail. Then lift using first your legs, and then once its a bit up you can lean your whole weight into pushing it back. Arm strength can help but not usually necessary
LOL. Good luck with that. If the bike lies flat like say, a Yamaha Tenere, or an Energica SS9, you're screwed, unless there happens to be a couple beefy police officers nearby.
If the bike has luggage or is a BMW with large cylinders and a crashbar, it'll be upright enough for the above technique to work. That 30 degrees or so makes a big different.
The Energica is especially bad, as being Italian[1], it sits near to vertical as makes no difference, and I had to chop 10mm out of the kickstand to get it to lean and not fall over in a stiff breeze.
[1] could be worse. A lot of Ducatis have a SPRING-LOADED kickstand that flips up if you look at it funny.
>Righting a heavy bike that's toppled over is not for the faint-hearted,
Ah, you've reminded me of a late, lamented dear friend of mine. He was of quite short stature, elderly, smoked like a chimney, but had always, his whole life, hankered after a big touring bike. So one day I drove him into town to pick up his new (to him) Goldwing. Probably a '93 vintage or thereabouts, so pretty chunky, with hard shell panniers & the works.
He was very confident that he could ride it, on the basis that he'd ridden bikes in his youth (i.e., BSA Bantams, circa 1960...).
As you can imagine, a Goldwing 40 years later wasn't quite the same. I followed him back home in my car, at a cautious distance.
My abiding memory is that at every traffic light we stopped at, he'd stop, then like a domino, he'd teeter for a moment and, almost gracefully - but not even bothering to make a futile attempt at stopping himself - he & bike would topple majestically over sideways. So I'd have to get out of my car, lift him + bike up (I was younger then too!), hold him steady till the light went green, give him a push to see him moving, then leap back into my car to follow.
Bless him. RIP Merv. Here's a pint for you, you old sod.
(Warning - name dropping post.)
I was chatting with Prof Robin Gandy* at a conference once and discovered our mutual interest in motorcycling. He had owned a BMW and 'dropped it' going round a roundabout slightly too quickly (us Mathematical Logicians are a wicked 'devil-may-care' bunch you know). Took off the cylinder head.
*Alan Turing's only actual research student.
I saw him at rossotti's a couple times when he was at Stanford (1968ish). At least once he was on a ratty old airhead with side-car mounts, but no sidecar ... apparently someone from Berkeley had loaned it to him for the duration.
The world was a very different place back then.
Me? I was still riding my childhood Bulltaco ... 125 turned into a 360. Death machine.
I suppose there's also the risk that it looks stable for a while, but the stress on whatever is now taking the weight eventually wears it and causes it to tilt further. I've seen it happen with wood, but it's probably much less likely with harder floor materials.
I'd think that the most obvious step is don't move racks unless you have confirmed that you should, and maybe don't do it late at night unless there's an emergency requiring it.
... we had a magnitude 5.7 earthquake on the Calaveras Fault near Coyote Lake, just South of San Jose. I was working for a laser company at the time, assigned to one of their manufacturing floors in Mountain View. When the quake hit, I was standing near a rack full of completed, but not yet boxed for shipping, laser levels. About 3 dozen of them. As everybody screamed and ran for the exits, I noticed the shelving was rocking precariously ... So I stupidly leaned on it, and managed to keep it from falling over.
I got a right royal public bollocking for it. I should have screamed and ran, I guess.
I also got bumped three levels in pay grade, not that anyone I worked with was allowed to know ... Seems that The Boss of the division knew about the shelving not being tied down after a company-wide insurance inspection, and had reported to The Board that it had been fixed. If the gear had hit the deck, he'd have been in deep shit. The following weekend, he and I bolted it down properly ... off the clock, but he bought lunch and slipped me a couple hundred bucks.
"All you kids who just sit and whine, you shoulda been there back in '79 ... " —TRB
A long time ago I worked for a world famous broadcasting corporation based in Britain, in telly outside broadcasts. At the time we were using Philips LDK5 cameras whose base ends had two cameras worth side by side in a ninteen inch rack with their hefty power supplies under the desks, 3U above ground level. These had protruding push button on/off switches. Being Philips they provided guards around them and being Philips they usually fell off. The spring was nicely strong and they didn't change state until they were released. This is important.
We were doing a routine Match of the Day one weekend and midway through the first half a colleague who was sat at his desk stretched out his legs and we all heard an ominous click. "Don't move!" we yelled at him. The sole of his shoe had depressed the guardless on/off button and he now faced the rest of the half holding the moderately strong action switch pressed in . Oh how we laughed.
To his credit he lasted the rest of the first half without releasing it but looked in some discomfort as he limped off for his half time cuppa.
The sole of his shoe had depressed the guardless on/off button and he now faced the rest of the half holding the moderately strong action switch pressed in . Oh how we laughed
Many years ago (very much pre-rackmount servers), we had a bench full of OS/2 LAN Server (remember that?) servers (in sort-of-desktop cases - not even tower cases) on.a bench. The one that handling the MSMail postbox shares was by the room door, with a monitor on top.
One day, I was walking out of the door and I noticed that the monitor had been left on (this was when screen burn-in was a thing and also, if the monitors were left on, the room got quite hot very quickly since it was a glass-walled room that faced mostly south-west. There were blinds but they were mostly to stop people seeing in clearly.) so, as I opened the door and started to walk out, I pressed in the monitor power switch - or what I thought was the monitor power switch (like yours it was a press and release type). Then realised that the case under my fingers had the edge *above* the switch and not below. The servers and the monitors (both from IBM had the same push-to-release power switches with the server one being about 5cm directly below the monitor power switch..
Cue some frantic calls to my colleages to inform people that MSMail would be down for "maintenance" and, while I stood there holding the switch in, to shut down the server involved (and the MSMail POP server since it really didn't handle the postbox share going away at all gracefully..
Service was eventually restored. And we went round taping stiff cardboard flaps over the power switches on the servers, just in case. Fortunately, my manager at the time had the attitude of "it would have happened eventually" and there were no negative consequences (apart from that we carried on using MSMail)
I think I've told the tale before, but it's Friday...
Back in my PFY days, boss and I decided to clean up the cabling on our server rack. The "server rack" was a console desk style rig with three shelves of servers. The "servers" were whitebox tower style PCs, mostly running Novell 3.11 on Pentium or PII era hardware.
Most of the PCs were AT style. For any youngins here, that meant that mains power went directly to the power switch on the front of the case. Press the button in, you'd hear a click, and the power would cut as soon as you released the button. This becomes important later on.
The back side of the rack was a rat's nest of cables due to years of ad-hoc additions. The plan was to pull an all-nighter, and systematically take down the servers, pull all their cables, and recable with some semblance of order. Boss started switching the KVM to correct server, downing the server, physically powering it off, and calling out which box was ok to pull. I was on the back side of the rack cutting zip ties and pulling cables.
We got a good groove going for the first few machines, and then I heard the initial click of a power button, followed by "oh shit!". Boss had pressed the power button on a server that was still up and running. Luckily he caught himself, since Novell didn't always recover gracefully from sudden power loss. Unluckily, the server was just out of reach of the KVM and its keyboard. More bad luck, the server behind his index finger was responsible for payroll.
He asked me to drop what I was doing, come over to the front and down the server so he could let the button go. I tried to start an impromptu renegotiation of my salary. He countered with an offer to not break my kneecaps if I didn't make him risk crashing the server. I had read enough of the BOFH to know that my position was untenable.
A few years ago we had the same problem but got round in a different way, this was also an as400.
We picked the space between the concrete sub floor and the tiles with teams of a4 paper. Under compression it is very strong and packed in tightly it should have been fireproof as well (that didn’t get tested when I was there)
Was working like that for a few years before I left the company
Try putting a fully loaded Juniper PTX5000 in a 800 wide x 1200 deep rack on the raised floor of a first floor data center which wasn't designed to house such a beast, now add a couple more and a couple of Ciena Core Directors and you now have a situation that we have to get approval from the architects (who are busy designing reinforcement of the support of the entire first floor) before we install anything (even a Cisco 4331!), still, consider the cost savings from the sites we closed and consolidated into this site!
... of not asking anyone else why the rack was different, and just trying to 'fix' it without doing any investigation or, and when we talk about heavy engineering the all important, performing a Risk Assessment. Plus, never move heavy engineering on your own. Had it toppled over onto him this would be an obituary rather than an amusement.*
My friend 'Steve' was often heard on the telephone bemoaning the lack of talking to the people who actually worked at a place before 'mending' something causing untold distress, damage and customer complaints. (When you turn off the power to a set of 24/7/365 servers that are in continuous use, clients do tend to notice.)
*There is a difference.
Yes, you'd have thought someone who had "bought into ISO standards, big time" would have had a tick box labelled "Conduct Risk Assessment". Of course everyone knows the words printed against tick boxes have no meaning, they're just to decorate the page before you add the ticks (and why don't they print the pages pre-ticked?).
Even so, it sounds more like a case of OCD than real adherence to standards.
"Of course everyone knows the words printed against tick boxes have no meaning, they're just to decorate the page before you add the ticks"
Yeah, I remember the discussion with the manager of the local dealer that serviced my car when I queried how well I could trust the service checklist which claimed the technician had "cleaned/greased sunroof rails" and "topped off automatic transmission fluid", neither of which my car had. He'd just ticked every box on the sheet. I wasn't oo bothered since it was a company car and I wasn't paying. The second time, I was more vocal in my complaint. The 3rd time, our fleet manager escalated it to the manufacturer and it never happened again. It even looked like the ticks had been done at different times during the process as they no longer looked identical all the down the page and there were dirty fingerprints on it :-)
The unit I worked for had a very large library collection of maps stored flat in map presses, which were steel cabinets with large, shallow drawers to store maps, each press was about 4' wide, 4' deep and 4 ' high. The cabinets on their own were impressively heavy; when filled with maps they were immovable. We had a group of 6 of these arranged as an island in the middle of the room..
Originally our map library was on the ground floor, but the unit moved to an upper floor. We moved the map presses (or rather, a removal company did so), and all was well - until I noticed that presses that were in contact at floor level had an inch-wide gap between them at the upper surface! I made a hasty phone call to the people responsible for building services, which got a VERY long considering pause! It turned out that the floor was strong enough, but the weight of the map presses was enough to deflect the floor slightly. But I think it gave the Building Services people a bad few minutes!
As the top of the map cabinets was used as a working space (you need lots of flat surfaces when working on maps) rearranging them next to the walls wasn't an option. They were still OK when I retired, about 20 years later!
Many years ago I worked in the Caltex data centre in Sydney which was on the 4th floor of the company’s them HQ on Kent Street (which is refurbed as apartments - I still chuckle every time I cross the Sydney Harbour bridge as I wonder how many millions it costs to sleep where my desk used to be). We embarked on a huge project in the early ‘90s to automate operations of the 3090 (if you don’t know what that is you’re too young to appreciate this story), which included replacing the 3480 cartridge drives with a StorageTek silo. It was craned it through the windows one weekend and placed on the reinforced raised floor at a very specific spot, which wasn’t exactly the most convenient in terms of everything else in the computer room. That spot was however right in the middle of 4 concrete columns that ran from the bedrock all the way up the 10 floor building. It was only spot that would hold the silo once it was loaded with the 6000 tape cartridges. Anywhere else and it would have ended up in the basement - one floor at a time.
I worked For a fruit-based phone maker. They used a mix of [Dirty flasher coats] and PCs to run their network.
High up manager comes in, decides to ethnically cleanse all (as he put it) "the windows filth" from the datacentre. For some reason he thought the PCs were there for people to 'play games' or something!
He unplugged various cables and manually removed the Twin PC servers.
Whole network including iCloud went down, as those PCs were being used to syncronize everything else, controlling load balancing, scheduling backups etc. They also ensured if you hadn't accessed your icloud data for a long time, it went to slower storage. (which is why if you haven't accessed stuff for a few months it can take a little bit to appear).
Whole system down, and the way he'd removed cables he'd damaged the motherboards, fibre connectors AND the ports (think angrily RIPPING them out).
Entire system was down for 1.5days until engineers could replace the PC motherboards and get everything up and running again.
Explanations? I want to see the psychiatric report on the manager. Someone who sees an in use computer system at work and literally rips out the network cables without asking what it is for needs a long lie down in a darkened room with soothing music (or something like that).
Sounds like an extreme example of Not Invented Here (NIH) syndrome, where you do not use anything your own company did not develop itself. Except for pencils, pens, paper, mugs, knives, forks, spoons, sporks, books, clocks, desks, chairs, telephones, language, money, ... motor cars, cardboard boxes, spectacles, clothes, torches, keys, fire extinguishers ... vacuum cleaners, toilets, sinks, soap, towels, windows, walls ceilings, floors, electricity, air conditioning ...
and pretty much everything you find described in the Encyclopaedia Britannica. After all it is not as if the Romans ever did anything for US*.
*If you don't know, shame on you.
Hmmm, as the processor AS400 was build by Mother Blue, I suppose the disc racks were also by the same Mother. And the Installation Manual by this company is know for very, very detailed installation guides (e.g. they tell you to way how one is supposed to fasten the screws), which explicatly show how to have the correct sheets under the machines, depending on the size of the machines.
Heinrich
Large Oz Government Department with large Amdahl mainframes. Water cooled. Being government they took the cheapest quote for the water cooling system maintenance. The company that got won the quote, still in business I see, found they could keep costs low by not doing any maintenance at all. So filters not regularly changed per spec. Eventually the pipes clog and freeze then burst spraying water directly at the side of one the Amdahls, which kept on working fine but was powered down before the water went above the false floor. No come back on any party...
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
