Server crashes traced to one very literal knee-jerk reaction A weekend of unwinding is behind us, so The Register returns to work on Monday with a fresh installment of "Who, Me?" – the reader-contributed column that reveals how you got in a tangle, and then extricated yourself. This week, meet a reader we'll Regomize as "Carter" who shared a story from the latter half of the 1990s when …
Ehlers Danlos I wonder.
I also wonder about anyone who has a server with such an easily accessible reset button, even back in the days of NetWare and "built from parts" servers I'd never connect the reset switch unless it was well recessed or covered because *nix and Novell didn't cope well with unscheduled resets
Seriously, downvotes for common sense cover up the reset switch FFS.
It's as if you lot have never learned that people accidentally press switches and reset stuff if they're not protected or disconnected.
And Ehlers Danlos is hyper mobility, the condition is potentially why the gangly youth could hit the reset switch by accident when it seemed so unlikely. Us, the EDS society symbol is a Zebra
I once (briefly, thank FSM) worked for a small, independent finance software company. The "IT Room" was a bunker in the basement full of boxes of old, knackered bits and cards. And the production server which was a tower unit. Situated next to the door. The power button sticking out nicely as if to say "press me!" It was not covered. I managed to hit that power button accidentally, downing the whole office, twice.
Despite my requests to be allowed to cover the button it remained prone when I departed several months later. Because they, er, didn't want me to make any changes to the server.
While I agree with you, and have practiced what I preached over the years (unplugging the reset button on PCs is simple, and cures many ills), try to remember that back a third of a century ago (or thereabouts) Microsoft was trying to convince TheGreatUnwashed that they could purchase a computer from any old shop on the highstreet, plug it into another computer, and be instantly running a network. No need to buy a "server", no need to hire a CNE (or similar), just plug it in and it would work. Modern Computing in 1992 was Wonderful with Microsoft and Intel running Windows for Workgroups!
So TheGreatUnwashed did[0] just that. But all those highstreet computers had reset buttons. And the owner/operators had no clue that the button could and should be disconnected when the machine was placed into the server roll that it was usually unsuited for to begin with. Thus all the stories we hear today about all the wild and wacky ways that the users found to reset them, usually at the most inopportune time.
[0] And, as we all noted at the time, it quite often didn't. Some of us made a living from this discrepancy.
The up side to those days was that most manufacturers, including the big names, were using standard form factor boards and cases so disconnecting the reset switch was fairly easy to do, custom motherboards ce quite a bit later.
I also bumped into malicious resets occasionally where some idiot would think it funny or necessary to push reset or otherwise disrupt a server, I solved a couple of those with CCTV cameras and locked cabinets
Yes that's another point that Jake didn't cover. The wild 90's were the times when it was common to reboot your computer many times per day to clear any and all kinds of unresponsiveness. And so people assumed that since it worked for their personal computer it would also work for the machine that was promoted to "server" class.
Which reminds me, in the mid-90's I was working for one of the major wireless telecoms companies, deploying the latest CDMA technologies. The system was not completely hardened or mature, so we hired humans to keep an eye on the controller and cell site equipment, prodding it if it fell asleep or stopped responding. These systems were built on top of Solarix/Unix. It seemed that in one of the cities we'd have about a half hour each morning when phone calls would sort of stall. Turns out the operator (ex-IBM machines) would issue a command to the system controller and if it took too long for it to come back he'd just reboot the whole thing. Our uptime and network availability improved dramatically after we let him go.
a server with such an easily accessible reset button, even back in the days of NetWare and "built from parts" servers
I've seen generic "white box" cases with reset buttons which protruded, and some for which you needed a pencil or ballpoint pen to activate.
PC keyboard-lock keys locked out only the keyboard. Minicomputer lock keys could lock out the entire switch panel, including HALT, CONT(inue), START, LOAD ADDR(ess), etc., and most-importantly, the power on/off function.
They allowed you to do any switchery needed get the computer up, then in the PANEL LOCK position, remove the key.
"Where can you find a paper clip to reset/default that box nowadays?"
Pencil tray in the top center drawer of my desk. Also, most of my toolboxen will have various sizes of paperclips rattling around in them somewhere. The desk down in the kitchen has a small square plastic box with a round magnet at the top that is full of various sizes of paperclips. Refills are in boxes on the "junk shelf" in the pantry, along side binder clips, various kinds of push pins and thumb tacks, staples, household glue and various kinds of tape, post-its, new decks of cards, a couple boxes of strike-anywhere matches, half a dozen types of string, toothpicks, and the like.
Someone I know *cough* ran an fsck on a 300 mb washing machine sized disk drive once then mistakingly pressed the power button to turn it off while the system was still running. Thankfully the console was within reach to shut the thing down before releasing the power button, saving the potential 30 minutes or so to do the whole process again.
I was 7 or 8, younger brother couldn't figure out why the family PC would sometimes turn off at random while he was using it.
The PC had a AT-style on/off push-button switch on the front, I remember you had to power it off with that after initiating shutdown in Windows. It was a pizza-box style design (think the form-factor was LPX iirc) and had been placed underneath the desk on the floor by our father.
Turned out when my brother was sat on the chair this put his big toe at the exact height of the power button, and when idly swinging his legs as kids do he'd power the PC off.
One of my previous tower PCs had a power switch right on the top of the front panel itself. The manufacturer had carefully designed the top of the panel so that it was slightly lower than the top of the PC case itself. This meant that placing anything on top of the PC wouldn't accidentally depress the power switch.
Of course, they never factored in cats jumping on top of the PC and switching it off!
No-one expects the feline imposition...
...aka, even if they *had* attempted to factor this into the design, it's pretty much a given that sooner or later the tenacity of the average feline to figure out ways of disrupting your work would win out.
Yes, I speak from experience...
Ah yes, cats.
That's why all the UPS's in the house have a piece of clear plastic taped on one edge to cover the power buttons on the top.
And on the home PC, the power strip with the main cords going to it has a Velcro strap holding several cords together to protect the power switch from the mouse. One time the PC's mouse fell off the back of the keyboard tray and hit the power button dead center.
Yeah, I need to manufacture custom mollyguards, or kittyguards actually.
In my Harry Potter datacentre (space under the stairs with router, switch, NAS, two Raspberry PiHoles) there's a UPS and a powerstrip with a switch.
Felines on murine patrol have on multiple occasions managed to switch off either of those.
I was once showing a group of international students around our high performance computing centre at the university. When we came to the Cray J932, with its impressive large green rectangular power led, and beneath it a red, very well recessed reset button, one of the students asked what would happen if he pressed that button. I stated a small metal claw would emerge and snip his finger off, and if that failed, I would remove the offending digit with pliers.
He decided not to test this.
I was just putting the finishing touches on a small cluster of vaxen at SLAC one fine Friday afternoon. The annual Big Game between Stanford & Berkeley was to be the following day. A couple of grad students started passing a football (American version) between themselves. In the glass room (that wasn't glass). Just as I was threatening mayhem if they didn't knock it off, the ball hit the Big Red Button. Needless to say, a bunch of very pissed off people couldn't attend the game the following day. The grad student's computer privileges were suspended for the rest of the academic year. Personally, I'd have hung them by the thumbs in the Quad as a warning ...
Special equipment required if the culprit is Buster Gonad (the boy with unfeasibly large testicles)!
You were lucky. If it was anything like the Honeywell-Bull console Emergency Power Off switches, it would have taken an hour to fix it because the entire console had to be disassembled to reach the switch to reset it from the inside of the cabinet...Ask me how I know (and no it wasn't me, but it was an approved action taken under direction of the system administrator because, hey, we'll never have this opportunty again) Hi Dan.
In my first job I was in charge of repair as well as Hw/Sw development. As an Epson dealer this meant printers as well as computers. There was a particular hinged printer part (I think it was a hinge, but I can't remember exactly) that made a perfect Molly Guard and the company owner and I discussed fitting it to all computers we supplied.
One of my first IT jobs as a contractor was at Norwich Union, primarily there to sort surplus IT equipment out & backfill.
First day on the job, twice within 30 minutes a whole line of PC's (They may have been dumb terminals) went down.
The root cause was daisy chained power bars & one ladies foot that liked to tangle up the power cable & pull it from the wall, how she managed this twice on that day of all days, but never before remains a mystery to me.
Repositioning the power strips resolved the issue.
Ah yes, the dependable Molly Guard. I have one on my power bar to prevent accidental turnoff of the main computer, and a separate one upstairs to protect the CCTV cameras.
I even used to have one on the keyboard to protect against accidental autoloads while gaming. Oddly enough, I never needed one to deal with my own "Molly" when she was a toddler.
(resident gangly youth) RGY would be variation on PFY - the BoFH could augment his team to have both. The PFY could open the window for bit of fresh air while the RGY uses his wayward gangly limbs to propel the current stinker into the carpark.
I assume the coroner by now has a custom form for Simon's workplace.
A frisbee being launched across a machine room that just happened to hit the large, unshielded red power button on the PDU at the other end of the room. Boredom on a night shift is a dangerous thing! Happened many, many years ago, and I was on the following shift and had to help pick up the pieces.
Many moons ago I took my daughter to SLAC on take your kid to work day. At the ripe old age of 9, she had been there many times before and knew the ropes, but I figured she deserved a day out of school.
She told me as we were walking in that it'd cost me ten bucks for her to not push any buttons. I gave her the money.
On the way back out, I told her that it'd cost her ten bucks for me not to tell her mother she was running a protection racket. She made a face and paid up ... and promptly told her mother as soon as we got home. They both still laugh about it. So do I :-)
I chose not to take either daughter into the machine room at work when they were growing up, even though I really would have loved to have shown them around.
Sadly, one would have been utterly compelled to press the forbidden buttons (and still would, as an adult) and the other would have been compelled to rub it in her sister's face if she'd been the only one allowed in.
Kids.
when at crawling/cruising age loved to press bright coloured or illuminated buttons when she could reach them - washing machine, dishwasher, child seat restraint release, twin servers with light blue front fascias and orange backlit on/off switches I used as fileservers (lots of drive bays). One day I heard a loud bang followed by a single cry from her. I couldn't place her location immediately - it seemed like moments that see was sat between my feet watching childrens cartoons. Whilst trying to find her I next heard tiny fists pounding metal and saw a sight that made me laugh and cringe at the same time. She was standing at the front of one of the servers, the one still working, pressing the power button listening to the fans all zoom up, then turning the machine off which went with a loud clunk and the fans wound down - repeat. The room had a smell of burned electrics I surmised was from the now dead machine which was no longer of interest. I calmly walked over and picked her up with the biggest of forced smiles and said "There you are. You okay?" She pointed to box and said "G'ang" - I interpreted as "go bang"! I distracted her with a different TV channel and some milk and she fell asleep. Used the opportunity to check out the machine and it was toast - motherboard and PSU burnt out. Both boxes were saved from being dumped after an office closure, and had travelled around the world with me for a few years afterwards. It took my little girl less than 5 mins to kill one of them.
My first exposure to business computing was when I worked for a major high street retailler in the late 80's.
Our tills and stock system was tied together with a couple of mirrored Compaq PCs acting as servers.
This system was pretty reliable, with very little downtown until suddenly the servers started glitching, sometimes it was the one on the left, sometimes the one on the right. The glitches would void any transaction happening at the time of the glitch. These glitches rarely resulted in a full scale crash, just a minor hiccup that cancelled any inprocess jobs on the database.
The second time this happened we called in the IT bods from Head Office. They ran tests, installed power conditioners, checked logs, scratched heads, but the glitches never occured when they were on site, so they couldn't trace their cause.
They called in Compaq's engineers, who did the same tests, and couldn't find any cause, until just as they were about to pack up, a glitch occurred.
One of the engineers was looking towards the door as it happened, and saw the possible cause; the store manager walked past, and as she did, her analogue cordless phone rang, and she answered it, the EMP, small as it was, seemed to be enough to interupt the nearest server.
The solution, swap her phone for a new DECT phone...
Picture a data center in the basement of a tall building in San Francisco's financial district. Card punch up against a wall, near the ancient Otis heavy goods lift. Every now and again, at seemingly random times, the punch generated errors for a couple characters. Nobody could figure out why, not even IBM's field circus dudes.
Until IBM was traipsing in and out one fine weekend, upgrading who knows what hardware, as only IBM could. Someone (ahem) noticed that the gibberish was being generated about ten seconds before the elevator doors opened.
Turned out that the motor for the lift was drawing so much current when it first started that it was inducing errors in the punch on the other side of the wall. Nobody put two and two together prior to this because the lift rarely went into the basement (that level was key-protected) ... until IBM was in and out that morning.
Once I figured it out, and could reproduce the problem at will, a little shielding (spec'd, provided and installed by IBM, gratis!) made it go away permanently.
Big power supply with lots of Caps to store a charge for a echo sounding boom being towed by a trawler up & down Gare Loch performing a survey for Faslane who wanted to put degaussing coils into the Loch for subs to travel through.
Survey over one of the team goes to hit the main shutoff\isolator for the power supply with a wooden boat hook, the boats owner commented.
"Och aye, Wassa matter wi' him, is the fella scared to turn it off by himself!"
"Damn right he is, there's enough stored charge in that thing to kill him, if it arcs over!"
"Ohhhh!"
Icon - Details may be a bit vague as it was 37 years ago, I was watching the boat go up n down the Loch from a hotel bar at Gareloch head most of that week.
I designed and built a DSP board and software back in the days when 50MHz was a decent clock speed and was desk testing it before it went into a shielded metal box (for integrating into an F1 Ferrari test rig). I could only get it to work perfectly on a Sunday. Come Monday morning when the company owner came in, it would throw random bus errors. We made a trip around the local industrial site together. It turned out that someone was building unregulated Taxi radios next door and they were the source of the interference. They immediately put up shielding and asked us not to mention the issue. (I think that the company owner also gained a brown envelope to cover my 2 wasted weeks of debugging.)
I'm sure many have experienced similar - so, Route 1 is always "is everything properly plugged-in?"; I've 'fixed' a colleague's PC by doing just that.
Every time they made a substantial movement they pulled on their desk, which pulled on the power lead trapped against the desk divider, which was only just plugged in to the floor socket. A gangly, 'Yoot' hitting a reset is no surprise.
As an ex service engineer I have a couple of classics..
One from I can't turn it on.. went in with a new PSU switch and opened the door flap (remember those on the Compaq towers) and powered it up & down. I asked it it was an intermittent fault, it was a didn't RTFM fault and he didn't know that the door existed.
so many kicking the power lead out from the power brick on the dell small form factor PCs
but the best has to be the phone that kept hanging up....
(I was there at the beginning but had moved offices for the end so was told the resolution)
It turned out that the phones were extremely sensitive to static, and if you moved in the chair it created a small static charge that wasn't enough to feel, but was enough to hang up the phone.
The solution was to put the chair in the skip
Sounds like an extract from The Hitchhikers Guide to the Galaxy when:
"spirits were brave, the stakes were high, men were real men, women were real women, and small furry creatures from Alpha Centauri were real small furry creatures from Alpha Centauri. And all dared to brave unknown terrors, to do mighty deeds, to boldly split infinitives that no man had split before"†
One might wonder whether Carter was employed by the telemarketing division of the Sirius Cybernetics Corporation.
RG58 aka coax, thinnet, 10base2 — a whinnet·ridden‡ technology that was a perpetual whack·a·mole of who in the coloured pencil department had disconnected their PC/Mac by unplugging the two RG58 cables (to the T-connector attached to the PC's network adaptor) from the wall and bringing their segment down.
For a fair while after 10baseT became de rigeur I still carried a couple of 50Ω terminators in my pocket. The clown that substituted 75Ω TV aerial coax (RG59) deserves, amongst other things, special mention. I never could get a coherent explanation of why or where got the cable from—not that the chap was ever distinguished by coherence at any time.
† Fit the Third ‡ Fit the Fourth (HHGTTG)
Ah the joys of coax cables.
We had a bit of kit that used coax video feed. Software devs report it as having vision issues, with interference on the screen. My boss picked the job up and after about a week of trying to diagnose things and swapping out a few PCBs was having little luck fixing the issue, so I offered to be a pair of fresh eyes. Noticed the coax cable wasn't a standard part for the unit, swapped it out - and all the vision problems went away.
The devs had a stash of the non-standard cables but no idea what spec they were. "We thought they were fine as the connector fitted and it worked when we tried plugging one in." The issue was resolved in the end by cutting the ends off the cables and chucking them in the skip.
And the Yoof of today, my three kids included, although the girl/young woman is a lot more tech savvy than her brothers, are fond of connecting cheap UK plug type power supplies to their power hungry laptops with a USB-C cable designed for a phone or, in one case an electric toothbrush, and then wonder why it won't charge their laptop or connect their backup drive!
I never could get a coherent explanation of why or where got the cable from
75ohm cable with BNC connectors could also be from video camera/monitors, so perhaps collected when left "unwanted" in a conference room?
Unfortunately the centre pins on 50 and 75 ohm BNCs are different diameters, so tend to either make poor connections or stretch the female socket centre connectors when plugs & sockets are mismatched.
CCTV video cables are 75-ohm RG-59/U, and use 75-ohm BNC connectors. They will mate with 50-ohm BNC connectors. Not perfectly; ISTR that the center pin diameter is a bit different.
The failure that results is less obvious than when you shove a USB B plug into a 10baseT ("RJ45" - NOT!) socket. Or insert a VGA plug by feel, upside-down. The pin layout is symmetrical, and there is enough 'give' in the D-shaped male shell that it will re-mold itself to match the female that you're inserting.
Or coming in on Monday morning to find that department "x" had decided to alter the office layout.....in a room where the power/network points were in recessed boxes in a concrete floor.
Of course they were no longer in the right place so there were random network & power cables stretched between desks not only bringing the network down but presenting a trip hazard!
They had to put it back to how it was before.
When we were decommissioning the old Fabian Avenue telco Central Office (now home to the Charleston Village condominiums), I was given the task of making sure the electrical power to the site was off. Not just at the breaker down at the street pole, but the physical breaker at the Colorado Avenue Sub Station in Palo Alto was to be pulled, thus making certain all power was deactivated until we could make certain everything was isolated.
This would involve taking out the entire Charleston Gardens section of Palo Alto for an hour or so mid-morning, mid-week when it would cause as little disruption as possible. (Charleston Gardens is a mostly residential section of Palo Alto, bordered by Middlefield, San Antonio and Charleston roads, if you care.) The neighbors were notified the week before, both by snail-mail and people physically knocking on doors to explain and hand out the small paper notice explaining what and why.
Come the morning of the Great Event, I was selected to physically make sure the power was off at the sub-station. Cell phones being a fad of the future, I drove down the Frontage road (West Bayshore) and arrived at the the sub station promptly at 10AM. Conversation went something like this:
Me: I'm here to see that the power to the Fabian project is off.
Site Engineer: We're all ready for you, I'll get to it in a second ... or you can just throw that switch (points).
Me: OK (throws switch).
Engineer: NOT THAT SWITCH!
Most of South Palo Alto: WTF‽‽‽‽‽
Management swallowed the story that the main breaker tripping was an un-foreseen knock-on effect of the smaller section being taken down ... The Engineer and I made a bee-line for Fred's (well known dive bar on the Palo Alto/Mountain View border) as soon as he cleared up the problem, which took into the late afternoon. We're still friends.
Well if comes to taking out a Solaris system, then the regonimiser should by default be Rodney (McKay) as he blew up 5/6ths of a star system.
Icon - Pet peeve I do so hate labelling any other planetary system as a solar system, our planetary system is the only one officially called “solar system".
1st week of a new job and we had to move some kit in the comms room.
Which featured a UPS with an emergency power off button that extended about 2mm from the case. Guest what I did…
A few years later had to move a Novell server, moved it and pressed the power button -this was an old Compaq profile int where ther power switch didn’t go off until it was released, took 10 minutes to get everyone off the system - those power switch springs are strong….
Lots of focus on the exposed reset button and the double-jointed junior but not so much on the logs...
None of the team had made any changes that could have made the box unstable, so Carter checked its log files."They showed nothing out of the ordinary," Carter told The Register. "No temperature spikes, no failing drives, no network or device errors."
and no indication that the reset button had been pressed either, apparently. Was that a BIOS setup omission or standard Windows behaviour at the time?
A reset button wouldn't be logged in any way back then. There was no ACPI to negotiate with or notify the system of the impending power loss. The reset button simply interrupted power to the whole computer without notice. As far as the "logs" are concerned, all is normal and then the power is gone.
Logging was also much more sparse back then as we didn't have effectively "infinite" storage to store it all as we do now.
This is what would have happened if Corporal Carrot went in to IT, instead of the City Watch.
One hell of a Helpdesk operator and sorting out any (un)helpful "just helping IT" sorts. Knows the name of every caller just based on listening to the voice, rather than caller ID. No shadow IT nor mailboxes filled with dubious material. Eventually rising to Chief IT Operations Officer just behind a solid, but honest IT grunt being appointed CIO and hating every minute of it.
First up: I wasn't there - but the tale goes that one of our network experts were sent to the Luxembourg site. On arrival to the IT room he opened the door to the supposed wardrobe and hung his coat on what turned out to be the the main breaker. It may be an urban/corporate myth but it did get retold over may years since the late 1980ties
For a couple of years I had an office to myself. And a Win 95 PC that worked well. Half the week I was out in schools so when they increased my staffing by1 they put her in the same office- using the same PC. But she was forever complaining that it kept crashing. Putting on my IT support hat I checked everything I could. Nothing. It worked fine for me. Kept crashing for her. Until the end of term.. Her school that day was closed for some reason or other so we were both in the office. And as I ws typing away she walked past and crash .
So we rebooted and tried again. Every damn time she got near.
The answer; The room had a dry atmosphere and she had a liking for those thick woollen sweaters with lots of sticking out strands.She was a mobile Van de Graff Generator.
I had an office full of such clients. Low humidity, lots of nylon. Typically leave their desk, come back, PC locked up. A very observant client noted a spark at sit-down. I taught them to spray diluted Downy Fabric Softener on carpet and chairs as needed. Augmented with added bonding from keyboard to PC, and even a grounding strap on a PC cart.
Had an incident in a computer room at a stock brokers where there was a power kill button on the wall next to the exit button. After a sparkie managed to kill power to the room while trying to get out, his next tyask was to put a molly guard over the power button :)
> a power kill button on the wall next to the exit button.
US safety code calls for an OFF switch for a gas/oil fired furnace away from the burning furnace, typically at the top of the cellar stair. Which is also an expected place for a light switch so you can go down cellar. This has always been a minor issue, but my father managed to turn-off the heat instead of the lights. Next day I had that switch guarded. (He also discovered, in the prior house, that the fluorescent lamp in the bathroom had erratic ground.)
Back in the late 90’s, I worked for a company in Maryland that was an Intel Solution Provider. This ment we sold Intel white box servers.
I was assembling a particular combination of Intel brand Lego parts when I found a design flaw. The chassis came with power supply installed and cables run to the appropriate locations. That particular server chassis had a 3.5” external drive bay in an odd vertical position on one side. The power lead to the floppy bay had a standard 4 pin molex connector and a standard “floppy” connector.
While configuring the system there was a point where a floppy disk was used to load a license or save a config file or such. When I pressed the eject button there was a pop and server powered off.
Turns out that when I punched the button things jostled just enough that the 4 pin molex connector in the floppy bay bumped into a protruding screw or rivet and sent 12V or 5V to ground as a dead short. I think this us why I started to see rubber/vinyl caps on molex connectors in some systems.
Nothing seemed to be damaged, except the RAID volume was corrupted and the OS would not boot. Luckily the step before the floppy step was to configure and test the tape backup. Not a great way to start a server’s life by testing disaster recovery.
I was once tasked with installing a server in a 'rack' in an Edinburgh university server room as part of an EU project the company I worked for was involved with.
The server room was in the basement and they had what looked like Dexion shelving instead of proper server racks. There was a narrow gap between the back of the racks and the wall.
After placing the server, I began the journey down this gap to start connecting it up. About half way down I felt my backside touch something. They had a PDU with a big rocker switch attached to the back of the rack, which I had inadvertently just switched off with my buttock. Swiftly switched it back on, completed my tasks and made a hasty exit. As I was leaving I heard one of the tech ask "Hey, anything wrong with DNS?"
In the days of the humble VCR when the only remote had a cable (so wasn't used), the cat would press "stop" with its nose.
Inconvenient if watching, a real pain if you were recording!
Not quite as bad as a friend's son, who worked out the slot was just the right size to disposed of that unwanted jam sandwich.
IBM knew about this risk back in the early 1960s, which is why the emergency power off switch on all IBM 360 and 370 computers required you to PULL on a red knob at the very top of the console panel.
But other people at IBM put the LOAD button (which means "reboot" in modern lingo) just a few inches above the desktop where you might bump it, hit it with you coffee cup, or push a telephone against it.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2026
