• Cbuttius (unregistered)

    Not the fr1st time we've had a story about heating this month is it.

  • C-Derb (unregistered)

    Brillant!

    Could be worse, I guess. We just had a fairly serious ceiling leak into our server room that wiped out one of our routers. After mopping up most of the water, the server room door was propped open by a large fan drying out the rest of the room. For about a week.

  • I'm not Dave (unregistered)

    Why would he have been called at 3 in the morning? Because the room was too cold?

  • (cs) in reply to I'm not Dave
    I'm not Dave:
    Why would he have been called at 3 in the morning? Because the room was too cold?

    Actually, since the exhaust was blowing directly onto the thermostat, the AC thought it was too warm. There was probably something watching either the thermostat or the AC cycle, and it notified him that, "Hey, the AC has been running for 6 hours straight and it's still 90 in here!"

  • AN AMAZING CODER (unregistered) in reply to CodeNinja
    CodeNinja:
    I'm not Dave:
    Why would he have been called at 3 in the morning? Because the room was too cold?

    Actually, since the exhaust was blowing directly onto the thermostat, the AC thought it was too warm. There was probably something watching either the thermostat or the AC cycle, and it notified him that, "Hey, the AC has been running for 6 hours straight and it's still 90 in here!"

    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF. The second one is not having a cover over the thermostat.

    I have a feeling there was a "legitamite" (as in not the blamed person's fault) reason Evi was there, but when this could-be problem was spotted Evi used it as an opportunity to rage about having his authority challeneged.

    We've all been there.

  • BOFH (unregistered)

    The real WTF is the Apple-keyboard.

  • Jason (unregistered)

    I'm very glad I don't have a job where I'd get called in at 3am because of a temperature issue in a small server room, when all the servers are still running fine.

  • Network Executioner (unregistered)

    You know what else 25 feet of un-taped down network cable's good for?

    Making a garrote.

  • Ken B. (unregistered) in reply to Jason
    Jason:
    I'm very glad I don't have a job where I'd get called in at 3am because of a temperature issue in a small server room, when all the servers are still running fine.
    So you'd rather wait to be called until after the servers start melting because it's too hot in the room?
  • C-Derb (unregistered) in reply to AN AMAZING CODER
    AN AMAZING CODER:

    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF. The second one is not having a cover over the thermostat.

    I don't understand what you expect a cover on the thermostat to do. The thermostat is supposed to monitor the ambient temperature in the room. If I point a blow dryer (or a computer exhaust fan) directly at the thermostat, how would a cover help the thermostat know that the rest of the room is actually fairly cool?
  • RandomPerson (unregistered) in reply to AN AMAZING CODER

    [quote user="AN AMAZING CODER"][quote user="CodeNinja"] Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF.[/quote] Wait... why buy a second thermostat if the first one can be monitored?

    [quote user="AN AMAZING CODER"][quote user="CodeNinja"] The second one is not having a cover over the thermostat.[/quote] You can't cover it completely, there's already enough of a cover over it? You need the air flow to measure the air temperature?

    Point an exhaust at a thermostat and it will change the temperature. Fix that problem, not the working thermostat.

  • (cs)

    Embedded systems don't run so hot, because sometimes there is no file system.

  • Jaybles (unregistered) in reply to RandomPerson
    RandomPerson:
    AN AMAZING CODER:
    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF.
    Wait... why buy a second thermostat if the first one can be monitored?
    AN AMAZING CODER:
    The second one is not having a cover over the thermostat.
    You can't cover it completely, there's already enough of a cover over it? You need the air flow to measure the air temperature?

    Point an exhaust at a thermostat and it will change the temperature. Fix that problem, not the working thermostat.

    Agreed. Clearly "AN AMAZING CODER" is not an amazing sysadmin. Doing either of those suggestions will not help the root problem at all. A second thermostat might give you a better idea of the room temperature (specifically, measure from two locations to see if the temp is uniform), but you'll still want to know that the A/C has been running non-stop because it thinks the temp is too high. And a cover for the thermostat is going to do nothing but stop a lazy person from changing the target temp on it. Good luck finding a cover that will completely stop heat transfer and also allow it to accurately measure the temp of the room. Physics is HARD!

  • (cs) in reply to Jaybles
    Jaybles:
    RandomPerson:
    AN AMAZING CODER:
    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF.
    Wait... why buy a second thermostat if the first one can be monitored?
    AN AMAZING CODER:
    The second one is not having a cover over the thermostat.
    You can't cover it completely, there's already enough of a cover over it? You need the air flow to measure the air temperature?

    Point an exhaust at a thermostat and it will change the temperature. Fix that problem, not the working thermostat.

    Agreed. Clearly "AN AMAZING CODER" is not an amazing sysadmin. Doing either of those suggestions will not help the root problem at all. A second thermostat might give you a better idea of the room temperature (specifically, measure from two locations to see if the temp is uniform), but you'll still want to know that the A/C has been running non-stop because it thinks the temp is too high. And a cover for the thermostat is going to do nothing but stop a lazy person from changing the target temp on it. Good luck finding a cover that will completely stop heat transfer and also allow it to accurately measure the temp of the room. Physics is HARD!

    But engineering is easy. Put an IR sensor inside the (IR-transparent) shield, focused on the far wall. If you don't have one, go see your local barbeque manic. He'll have one for his grill.

  • Hindrim (unregistered) in reply to Network Executioner
    Network Executioner:
    You know what else 25 feet of un-taped down network cable's good for?

    Making a garrote.

    Even better, hangman's noose for the next morning.

  • Bob (unregistered) in reply to Jaybles

    Oh come on. His name is "AN AMAZING CODER". In caps. Obvious troll is obvious.

  • Yanman (unregistered)

    Evi is a female name. I demand that this comment be featured.

  • n_slash_a (unregistered) in reply to BOFH
    BOFH:
    The real WTF is the Apple-keyboard.
    +1
  • Jack (unregistered) in reply to Ken B.
    Ken B.:
    Jason:
    I'm very glad I don't have a job where I'd get called in at 3am because of a temperature issue in a small server room, when all the servers are still running fine.
    So you'd rather wait to be called until after the servers start melting because it's too hot in the room?
    Yes. It is exactly like the disk-full situation that has played out countless times over my career. Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
  • (cs) in reply to Bob
    Bob:
    Oh come on. His name is "AN AMAZING CODER". In caps. Obvious troll is obvious.
    Do you prefer subtle trolls? We have those here too.
  • VictorSierraGolf (unregistered) in reply to Jaybles
    Jaybles:
    RandomPerson:
    AN AMAZING CODER:
    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF.
    Wait... why buy a second thermostat if the first one can be monitored?
    AN AMAZING CODER:
    The second one is not having a cover over the thermostat.
    You can't cover it completely, there's already enough of a cover over it? You need the air flow to measure the air temperature?

    Point an exhaust at a thermostat and it will change the temperature. Fix that problem, not the working thermostat.

    Agreed. Clearly "AN AMAZING CODER" is not an amazing sysadmin. Doing either of those suggestions will not help the root problem at all. A second thermostat might give you a better idea of the room temperature (specifically, measure from two locations to see if the temp is uniform), but you'll still want to know that the A/C has been running non-stop because it thinks the temp is too high. And a cover for the thermostat is going to do nothing but stop a lazy person from changing the target temp on it. Good luck finding a cover that will completely stop heat transfer and also allow it to accurately measure the temp of the room. Physics is HARD!

    I seriously hope AN AMAZING CODER isn't actually a sysadmin, or we'll end up reading a story about melted servers, an AC saying "Huh, what are you talking about? The room is so cool, I even didn't need to work for the last two weeks..." and a sysadmin constantly replacing "broken" thermostats...

  • Network Executioner (unregistered) in reply to Hindrim
    Hindrim:
    Network Executioner:
    You know what else 25 feet of un-taped down network cable's good for?

    Making a garrote.

    Even better, hangman's noose for the next morning.

    The Darwin Award Nominee doesn't learn anything if they're dead.

    CAPTCHA: tristique - What Mystique was after she tried a threesome

  • Archon (unregistered) in reply to cellocgw
    Put an IR sensor inside the (IR-transparent) shield, focused on the far wall. If you don't have one, go see your local barbeque manic. He'll have one for his grill.
    And then someone will point the exhaust at the far wall or put some new hardware along the "unused" wall and then everything goes haywire for less obvious reasons.

    Change the thermostat to a more complex device, molly-guard everything... you're trying to idiot-proof a server room. It's supposed to be a place idiots are barred from, which is the entire point of the story.

  • Mike (unregistered) in reply to Jack
    Jack:
    Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
    I spent about six months writing a monitoring system that would probe multiple indicators regularly, graphically displaying trends in both response time and error messages, so you could readily see when things were bogging down, spot patterns, and deal with them before you had a full on outage.

    Cut to about one hour after go live, when we were training the operators who would use it. #1 requested feature: an alarm to page us when something actually does go down. They didn't care about anything prior to that, but get an outage, and they all wanted to be the heroic first responder.

    I don't really blame them. Years ago I saw an outstanding sys admin, who anticipated all problems and resolved them before they occurred, get fired because he "never did anything". I learned. The people who pay your salary need smoke now and then to remind them why you exist.

  • Jockamo (unregistered) in reply to Jason
    Jason:
    I'm very glad I don't have a job where I'd get called in at 3am because of a temperature issue in a small server room, when all the servers are still running fine.

    I'm very glad I don't have a job.

  • Secure your Job (unregistered) in reply to Jack

    I can answer that one for you. Job security. If things hum along w/o incident, upper management figures they can cut staff. If there are pages and alarms from time to time, you obviously need the staff to deal with all the issues, even if they could have been avoided.

    This is why it took 3 Unix admins to monitor 4 racks of equipment and yet only 1 Windows admin to monitor an equivalent rack of equipment. None of the unix machines had basic scripts to do things like rotate/compress log files, alert on disk space BEFORE it was an issue, etc.

    When management (for other reasons) finally got rid of the Unix folks, the Windows "team" (1 person) found themselves inundated with pointless alarms that were easily fixed with some scripting to avoid the problems in the first place. Before long, the 1 person was doing the work that had required 3 people to do. (and no, didn't get paid 3x as much)

  • The Del Operator (unregistered)

    This reminds me of the last house I lived in. Being the only computer/tech person in the house I kept my desktop, laptop, and server in my room. This made my room several degrees warmer than the hallway outside my door and almost 10 degrees warmer than the rest of the house. Unfortunately for my roommates and our electric bill, the thermostat for the house was also right outside my door. This meant while my room was a nice 75'ish, the living room was probably closer to 65. We knew something was up when we were wearing our jackets in the living room during the middle of West Texas summer.

  • Spewin Coffee (unregistered)

    Pff. 25 feet is nothing. I've had 400 feet of loose Ethernet cables running down hallways and across doorways creating all sorts of tripping hazards. Fortunately, no one died. I only tripped on one of the cables one time over a two year timeframe.

  • np (unregistered) in reply to The Del Operator

    And why not just set the temperature on the thermostat to start air-conditioning at 75 and only cool to 72 or something?

  • ¯\(°_o)/¯ I DUNNO LOL (unregistered) in reply to Archon
    Archon:
    Change the thermostat to a more complex device, molly-guard everything... you're trying to idiot-proof a server room. It's supposed to be a place idiots are barred from, which is the entire point of the story.
    Aren't we in that universe which is famous for creating bigger idiots?
  • (cs)

    Of course, there is the trick that was published eons ago before there were computer controlled thermostats. Put a small night light right below the thermostat of the house and have it on a timer. It would raise the temp the thermostat thought it was and cut off the heat (in the winter) during the time when everybody was snug in bed (and really didn't need as much heat). It was a simple solution to a couple of problems. Now we have $100 thermostats that do the same job. Win some, lose some...

  • Fred (unregistered) in reply to Mike
    Mike:
    Jack:
    Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
    I spent about six months writing a monitoring system that would probe multiple indicators regularly, graphically displaying trends in both response time and error messages, so you could readily see when things were bogging down, spot patterns, and deal with them before you had a full on outage.

    Cut to about one hour after go live, when we were training the operators who would use it. #1 requested feature: an alarm to page us when something actually does go down. They didn't care about anything prior to that, but get an outage, and they all wanted to be the heroic first responder.

    I don't really blame them. Years ago I saw an outstanding sys admin, who anticipated all problems and resolved them before they occurred, get fired because he "never did anything". I learned. The people who pay your salary need smoke now and then to remind them why you exist.

    This is a very unfortunate tradition in IT (and indeed any discipline where experts work for idiots). So many times I have seen awards given to people who swan in when something is broken and fix a probelm that they they could have avoided in the first place. Rarely (read never) do I see someone lauded for the proactive approach that saved a disaster.

    To be a little bit fair to the idiots, they only have the expert's word that a disaster would have happened had they not intervened.

    Taking it all a little further (and admittedly I work for a government department that is spectacularly good at projects that plan to do nothing and still run behind time and over budget) people seem to have worked out that successful projects are dismissed as 'trivial' because there were no issues encountered. It is important, then, to make sure a major (even minor TBH) project causes a major outage and affects everyone, so you can prove that your staff are capable of fixing major issues. Therefore, it is important not to try too hard to avoid invisible issues - aside from everything else, if a deployment runs smoothly, the client will ask why it took so long to get to that point - smooth running shows that we spent too much time in phases like planning and gave too much thought to potential issues - this time would far better be spent recovering from the issues caused by not identifying the potential problem.

  • Fred (unregistered) in reply to Secure your Job
    Secure your Job:
    I can answer that one for you. Job security. If things hum along w/o incident, upper management figures they can cut staff. If there are pages and alarms from time to time, you obviously need the staff to deal with all the issues, even if they could have been avoided.

    This is why it took 3 Unix admins to monitor 4 racks of equipment and yet only 1 Windows admin to monitor an equivalent rack of equipment. None of the unix machines had basic scripts to do things like rotate/compress log files, alert on disk space BEFORE it was an issue, etc.

    When management (for other reasons) finally got rid of the Unix folks, the Windows "team" (1 person) found themselves inundated with pointless alarms that were easily fixed with some scripting to avoid the problems in the first place. Before long, the 1 person was doing the work that had required 3 people to do. (and no, didn't get paid 3x as much)

    This is true, but perhaps the simplest fix is a GUI monitor. When you need to look like you're valuable, you change some bits to red. After a while you change them back to green. Now people will know that disaster was near, and it was averted....

    True story: After implementing some new (resource intensive) libraries to do some data matching we found that average heartbeats had gone up from around or even under 1 second to above 3 seconds. This meant that the monitoring web page the business had access to showed a lot of red (we showed blue for heartbeats below 3 seconds, and red for ones above it). Naturally, so much red means that the world is about to end, so the poor service desk were inundated with calls from the business insisting a high severity issue be raised to resolve the issue. We moved the threshold to 6 seconds and averted the disaster, because the same data now showed a lot less red.

    Corollary: Red is bad, yellow is a little bad, green is good, blue is all systems normal. It's not about the data, it's about the way you present it for the idiots to interpret.

  • Cheong (unregistered) in reply to AN AMAZING CODER
    AN AMAZING CODER:
    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF. The second one is not having a cover over the thermostat.

    I have a feeling there was a "legitamite" (as in not the blamed person's fault) reason Evi was there, but when this could-be problem was spotted Evi used it as an opportunity to rage about having his authority challeneged.

    We've all been there.

    Actually when planning server rooms, we need to plan air flow direction to make sure cold air from A/C get evenly blowed to the entire room. Also it helps to place a few thermosensors to the ends of air flow to make sure the temperature is just right.
  • Captcha:causa (unregistered) in reply to RandomPerson
    RandomPerson:
    AN AMAZING CODER:
    CodeNinja:
    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF.
    Wait... why buy a second thermostat if the first one can be monitored?
    Pardon my ignorance but... don't most if not all modern computers have temperature sensors in them already? Wouldn't it make sense to pay more attention to these?
  • Le Forgeron (unregistered) in reply to Captcha:causa
    Captcha:causa:
    Pardon my ignorance but... don't most if not all modern computers have temperature sensors in them already? Wouldn't it make sense to pay more attention to these?

    Yep, but the only small issue is that the A/C is not provided with the right interface to connect to all the computers in the room to fetch their temperature and then combine them somehow to compute the function of a single thermostat.

    What A/C need: bluetooth ! With bluetooth enabled computer to reports their various sensors.

    (beware of traveling system, or system in the next room...)

  • Mijzelf (unregistered) in reply to pauly
    pauly:
    Embedded systems don't run so hot, because sometimes there is no file system.

    Brilliant. So obviously a FILE_NOT_FOUND error cools down the system.

    Captcha: ingenium, it's an ingenious quote.

  • John (unregistered) in reply to Fred

    ...and you eloquently explain why no one is prepared to make changes now to avoid making anthropological global warning into a complete disaster.

    captcha: suscipit - the process of working out that hot drinks should be drunk carefully.

  • John (unregistered) in reply to Fred
    Fred:
    Mike:
    Jack:
    Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
    I spent about six months writing a monitoring system that would probe multiple indicators regularly, graphically displaying trends in both response time and error messages, so you could readily see when things were bogging down, spot patterns, and deal with them before you had a full on outage.

    Cut to about one hour after go live, when we were training the operators who would use it. #1 requested feature: an alarm to page us when something actually does go down. They didn't care about anything prior to that, but get an outage, and they all wanted to be the heroic first responder.

    I don't really blame them. Years ago I saw an outstanding sys admin, who anticipated all problems and resolved them before they occurred, get fired because he "never did anything". I learned. The people who pay your salary need smoke now and then to remind them why you exist.

    This is a very unfortunate tradition in IT (and indeed any discipline where experts work for idiots). So many times I have seen awards given to people who swan in when something is broken and fix a probelm that they they could have avoided in the first place. Rarely (read never) do I see someone lauded for the proactive approach that saved a disaster.

    To be a little bit fair to the idiots, they only have the expert's word that a disaster would have happened had they not intervened.

    Taking it all a little further (and admittedly I work for a government department that is spectacularly good at projects that plan to do nothing and still run behind time and over budget) people seem to have worked out that successful projects are dismissed as 'trivial' because there were no issues encountered. It is important, then, to make sure a major (even minor TBH) project causes a major outage and affects everyone, so you can prove that your staff are capable of fixing major issues. Therefore, it is important not to try too hard to avoid invisible issues - aside from everything else, if a deployment runs smoothly, the client will ask why it took so long to get to that point - smooth running shows that we spent too much time in phases like planning and gave too much thought to potential issues - this time would far better be spent recovering from the issues caused by not identifying the potential problem.

    This is the quote I shoulda had.

  • (cs) in reply to Mike
    Mike:
    Jack:
    Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
    I spent about six months writing a monitoring system that would probe multiple indicators regularly, graphically displaying trends in both response time and error messages, so you could readily see when things were bogging down, spot patterns, and deal with them before you had a full on outage.

    Cut to about one hour after go live, when we were training the operators who would use it. #1 requested feature: an alarm to page us when something actually does go down. They didn't care about anything prior to that, but get an outage, and they all wanted to be the heroic first responder.

    I don't really blame them. Years ago I saw an outstanding sys admin, who anticipated all problems and resolved them before they occurred, get fired because he "never did anything". I learned. The people who pay your salary need smoke now and then to remind them why you exist.

    You don't need smoke; just annoying emails informing them of every maintenance and upgrade task.

  • (cs) in reply to herby
    herby:
    Of course, there is the trick that was published eons ago before there were computer controlled thermostats. Put a small night light right below the thermostat of the house and have it on a timer. It would raise the temp the thermostat thought it was and cut off the heat (in the winter) during the time when everybody was snug in bed (and really didn't need as much heat). It was a simple solution to a couple of problems. Now we have $100 thermostats that do the same job. Win some, lose some...
    That sounds great, but in my house the receptacles are at the standard location about 12" from the floor while the thermostat is about 5' from the floor. You'd have to have quite a large "night light"... really, it would have to be a socket on a long cord that you could actually tack on next to the thermostat. Also, programmable thermostats are less than $30 now.
  • Ken B. (unregistered) in reply to Jack
    Jack:
    Ken B.:
    Jason:
    I'm very glad I don't have a job where I'd get called in at 3am because of a temperature issue in a small server room, when all the servers are still running fine.
    So you'd rather wait to be called until after the servers start melting because it's too hot in the room?
    Yes. It is exactly like the disk-full situation that has played out countless times over my career. Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
    Ah... And that also takes care of the "why do we even pay you to maintain the systems, when they never have problems to begin with" issue.
  • One does not simply wakka wakka wakka into mordor (unregistered)
    The Darwin Award Nominee doesn't learn anything if they're dead.
    The implication being they do if they're not? Pshaw.
  • Tonsil (unregistered)

    At an old job, we worked out of a double-wide trailer that served as a lab. It was situated in a river valley next to the water, and as such could get very cold in winter. One day, when it was particularly unbearable, we discovered that the one tech had a space heater in her office right next to the thermostat for the whole building.

  • Andrew Edelstein (unregistered) in reply to Mike

    By the same token, when taking a Netware class aeons ago, the instructor told us about a time he was in a senior staff meeting. A director had submitted a request to the president to give an employee a raise. The president responded "I don't recognize his name. Who is he?" The director explained he was the Netware admin. The president's response was "Ah, that's why I don't know his name. The Netware never breaks. Approved!"

  • the beholder (unregistered) in reply to Ken B.
    Ken B.:
    Jason:
    I'm very glad I don't have a job where I'd get called in at 3am because of a temperature issue in a small server room, when all the servers are still running fine.
    So you'd rather wait to be called until after the servers start melting because it's too hot in the room?
    Just as Jason, I'm also glad that I don't have a job where I would be called at 3AM because of whatever it is happening in the server room.

    In other words: I'm glad it is someone else's fucking job to take those calls while I sleep soundly.

  • (cs) in reply to Mike
    Mike:
    Jack:
    Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
    I spent about six months writing a monitoring system that would probe multiple indicators regularly, graphically displaying trends in both response time and error messages, so you could readily see when things were bogging down, spot patterns, and deal with them before you had a full on outage.

    Cut to about one hour after go live, when we were training the operators who would use it. #1 requested feature: an alarm to page us when something actually does go down. They didn't care about anything prior to that, but get an outage, and they all wanted to be the heroic first responder.

    I don't really blame them. Years ago I saw an outstanding sys admin, who anticipated all problems and resolved them before they occurred, get fired because he "never did anything". I learned. The people who pay your salary need smoke now and then to remind them why you exist.

    Well that depends. If the managament is worth their salt and you drop a few hints at the yearly salery negotiations that when they do not notice the maintenence work, it is because you do your job properly, then they will get it. However if they are clueless, they will not, but then you are perhaps better off finding an other place anyways...

    I guess I'm lucky, as my managers knows that when all the systems I supervise work and noone notices how many systems we actually have because they work, then I'm doing my job effectively.

    Yours Yazeran

    Plan: To go to Mars one day with a hammer.

  • (cs) in reply to operagost
    operagost:
    Mike:
    Jack:
    Why TF would the operators/admins monitor disk space and forestall a crisis when they can just wait until everything blows up?
    I spent about six months writing a monitoring system that would probe multiple indicators regularly, graphically displaying trends in both response time and error messages, so you could readily see when things were bogging down, spot patterns, and deal with them before you had a full on outage.

    Cut to about one hour after go live, when we were training the operators who would use it. #1 requested feature: an alarm to page us when something actually does go down. They didn't care about anything prior to that, but get an outage, and they all wanted to be the heroic first responder.

    I don't really blame them. Years ago I saw an outstanding sys admin, who anticipated all problems and resolved them before they occurred, get fired because he "never did anything". I learned. The people who pay your salary need smoke now and then to remind them why you exist.

    You don't need smoke; just annoying emails informing them of every maintenance and upgrade task.

    I was dropped in the deep end maintaining a project which required a lot of manual intervention and cost the company so much to run that we barely broke even. The obvious solution (automate) was, er, approved when I raised it. Unfortunately this resulted in the work of three people no longer being needed. As they could not retrain these people to do anything more complicated than the simple (now automated) tasks they had been doing, they had to be let go. As for me, I moved on to another project.

    Ten years later that project was still a major money-spinner and, despite every effort, the customer could not be persuaded to upgrade to a project that ran on non-obsolete hardware - because it worked sufficiently smoothly and trouble-free that they could see no reason so to do.

  • A Gould (unregistered) in reply to np
    np:
    And why not just set the temperature on the thermostat to start air-conditioning at 75 and only cool to 72 or something?

    Because if you're blowing an exhaust directly on (or even generally near) the thermostat, that corner of the room will never hit 72.

    Had this happen twice at my company:

    1. Bosses set up a rack of three thermal printers (the kind you use to print shipping labels) on the same wall as the thermostat. It would be freezing outside (and inside!), but the AC would still be firing at full-blast trying to bring the temperature down at that square foot where the 'stat was.

    2. New building uses zone AC. Two offices share a thermostat. Office 1 has a manager who works days. (One guy, one computer/phone/monitor.) Office 2 is shared by two guys working afternoons/nights, each has dual-monitors, plus a fax machine and thermal printer. This room kicks out a crazy amount of heat. The 'stat was in office 2, which meant at night, #2 is nice and cool (no-one there), the AC won't turn on even though the windows are starting to fog up in office 1. They move the 'stat over to office 1, and now office 2 is freezing because it's trying to keep #1 cold. (To my company's credit, they then just added another zone, and everyone is happy).

    Moral: the location of the thermostat is important.

  • Paul Neumann (unregistered) in reply to RandomPerson
    RandomPerson:
    AN AMAZING CODER:
    CodeNinja:
    Something monitoring the A/C's thermostat instead of having it's own thermostat is already the WTF.
    Wait... why buy a second thermostat if the first one can be monitored?
    CodeNinja:
    The second one is not having a cover over the thermostat.
    You can't cover it completely, there's already enough of a cover over it? You need the air flow to measure the air temperature?

    Point an exhaust at a thermostat and it will change the temperature. Fix that problem, not the working thermostat.

    An ISP I worked at had installed a new rack on the end of an aisle in the server room holding only 6 2u boxen. This caused enough of a change in the airflow, that the primary A/C system thermostat was reporting a "yellow" level temperature. The backup A/C was still reporting "green" level temperature; The environment monitor thermometer (note the difference everybody) was also reporting "green" levels. The assumption was that the new boxes were just blowing a little extra heat in the air stream directed at the therm. As such, the yellow level was bumped up to become green again and made the world well.

    Fast forward one week when the both A/C thermostats jumps to red level and the server room has become ~97 degrees in the period of 30 minutes. Further inspection showed that the primary A/C had been running non-stop the entire week causing the condenser coils on the roof to freeze over solid (in mid-summer, mind you). The secondary A/C system took over without alarm (it had been disabled because the secondary system would often kick on during warmer than usual days) and had frozen up as well. The primary condenser and a few other pieces were destroyed causing about $13,000 damage to the $75,000 system.

    The moral? Treat warnings as errors, never ignore errors simply because they are common, and (drum roll) thermostat placement is CRITICAL!

Leave a comment on “A Heated Situation”

Log In or post as a guest

Replying to comment #385713:

« Return to Article