My cool stories

    Representatives of one western corporation came here again. We drank it. They asked about our barbaric features of IT. Well, I told a couple of stories. Chukhna did not believe it, she says, I was lucky, and this is all a statistical outburst. For some reason I don’t think so - everyone has such stories with a suitcase.



    Once a colleague from L2 support says: take your purse to the object, since you are going to the same city. Ok, I say, no question, of course I will. I arrive at the office with a briefcase, and there are three security guards and this one from L2. They have a black sports bag, hefty. It opens with difficulty, but I still open it, look in - and there’s one-to-one RAM dice for threshing servers of one company, conditionally “three letters”.

    I approximately estimate the value of the “handbag”: I say that, friend, there are about three hundred thousand dollars, right? Not, he says, there’s half a million. Come on, come on. Here are the guys leading you to the airport entrance, and then you yourself. I did not get out of the arrival airport until the person meeting took my purse.

    One university in the region had an expensive pulsed laser, and everyone knew that it was the most valuable thing there. One fine day, some comrades unclenched the grate on the window (made a loop between the rods, inserted the armature and twisted like a lever until a sufficient hole was formed). And they stole the programmer instead of the laser, because it looked clearly much more impressive, was connected to a computer.

    The cost of the programmer is cheap, but this is no more. He’s been there almost since the days of the Soviet CNC machines, which are on punch cards. And the laser is backward compatible with it and with new expensive systems. Then the scheme was like this: they modeled a part on a computer, gave it to the programmer, it gave it to the laser. We came to set up a direct connection. Draw circles, squares and more complex shapes, reverse-engineering format. In the end, it turned out, but still not completely good. As far as I know, the stolen programmer never surfaced, and in the end they got somewhere the same a couple of years after the modernization of one of the industries in the region.

    There was such a case: on a large storage system, two disks from RAID flew out at once. The manufacturer suffered for 4 days, and then found data loss. Having received freedom, the administrator of the injured insurance took these discs to the recovery laboratory and got one corrected back, and then, using the chisel and some mother, ignoring the warnings, nevertheless made a rebuild and collected consistent data. It was precisely for this case that our chukhn was offended, by the way. They say that we are all professionals there, this could not be.

    Another friend of mine had a similar case - an Indian from global tech support broke the array, and when the local team started asking what idiot he did, they were presented with a fact. I must say that they admitted their mistake, but they could no longer help and, of course, there was no compensation. They believed in the Indian; all know the Indians well.

    Or here: one of my friends admin had such a story. He came to a large state-owned company (in the region), began with an inventory. Found flash storage, which is not listed anywhere. The array is fresh, 3 years from production, 2 years from purchase. It costs like a pretty big piece from an airplane. Somehow they raised the story, found out that the previous specialists either did not buy support, or simply forgot something, and as a result of the transfer from balance to balance to another city, the piece of iron was left unaccompanied. They couldn’t raise a backup for it, they just didn’t even announce LUNs. As a result, the backup went to Google services (this is with a Russian state-owned company), and a lot of taxpayer money quietly stood idle in the server room.

    In another state-owned company in the hot southern region, they really did not want to pay for backup services. Simply physically, we could not explain to their supervisor why it is necessary to pay for work when there is no breakdown. Well, like, there will be a problem - he will pay these unfortunate 3,750 rubles a month, but for now he does not believe that this is necessary. Naturally, according to the law of the genre, they catch a virus through accounting, put their RDP farm with 1C, lose their base on a winlocker with a key of 1024. That same day, the manager calls and says that everything has fallen, and he is ready to pay these 3,750 rubles. How to fix it.

    Once again I saw how in a fairly large company they mixed up the IP of the test server and the production server. And for several months they drove tests directly on real customer data. They realized the problem at the moment when one of the clients saw a strange movement in the account (in minus). The kickbacks were epochal - it was necessary to raise the entire history of tests and return, in fact, with your hands, as it was.

    Or there was a story. The client (government agency) has dropped one of the loops of the array. We came from EMC for support, diagnosed a problem with the LCC card. As a result, after dancing with a tambourine, it turned out that replacing the LCC card did not help. The base had to be raised very quickly - there were serious financial losses for the client.

    The Russian IT guy who was at the site did not agree with the diagnosis of a potential adversary and proposed a different solution. There was an IO-module from another machine room, put it during the development of this version. Everything rose immediately. But since they didn’t do according to plan, EMC sent everyone to ... Well, in general, they refused to change the faulty IO module to a new one. They say that the problem is not in it, and dancing with a tambourine was useless, they just spent the spare part. And he costs a lot of money.

    Fortunately, the manager wrote a direct letter to their (client) principal, that if the problem is not in the module, then he will change it back to the one that was. Only the base will fall again, most likely. They’ll do it right tomorrow. And he sent the insurance team, attaching an EMC letter. The next morning, EMC agreed on a replacement immediately and promised to bring a new IO in 24 hours.

    Well, of course, the best thing that happens in life is when development rolls out a release somewhere on December 30, and at that time someone updates either the router firmware or the Oracle revision to the heap. Holidays are much more joyful. Moreover, the SLA 24 hours from the manufacturers of this damn firmware means "an Indian will call you and ask you 100 stupid questions, he won’t understand anything, but he will say that the problem is on your side."
    It was the same.

    Also popular now: