The second life of system units
- system unit identification;
- internal inspection;
- kernel operation check;
- testing equipment.
Step 1. Identification of the system unit
This step can be carried out directly in the warehouse where the system managers are stored, if lighting allows, a table and chair will also be required to fill out the passport. May require consumables for marking, as well as wet wipes or a damp cloth for pre-cleaning external surfaces. In most enterprises, inventory numbers are applied to equipment. If so, and inventory numbers are clearly visible, then this is quite enough for this step. In all other cases, identification data (for example, the number in order) should be applied to the system unit case, so that it is easy to distinguish one system unit from another. It may be necessary to duplicate the identification data inside the system unit. We enter the identifier in the passport (inventory number or number in order), you can also give the names (brand, model, if any), but this is optional. Next, an external examination is done. In principle, the gradation can be as follows: satisfactory (the appearance allows the user to put the system) and unsatisfactory (the appearance does not allow the user to put the system). The results of the external examination are entered in the passport.
Step 2. Internal inspection
Before starting an internal inspection, you need to clean the system unit of dust. At many enterprises, sysadmins neglect the annual procedure for cleaning systemists from dust, so this step is primarily devoted to them.
This procedure will require a room free of employees (you can do the cleaning on the street, if conditions and weather permit). The room must have a table, a chair and an electrical outlet if a vacuum cleaner is to be used. For this procedure you will need: a small paint brush, a wooden stick (you can buy instant noodles that come with two whole wooden sticks), an enema like a pear (like a heating pad, it will not work for us), and naturally a Phillips screwdriver. It will be very wonderful if an old Soviet vacuum cleaner is available, which could be used to whitewash it; if it is not, then the usual one will do. You can do without a vacuum cleaner, using an enema, but with large volumes your wrists will hurt. Compressors and cans of compressed gas can be used. Overalls and protective equipment will also be required: technical gown (overalls), work gloves, respirator and safety glasses. Whether or not to use protective equipment is up to you. In addition, the following consumables will be required: thermal grease, a rag, or wet wipes.
We remove the side wall of the case, after which with a brush (cloth or rag) we clean the horizontal surfaces (not clean, but not to dust). The collected “moss” should preferably be placed in a container that can be made from a plastic bottle (greetings to Andrei Alexandrovich Bakhmetyev). Next, we clean the wires and connectors coming from the power supply, disconnect the connectors, unscrew and remove the unit. Remove the power supply cover and clean the “Augean stables”.
I draw your attention to the fact that if you are cleaning with a stream of air, it is better to fix the fan wings, for example, with the same Chinese stick. This is done in order not to damage the fan motor and moving parts, if this is not done at best, the fan will make noise, and at worst the electric motor will not work. After cleaning the power supply, you can visually inspect: the capacitors for swelling, the rest for burnout, write down the results of the inspection of the power supply in the passport. It is strongly not recommended to use the power supply for swollen capacitors or other elements with signs of burnout.
We put the power supply aside, now we will remove the memory strips (very often they are located close to the radiator, so you can inadvertently touch it). If thermal paste has not changed for more than 5 years, then it is better to remove the radiator from the motherboard, in other cases, try to remove the fan from the radiator. Haste in this case will lead to a loss of time. Therefore, if in doubt, it is better to remove the radiator. This procedure may require removing the motherboard from the chassis of the chassis, as some coolers are mounted on a lining under the motherboard. In this case, the cleaning will be done better.
After cleaning the cooler and radiator, do not rush to put the fan (radiator) in place, clean the elements, especially the condensers located next to the radiator, and inspect them. Record the results of the inspection in your passport. In particular, you should clean the RAM slots with a stream of air (then an enema can come in handy). Now you can put in place the fan (radiator).
Before cleaning the rest of the motherboard (if you did not remove the motherboard from the chassis chassis), disconnect all connectors of the drives and hard drives, remove all non-integrated cards and strips (video card, network card, USB stick, etc.), so as in the next step, we still have to shoot them. All cards that have been removed should be cleaned, inspected, and the results of the inspection, as well as the model and brand, noted in the passport. After cleaning, we examine the remaining elements of the motherboard, including the chipset (it often has microexplosions on it, for example, as a result of a short circuit on a USB device).
If drives and hard drives interfere with the cleaning of housing niches, they should be removed. After removing the drive (hard drive) with a brush, we clean the surface of the device, make an inspection and enter the make and model in the passport. We finish cleaning by wiping with a damp cloth or a cloth of large dusty surfaces. Before installing RAM, clean the contacts with an elastic band or a dry cloth. Before installing the power supply in place, check it. There are several verification options: the first option is a Power Supply Tester check, the second option is a multimeter test with the green and black contacts closed under load (for example, a fan) and the third option is connected to test (known in advance that it is working) equipment (in In this case, a motherboard with a processor and RAM with the same power connector is suitable, as on the power supply). We enter the results in the passport. Please note that a motherboard with swollen capacitors is best avoided.
Step 3. Checking the kernel
Under the core is understood: power supply, motherboard, processor, RAM. For a motherboard without an integrated video adapter, the core includes a video card. To complete this step, you will need a table, chair, uninterruptible power supply, monitor and keyboard. Consumables may require a CMOS battery (CR2032), and a cabinet speaker.
At the start, we listen to the sound signals from the connected speaker. If the screen still doesn’t light up, we try to connect a test (we know in advance that it is working) power supply, if the result is negative, reset the BIOS with the Clear CMOS jumper (Clear BIOS, Clear RTC, etc.).
If the kernel "started with one peak", then in the BIOS we look at the date for replacing the CR2032 battery (if you did not reset the Clear CMOS settings), kernel characteristics: model and processor speed, number of cores, type and amount of RAM, motherboard model . You should also look at the temperature of the processor, if the processor is very hot, then you may have installed the cooler incorrectly.
At the same step we listen to the “noise” of the fans. The results are written in the passport: whether the kernel is working properly, the characteristics of the core, the need to replace the CMOS battery and cooler fans, the case and the power supply.
After that, turn off the system and install all the removed cards, connect the cables and loops, in general, as before cleaning. We try to turn it on in full running condition. The purpose of this measure is to prepare a systematic for testing. A situation is possible when some equipment begins to "hang up" the system. Two options are possible here, either the power supply does not “pivot” this equipment, or this equipment is faulty. In both cases, we must turn off this equipment, and make a mark in the passport. If possible, connect this equipment to a working system in order to verify its functionality.
Step 4. Testing the equipment
To complete this step, you need the same as in the previous step plus: a mouse, a boot disk with a set of test programs, a network switch or router for checking the operation of a network card, speakers for checking a sound card, a USB device, for example, a USB flash drive.
We need to test the central processor, and it is advisable to see a graph of the temperature of both the CPU itself and the motherboard, as well as power fluctuations. We need to test the RAM for the presence of broken cells and the hard drive for bad sectors. It is also advisable to test the chipset, video adapter and other devices for performance. This is best done from a boot disk, because in this case we will not interfere with services, startup programs and antivirus programs. There is a drawback of this approach - the lack of the necessary device driver on the boot disk will not allow you to check fully.
There are a lot of programs for stress testing, I will list only the most famous: AIDA64 (formerly Everest), SiSoft Sandra, S&M, OCCT. In addition to stress testing, you can use the good old Memtest86 to test RAM. But you can get confused with hard drive verification programs. Firstly, it is better to use official diagnostic programs from manufacturers of hard drives. Secondly, the Victoria disk verification program is still popular, and there is another popular program - mhdd. Thirdly, there are other programs that check for bad blocks, for example, DiskGenius, a universal utility for partitioning and restoring data.
Keep in mind that testing can be very time consuming. It should also be clearly understood that if all tests are successfully completed, then with a certain degree of probability it can be considered equipment in good condition. Therefore, it is necessary to test, but without fanaticism. You can approve the following policy: if the stress test is successful, then you can consider the equipment to be serviceable, and use specialized tests only if you suspect it. In any case, one must be prepared that it will fail at any moment. Although practice shows that old computers are more stable than new ones.
Summary
And since the passport contains comprehensive information about the current state of the system unit, it is time to make a decision. Of course, you will make the decision, but here are only recommendations. Let's divide the systemists into groups: top grade, first grade, second grade, third grade and substandard. The highest grade will include systemics having:
- satisfactory appearance;
- inspection of equipment - without comment;
- low-noise fans and coolers (or replaced);
- the characteristics of the processor, RAM and hard drive correspond to the optimal requirements for the software used in your enterprise;
- stress test passed without comment.
The first grade differs from the highest one only in the point: the characteristics of the processor, RAM and hard disk correspond to the minimum requirements for the software used in your enterprise. The second grade differs from the first one only in the point: the characteristics of the processor, RAM and hard drive do not meet the minimum requirements for the software used in your enterprise. The third grade includes system drivers with unsatisfactory appearance, with noisy fans, for example, which cannot be changed due to their specificity, but with a working core that has successfully passed testing (drives and hard drives may be malfunctioning). Everything that was not included in the previous groups will relate to non-condition, which can be analyzed, that is, remove working equipment.
The third grade can be used as miniservers: firewalls, routers, VPN gateways, file cleaners, alarms, etc. Or use as test equipment. The second grade can be used instead of terminal clients, or for undemanding jobs. From systemists of the highest and first grades, you can create a replacement fund.
The replacement fund can be used as a temporary replacement for the working system unit in order to reduce or eliminate employee downtime in the following cases:
- failure;
- suspected viral infection;
- long-term software updates, for example, switching to another operating system;
- scheduled maintenance, e.g. dusting.
Some enterprises allow themselves the creation of training (test) laboratories, which are equipped with system engineers from the first and highest grades. But the option of non-immediate transfer of systematic units is rarely used now, probably, because nobody needs such systematic systems. Perhaps readers will prompt some other use cases.