Geekly Articles each Day
📜 ⬆️ ⬇️

The history of one infrastructure. MS solutions. Part 1

Lyrical introduction ...


After reading a few topics on this topic (" Why I love Microsoft. Zombie Notes " and " We build an infrastructure based on MS products "), I decided to share my stories, accumulated experience and knowledge ... For many years I have been working at an industrial enterprise, first as an engineer, and then and the head of the IT department. Having experience of successful implementations in my own factory, experience in designing and implementing large infrastructure projects in other companies, I initially wanted to write another “success story” about how we effectively use infrastructure solutions from the well-known company MS, what great they are for IT professionals , for users and for business in general. But in the end I decided to try to convey the atmosphere of building the IT infrastructure of the enterprise, to show the factors that led us to the current state, and also to tell about key errors. And in focus, of course, I will try to keep decisions from MS.

Start-up capital…


When I came to work at the plant in 2003 for the vacant position of an engineer for automated systems (AS) in the department of the chief designer (OGK), I got to manage, it’s terrible to remember, about 60 personal computers (PCs), 3 “servers” and several SOHO switches class, scattered in the offices of the plant.
Briefly described what I got:
One of the servers worked since 1999 under the control of Windows NT 4.0 and was the primary domain controller (PDC) and file server, it had two processors and even a SCSI disk. The second worked since 2002 under control of Windows Server 2000 and was a terminal server for 1C applications, also had two processors, a SCSI disk and 256MB of RAM, and somehow managed to cope with the load. For these two “important” servers, there was an old UPS that could last ten minutes or so. The third was a regular PC and was running Windows 2000 Professional, was my work computer, a gateway to the Internet (organized by WinRoute) and a mail server (organized by MDaemon), often hanging, restarting and losing an ADSL modem.
Personal computers of users, for the most part, have been working since the introduction of the domain controller and were of the Pentium I class with the Windows 98 operating system (OS) (those that were higher, worked on the basis of Windows 2000 Professional).
Switches are different, 10 / 100Mbps, the largest accumulation of which was in the accounting department, sometimes hung up and required a reboot.
All software (software), needless to say, was unlicensed. The technical documentation of this economy was absent as a class, however, and there were almost no requirements for IT management. It should be clarified that there was no IT department at the enterprise (there was also a software engineer in the general department who was engaged in a software product solving a specific production task), there was no IT budget, and I had almost no knowledge and experience ...

Something like this ... Typical, for many domestic enterprises, the situation with IT.

Adaptation…


But I was lucky - everything worked tolerably until you touch it, i.e. There were no serious accidents, therefore, I had a lot of free time, which I spent on studying the theoretical foundations and searching for effective solutions, because understood that it could be better. I was also lucky with the direct manager - an extremely technically literate specialist, tried to understand the essence and supported my initiatives to improve the IT situation (and I must say, I am a very proactive perfectionist - I work only to be proud of the results). And, therefore, I took up the development of the direction entrusted to me.
One of the strategic objectives of IT development, or rather, the desire of management, was to see the network switching between the plant management and the workshops of the plant complex. What was a difficult task due to the considerable territorial extent and the lack of adequate allocated resources. Another wish of the IT management was to get access to their e-mail from anywhere via the global network. From the wishes of users of 1C applications (and these are mainly numerous accounting staff), it was to see acceptable performance. The rest of the users of the information system (IS) wanted only uninterrupted access to Internet resources.
I, working through the questions, came to the conclusion that we needed:

All my thoughts on the development of IT, I had a rule to issue (in expanded form and with an indication of the estimated cost) in the form of memos and put on the table to the CEO.
')

"Large" implementation ...


And now, after a year and a half of this sluggish inactivity, like snow on my head, comes the news that we, they say, are introducing one well-known ERP-system! The issue is resolved at the highest level, and is not discussed. I need technical preparation of the infrastructure for implementation, i.e. FOCL organization at the plant, preparation of switching and server equipment (all this is in the strictest saving mode, of course).
The project has been agreed, launched, work has begun to boil. A room was allocated for the organization of the switching and computing center (server). Equipment was selected and ordered - servers, rack, switches, fiber-optic cable (FOC) segments (required length and with soldered connectors), converters, UPS.
The rack with the equipment is installed, the communication lines are transferred to the server. Deployed a domain based on Windows Server 2000 (2 CDs), SQL Server 2000 and an application server for an ERP system are deployed on two other servers. The communication engineers of the enterprise installed the FOC segments. Everything is tested and put into operation. It is strange that the first pancake is not lumpy and half a year of work was not in vain ...

I'm resting, waiting for the deployment team to deploy its ERP system. At the same time I am working on issues of further development of the infrastructure, I come to the conclusion that we need:

As usual, the project has been agreed, launched, work has begun to boil. Purchased two servers, UPS. Everything is deployed based on Windows Server 2003, tested and run. One more strategic task of the management is realized - ubiquitous access to the mailbox resources. Everyone is happy with the new “buns”, especially the current list of contacts. The second pancake is also far from being lumpy and another six months of work ... In parallel, these projects have replaced a lot of “technically not relevant” PCs. As a platform for personal computers selected Windows XP Professional.

I rest again, no longer waiting for the deployment team to deploy its ERP system, since management reported that the implementation project is being postponed indefinitely, funding is suspended, it is allowed to use idle servers.
Work has begun to boil. Database "1C" transferred to SQL Server 2000. Users finally transferred to the use of native versions of "1C". From the application server, after some modifications of the disk subsystem, it turned out the file storage server. The previous file dumper has become a structured repository with access control. In the end, everyone is happy, productivity has increased.

I am working on issues of further development of infrastructure, I come to the conclusion that we need:

As usual, the project has been agreed, launched, work has begun to boil. Purchased two servers, new switching equipment (because the existing one did not support VLANs), UPS. Everything is deployed based on Windows Server 2003, tested and run. Users have been trained to work with Office Communicator 2005 and are very satisfied. Much of the unwanted correspondence filtering load has been removed from the Exchange Back-End server.
All old equipment is decommissioned and placed in the far corner of a special room.

Evaluation of the first results ...


We work with what we have. One fine winter day, the main FOC main line fails - an investigation revealed that the cause was unqualified cable installation, the cable was rubbed against the corner of the building. An urgent need to restore the connection. Installation of optics is difficult due to severe frosts. Need to do something. The supplier offered to quickly organize a temporary (backup) channel using Wifi 802.11g equipment, for a while, until it becomes possible to mount the FOC. Estimated, counted and agreed. We ordered two expensive “professional” access points and directional street-mounted antennas (all equipment from one well-known manufacturer). Quickly arrived, quickly mounted and launched in bridge mode. Earned, however, not as fast as we would like. And, as it turned out during the operation, not so reliably - every day and not once the connection was lost and, after a while, was restored. Three months worked with such a backup channel and subsequently replaced with an FOC. Radio points have been dismantled and installed as client - work without complaints, to this day.
Looking ahead, I’ll say that we didn’t implement that ERP system, but perhaps we are going to introduce its new version when the economic situation in the enterprise improves significantly. And now, summing up under these milestones, we can say what mistakes were made when designing:

Inefficient ... These are the results of the lack of IT financing practices in the enterprise. Allocating money to my direction for management was something new, unusual, in wonder. It is now, I will not even begin a project, if the allocated resources do not allow "to do everything correctly, as it should be."
The second factor was, of course, the lack of necessary competencies among IT representatives (that is, me) and, as a result, there was no clear vision of the final result and there was no acceptable level of project planning (“let's start and we'll see”). The embedded infrastructure services were the result of my experiences. But I want to note that there were no serious deviations in the context of the effectiveness of using the MS software for organizing the infrastructure. It happened so well that MS products got accustomed, neither IT specialists nor IP users had any misunderstandings and disagreements with them. The management also had enough services and was aware of the main deviations.
I will say that for some time, the infrastructure was in this state, but further work on the bugs fixed everything.

Continued ...

Source: https://habr.com/ru/post/91450/

More articles:


All Articles