Server installation for 1s.

The topic of installing MS SQL Server is usually ignored. Indeed, it is difficult not to install this DBMS, even doing it for the first time, and it is equally difficult not to launch 1C:Enterprise Server in conjunction with it. However, there are a number of non-obvious subtleties that can significantly poison the life of an administrator, which we will talk about today.

MS SQL Server ranks first in the number of implementations in conjunction with 1C:Enterprise, this is largely due to the low barrier to entry; a person without experience is quite capable of mastering this combination, purely using the Next - Next - Finish method. And, what’s most interesting, all this will work. Let's say more, in the overwhelming majority of cases, the default SQL server settings are more than enough to ensure productive operation of the 1C: Enterprise server and touching them is not only unnecessary, but even harmful.

First of all, you should remember about the system base tempdb, which is actively used by 1C for storing temporary tables and intermediate results. Moreover, it is used immediately by all 1C databases running on the server. And since by default it is located in the SQL server installation folder, i.e. on the system disk, then when the load increases, it is tempdb becomes a bottleneck for the entire server. Very often this leads to situations: you bought a fast HDD / SSD, there are enough disk resources, but 1C is slow, which can cause serious difficulties for novice administrators.

Second point. Comparison encoding tempdb must match the encoding of the comparison of information bases, otherwise this may in some cases lead to unexpected results, even serious errors in calculations.

At the same time, these difficulties are not at all difficult to avoid; you just need to spend a couple of extra minutes during installation or carefully review the settings of an already installed server.

Installing MS SQL Server to work with 1C:Enterprise

As we have already said, installing a SQL server is extremely simple, and we will not describe this process in detail, focusing only on the necessary settings. Let's start with the selection of components, since 1C does not use most of the SQL server mechanisms and if you are not going to use them for other purposes, then we leave only Database Engine,Client Facilities Communications And Controls(optional).

Management tools can not be installed on the server, but installed separately on the administrator’s workstation and managed from there all available MS SQL servers.

You should also check the sorting parameters; if you have your regional settings configured correctly, then most likely you won’t have to change anything there, but it is advisable to check this parameter, it should be there Cyrillic_General_CI_AS.

IN Server configurations please indicate Mixed mode authentication and set the password to the SQL superuser - sa. Also indicate below the administrators of this instance of the SQL server; by default, there is already an account under which the installation was performed, but if other users must also administer this instance, then it makes sense to indicate them right away.

Next bookmark - Data directories- requires the closest attention. Be sure to specify the storage location for user databases and database tempdb space on a performance array or a separate disk. Despite the fact that the location of the database can be specified when creating it, setting the correct default settings saves you from unnecessary work, as well as from the situation when the database is created using 1C tools and ends up in the default directory, i.e. on the system disk. You can also immediately specify the directory for storing backups.

You can leave the remaining settings as default and complete the installation.

Setting up MS SQL Server to work with 1C:Enterprise

If you are dealing with an already installed SQL Server instance, make sure that the comparison encoding Cyrillic_General_CI_AS, otherwise the data should be downloaded using 1C tools, and the server should be reinstalled (or another instance should be installed if this one is used by other services).

To do this, open Management Studio, select the required SQL server instance and right-click on it and go to Properties.

Then go to the bookmark Memory and indicate the amount of RAM available to the SQL server, otherwise the SQL server will try to recycle all available memory. In a situation where the roles of the SQL server are combined with other roles, and in small and medium-sized implementations it is usually located on the same machine with the 1C server, you should subtract from the total amount of memory needed by the system and the 1C server, giving SQL what remains.

It is difficult to give unambiguous recommendations here; it all depends on the volume of data being processed; in practice, it makes sense to allocate half of the free memory to the SQL server, subsequently adjusting this value based on its actual load.

The next setting will be related to security. To connect 1C to the server, the account is most often used sa, which, to put it mildly, is unsafe, as it gives the person who enters under it full access to the SQL server. Considering that the administration of 1C databases is often carried out by third-party specialists, it makes sense to create a separate account for them.

To do this, open Security - Logins and create a new name (account), specify authentication SQL server and set a password.

Then go to the bookmark Server roles and allow dbcreator, processadmin And public.

Then use this account to connect to the SQL server from 1C.

Another setting applies to already created databases, open the properties of the desired database and go to the tab Files. Find the option Auto-grow/maximum size for the data file. By default, it is 1 MB, which is very non-optimal; when actively working with the database, the DBMS will only be concerned with increasing the file size; in addition, when several databases are actively working, this will lead to significant fragmentation of the data file. Therefore, based on the size of the database and work activity, set a higher value that will not lead to a constant increase in the database file.

Transferring the tempdb database

To conclude our article, let us again turn to the database tempdb, there are often situations when the file of this database needs to be moved to another location. For example, the server was installed with default settings and tempdb located on the system partition, or you purchased an SSD and want to transfer not only the databases there, but also tempdb(which is the correct solution). Also under heavy load tempdb It is recommended to put it on a separate disk.

To change the file location tempdb open Management Studio, select Create a request and in the window that opens, enter the following text, where E:\NEW_FOLDER- new location for the base:

Use master
alter database tempdb
modify file(
name = tempdev,
filename = N"E:\NEW_FOLDER\tempdb.mdf")
go

alter database tempdb
modify file(
name = templog,
filename = N"E:\NEW_FOLDER\templog.ldf")
go

Then click Execute, after executing the request, restart the SQL server, database and log files tempdb will be created in a new location, files in the old location must be deleted manually.

We will end here today, finally reminding you not to forget about database maintenance and backups.

In most cases, to install 1C:Enterprise 8.x in the client-server version, it is enough to run the 1C:Enterprise 8.x installation program. In this case, the 1C:Enterprise server receives standard parameter values ​​necessary for its normal functioning.

Let's look at installing the 1C:Enterprise server in more detail. During the installation of the 1C:Enterprise 8.x server, the 1C:Enterprise 8.x installation program performs the following actions:

* Copies the 1C:Enterprise server boot modules to the directory specified by the 1C:Enterprise installation program as the final folder.
* If "Create user USR1CV81" is selected during installation, creates user USR1CV81. The 1C:Enterprise 8.1 server runs on behalf of this user if it is launched as a service. It has access only to those resources that the 1C:Enterprise server needs. It is important that the 1C:Enterprise server requires two directories to operate: a general directory with server data (usually "C:\Program Files\1cv81\server") and a directory of temporary files (usually "C:\Documents and Settings\usr1cv81\Local Settings \Temp" or "C:\WINNT\Temp"). User USR1CV81 receives rights to the shared directory with server data. The temporary files directory is usually accessible to all users.
* If during the installation process "Install 1C:Enterprise 8.1 server as a Windows service" is enabled, then it registers the 1C:Enterprise server agent service in Windows and starts it. At the first launch, a cluster of 1C:Enterprise servers is created with default settings. It has one worker server and one worker process. The working server address matches the name of the computer on which the installation was performed.

USR1CV81 or USR1CV82 user and his rights

1C:Enterprise Server is a server application whose operation should not depend on which user is logged into the server computer in interactive mode, if anyone is logged in at all. Therefore, when installing a 1C:Enterprise server, it is advisable to create a special user USR1CV81, endowed with the minimum rights required for the 1C:Enterprise server, and not intended for interactive login. The 1C:Enterprise server is presented to the Windows system by user USR1CV81.

Let's take a closer look at the rights set for the user USR1CV81. 1C:Enterprise server uses the following directories:

* The directory of loading modules is located in the directory specified by the 1C:Enterprise installation program as the final folder. It contains the loading modules of the 1C:Enterprise server. User USR1CV81 requires rights to read data and run programs from this directory and its subdirectories. It receives these rights implicitly by being included in the Users group.
* The server data directory is usually named "C:\Program Files\1cv81\server". User USR1CV81 requires full rights to this directory. The 1C:Enterprise installation program, when creating the user USR1CV81, gives him rights to this directory.
* The temporary files directory is usually named "C:\Documents and Settings\usr1cv81\Local Settings\Temp" or "C:\WINNT\Temp", which is determined by the value of the user's environment TEMP variable or the system environment TEMP variable. You can view the value of this variable in the System Properties dialog (Start -> Settings -> Control Panel -> System -> Advanced -> Environment Variables). The 1C:Enterprise installation program gives user USR1CV81 full rights to this directory. Typically, when installing Windows, the temporary files directory is accessible to all users by including the CREATOR OWNER group in its access list. However, this access is not full. In particular, searching for files in this directory is not available to all users. Setting user USR1CV81 full rights to the temporary files directory allows the 1C:Enterprise server to perform all the operations it needs. You can view the access list in the directory properties dialog on the Security tab. The presence of the CREATOR OWNER group allows access to the directory by any user who creates any files in this directory or owns any files in this directory. In this case, in the access list of the created file, instead of the CREATOR OWNER group, the user who created the file will be written. Among the users who are allowed access to this directory, there must be user USR1CV81, who has full rights to this directory.
It is important to keep in mind that the temporary files directory for a given user (including user USR1CV81) is determined by a combination of that user's environment variables and system environment variables. To find out this directory, the 1C:Enterprise installation program requests the user context USR1CV81. To do this in Windows 2000, the user on whose behalf the 1C:Enterprise installation program is launched may need the following privileges: Act as part of the operating system and Bypass traverse checking. You can check user privileges using the Local Sequrity Settings utility in the Local Policies -> User Rights Assignment branch. When installing new software, the installer usually gains these privileges automatically.

Registering a 1C:Enterprise server as a Windows service

1C:Enterprise Server is a simple Windows console application and can be launched interactively. However, for constant use this is inconvenient, since it requires the launch of the 1C:Enterprise server from the login of an inactive user to the server computer. To eliminate this dependency, the 1C:Enterprise server can be launched as a Windows service. To do this, it must be registered in the Windows service manager.

To view a list of Windows services and their parameters, use the Component Services utility (Start -> Settings -> Control Panel -> Administrative Tools -> Services). The 1C:Enterprise server is represented in the list of services by the service "1C:Enterprise Server Agent 8.1". The service parameters determine the launch of the 1C:Enterprise Server Agent process (ragent), the user under whose name it is launched, and the method of restarting in emergency situations.

In the properties dialog of the "1C:Enterprise 8.1 Server Agent" service, on the General tab, the line for launching the ragent process, which is the 1C:Enterprise Server Agent, is shown. Typically this line looks like:


It states that:

* the Server Agent process is the boot module "C:\Program Files\1cv81\bin\ragent.exe";
* the ragent process runs as a Windows service and must be managed by a service manager (-srvc);
* used as 1C:Enterprise Server Agent (-agent);
* when starting the service for the first time, a cluster must be created with default parameters and main IP port number 1541 (-regport 1541). Using this port, client applications must connect to infobases registered in the cluster;
* The server agent IP port must be numbered 1540 (-port 1540). Using this port, the Cluster Console must connect to the central server to perform administrative functions;
* when cluster processes are launched on this server, they will be dynamically assigned IP ports from the range 1560-1591 (-range 1560:1591).
* general cluster data will be located in the "C:\Program Files\1cv81\server" directory (-d "C:\Program Files\1cv81\server").

The "1C:Enterprise 8.1 Server Agent" service can be added or removed not only when installing or uninstalling 1C:Enterprise using the 1C:Enterprise 8.1 installation program, but also manually. To do this, you can run the ragent utility from the command line, specifying the appropriate parameters.

To create a service, you need to specify the -instsrvc parameter and the following parameters: -usr - the name of the user under whose name the service should be launched, -pwd - the password of this user. In this case, the remaining parameters will become the launch line parameters of the 1C:Enterprise Server Agent as a service. For example, for standard registration of the 1C:Enterprise Server Agent service in debug mode, the set of parameters should be as follows:

"C:\Program Files\1cv81\bin\ragent.exe" -instsrvc -usr .\USR1CV81 -pwd Password -regport 1541 -port 1540 -range 1560:1591 -d "C:\Program Files\1cv81\server" - debug

To remove a service, you need to specify the -rmsrvc parameter. For example:
"C:\Program Files\1cv81\bin\ragent.exe" -rmsrvc

Sometimes it is useful to change the Server Agent startup line or other parameters of the Agent service, for example, enable debugging mode, or create several services of different versions. The service properties dialog does not allow you to edit the service application launch line and some other parameters, for example, the service identifier. To edit, you will need the regedit utility, designed to view and edit the Windows system registry.

Attention!
Editing the Windows system registry requires extreme caution, since erroneous changes to it can render the operating system inoperable.

Run the regedit utility (open Start -> Run and type regedit) and select the branch:


Among its parameters is the ImagePath parameter, the value of which is the startup string of the 1C:Enterprise Server Agent. Here you can add new launch string parameters or change the values ​​of existing ones. A complete list of possible parameters is given in the book "1C:Enterprise 8.1 Client-Server" documentation.

If you need to register several independent 1C:Enterprise Server Agent services, you need to specify them with different boot modules, different ports and different cluster data directories. You also need to register them with different service identifiers. This can be done like this:

* Create the first service:
"C:\Program Files\1cv81\bin\ragent.exe" -srvc -agent -regport 1541 -port 1540 -range 1560:1591 -d "C:\Program Files\1cv81\server"

* Using the regedit utility, change the identifier of the registered service. To do this: select a branch
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\1C:Enterprise 8.1 Server Agent

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\1C:Enterprise 8.1 Server Agent First
* Create a second service:
"C:\Program Files\1cv81_10\bin\ragent.exe" -srvc -agent -regport 1641 -port 1640 -range 1660:1691 -d "C:\Program Files\1cv81_10\server"

* Perhaps his ID should also be changed. To do this: select a branch
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\1C:Enterprise 8.1 Server Agent
and change its name, for example to:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\1C:Enterprise 8.1 Server Agent Second

What can't the 1C:Enterprise installation program do?

As already mentioned, the 1C:Enterprise installation program copies the 1C:Enterprise boot modules and performs the necessary registration in COM and in the Windows service manager. The above is information necessary to understand the internal mechanisms of this registration. If not only the server is installed on the server computer, but also the client part of 1C:Enterprise, then it is ready to work immediately after installation (and connecting the security keys).

In order for the 1C:Enterprise server to be accessible from other computers on the local network, you need to check the network settings on the server and client computers, as well as for the network as a whole. TCP/IP is used to transfer data between client applications and the 1C:Enterprise server, as well as between server cluster processes. The operation of 1C:Enterprise in the client-server version depends on its correct configuration.

Processes of a 1C:Enterprise server cluster connect to each other at the addresses defined as the values ​​of the “Computer” property in the properties dialog of working servers. The cluster requires that the value of the Computer property be either an IP address in dot notation or a symbolic address from which the IP address can be determined using the gethostbyname function defined in the TCP API. The IP address is determined either based on the local table of symbolic addresses (C:\WINNT\system32\drivers\etc\hosts) or on the address tables in available DNS servers. If the symbolic address of the working server does not determine its IP address or determines it incorrectly (for example, the IP address does not match the actual IP address of this computer), then the cluster will not work. It is important that the names of computers and their addresses defined in Windows on each of the working servers in the cluster do not contradict their names in DNS.

On each working server, cluster processes use the following ports: IP port of the working server (usually 1540); IP ports from the workflow IP port ranges (usually 1560-1591). Additionally, the cluster's central server uses the cluster port (usually 1541). If the system uses firewalls, then data transmission on these ports must be allowed. Instead of allowing ports from the list above, you can allow data transfer to cluster processes (ragent, rmngr, rphost).

The connection between the 1C:Enterprise client application and the server is performed in 2 stages. It first establishes a connection to the cluster manager. This uses the central server address (symbolic or numeric) and the cluster port (usually 1541). Next, the client application establishes a connection with one of the worker processes. The value of the "Computer" property of the corresponding working server and the working process port, which is selected from the range of IP ports of the working server, are used as its address. Data transmission to these ports must be allowed in all firewalls along the route from the client application computer to the computers of the 1C:Enterprise server cluster. The IP address of server processes is determined using the gethostbyname function on the client computer. It is important that the names of the central and working servers and their addresses defined in Windows on each of the servers in the cluster do not contradict their names in the DNS accessible to the client computer.

And one last thing. Obviously, in order to successfully access the 1C:Enterprise server from other computers, it must be on the network and the necessary settings must be made for this. Connecting to a network and setting up methods relate to the administration of networks based on Microsoft Windows and are described in the corresponding instructions.

Features of setting up a SQL server

1C:Enterprise in the “client-server” version uses a SQL server to store data. In this case, only 1C:Enterprise Server accesses the SQL server. 1C:Enterprise clients do not have direct access to the SQL server. Installing and configuring a SQL server is described in detail in the Microsoft SQL Server documentation. For the successful operation of 1C:Enterprise Server with a SQL server, you need to pay special attention to the following settings.

* Required SQL server components. To access the SQL server from the 1C:Enterprise Server side, Microsoft Data Access 2.6 or later components must be installed on the 1C:Enterprise Server computer.
* User authentication by SQL server. Access rights to SQL server databases are determined by the user on whose behalf the databases are accessed. From the computer on which the SQL server is installed, launch the SQL Server Enterprise Manager utility, find the Local node (Console Root -> Microsoft SQL Servers -> SQL Server Group -> (Local)) and open its properties. On the Sequrity tab you can see that the SQL server supports two methods of user authentication: SQL Server and Windows and Windows only. Windows authentication will allow the 1C:Enterprise Server to access the SQL server only on behalf of the user USR1CV81, which does not allow differentiating access rights to different infobases served by one 1C:Enterprise server. It is recommended to select the SQL Server and Windows mode. In this case, access to a specific infobase will be performed on behalf of the user who was specified as the SQL server user when creating this infobase. It is important that this user must have not only full rights to the infobase database, but also rights to create databases in the SQL server and to read Master database tables.
* Network protocols for accessing the SQL server. If 1C:Enterprise Server and the SQL server are located on different computers, then you need to configure the network access protocols to the SQL server. This can be done using the SQL Server Client Network Utility. On the General tab, you can select a list of network protocols used to access the SQL server. The fastest and most versatile is to use the TCP/IP protocol. When using other protocols, you must keep in mind that some of them, for example Named Pipes, perform additional authentication using Windows tools when exchanging data with the SQL server. In this case, to successfully work with the SQL server, the user USR1CV81 must be registered on the computer with the SQL server, with the appropriate rights. The access protocol for this SQL server can be changed on the Alias ​​tab.

In addition to the article

There is no doubt that the combination of MS SQL Server + 1C: Enterprise 8 server is the most popular and frequently used combination in its niche. For its high-quality support, an understanding of both products is desirable. At the same time, in practice, a support specialist usually either specializes in administering MS SQL Server and is not familiar with the features of the 1C: Enterprise 8 server, or, conversely, specializes in administering the 1C: Enterprise 8 server and is not familiar with the features MS SQL Server.

This article was written to help both those and other specialists, designed to save your time and draw your attention to the most important details when using these software products together.

To make the information easier to understand, case studies, notes and tips are provided (in italics).

Three-link circuit

As the reader may already know, the database in the case under consideration has a three-tier architecture:

Link 1: MS SQL Server DBMS. “Stores” and maintains the database, ultimately performing all types of database operations. Thus, the performance of the database, the speed and parallelism of reading and writing data are largely determined by the performance of MS SQL Server.

Link 2: Server "1C: Enterprise 8". Serves as an intermediary in the interaction between clients (users) and MS SQL Server. All client requests are sent to the server, which “translates” them into the MS SQL Server query language, receives the results of executing these queries, and sends the results to the client.

There is only a small part of the operations that are performed at the 1C: Enterprise 8 server level, without accessing MS SQL - this is, in particular, tracking the so-called “managed locks”, reading and writing “session parameters”. In such cases, access to the DBMS is not required, since these operations are performed not with database data, but with auxiliary server information.

Link 3: Client part of "1C: Enterprise 8". Accesses the 1C: Enterprise 8 server, receives results from it (that is, for example, data samples), and is responsible for the user interface.

"I wanted the best."

After reinstalling the 1C: Enterprise 8 server, users complain of a sharp drop in performance. The 1C: Enterprise software implementation specialist who carried out the reinstallation is only surprised - they say he wanted the best, the system should have started working faster... Analysis of the situation showed that the 1C: Enterprise 8 server was allocated too many resources: it processes (see point 3) rphost occupied 15.5 GB of the server’s 16 GB of RAM, as a result, there was practically no available RAM for the compliant MS SQL Server.

The result is a constant "swap", an unnecessary load on the disk subsystem, and extremely slow execution of database operations - due to the fact that MS SQL Server does not have time to process requests coming from the "overclocked" 1C: Enterprise 8 server.

Product Compatibility

Current information about versions of MS SQL Server recommended for use in conjunction with 1C: Enterprise 8 should be found at this link http://v8.1c.ru/requirements/.

At the time of writing this article, 1C developers recommend the following options:

      1. SQL Server 2008 R2.
      2. SQL Server 2008, requires installation of Service Pack 1 (SP1).
    3. SQL Server 2005, requires installation of Service Pack 3 (SP3).



It is technically possible, but not recommended, to use MS SQL Server 2000; it requires installation of Service Pack 2 (SP2), and installation of Service Pack 4 (SP4) is desirable.

Please note that this version is currently deprecated and does not have a 64-bit version for x86-64 architecture.

Note:

It is necessary to pay attention to the operating system settings: for example, for effective operation of M SQL Server 2008 under Server 2008R2 OS, it is necessary to disable the balanced power supply mode and switch to maximum performance mode.

Installing the client-server version of "1C: Enterprise 8"

"1C installed"

For one of the customers, the installation of 1C: Enterprise 8 was carried out by a system administrator who had no experience working with 1C: Enterprise 8. And although, according to him, he “installed 1C”, there was no client part on the user computers, and there was no server part on the server. Analysis of the situation clarified the picture - the 1C: Enterprise 8 kit included 2 disks - installation of the platform and installation of database templates. The administrator did not delve into the installation procedure - and installed database templates, rather than executable files, platform components.

Of course, this is an atypical example of an exceptionally inattentive attitude to work.

When installing "1C: Enterprise 8" you should take into account that the following are installed separately:

      The 1C: Enterprise 8 platform is an executable application, an integrated environment for the development and operation of databases. When you launch it, you select one of two operating modes - “Enterprise” (user database shell) or “Configurator” (integrated development environment). A more complete description can be read at the link
      "1C: Enterprise" configuration templates are a file of the platform's internal format, with the help of which the platform can create a pure or demo database of the structure contained in the template. You can also use the update pattern to update the structure of an existing database that is already filled with data.
      When installing the platform, you should pay attention to the selection of components:





The 1C: Enterprise component may not be installed on the server(s).

In this case, the server will provide client computers with access to 1C: Enterprise databases, but working with the database in user mode directly from the server will not be possible.

Note:

The 64-bit version of the platform does not contain a client part. Therefore, when installing on a server, 64-bit server components are installed separately, and 32-bit client application components are installed separately.

The "1C Server: Enterprise" component is needed to connect to MS SQL Server - it is an application server, a connecting link between the platform on client workstations and MS SQL Server.

It is possible to install the component in the mode of a simple application or system service, and, of course, the second option is recommended.

When installed "as a service", this component will be launched and executed on behalf of the selected user:




After loading, the component spawns several processes, such as: “server agent”, “server cluster manager”, “server worker processes”.

Database queries are executed by worker processes, and the server cluster manager distributes the load between them.

Server work processes can be managed (added, deleted, set limits on RAM usage, declared primary or backup) if the 1C: Enterprise Server Administration component is installed.



Note:

For the 32-bit version of the server, it is recommended to install working processes in such a number as not to leave RAM unused - each of them has a noticeable limitation on the use of RAM, from 2 to 4 GB depending on the system configuration.

For the 64-bit version of the server, two worker processes are theoretically sufficient - one worker and one backup. However, in practice, to ensure reliability and stability of connections for a significant (several hundred) number of users, a larger number is required, it depends on many factors - on the number of users, the content of the database and the volume of queries performed, so the authors believe that the number of processes in this case should be selected experimentally.

"Ouroboros"

After unsuccessful optimization of the 1C: Enterprise 8 server settings, users reported extremely slow system operation, and the system administrator noted a constant 100% processor load on the server.

An analysis of the situation showed the source of the problem - during configuration, too small a limit was set on the use of RAM by working processes.

But the fact is that this limitation works as follows:

When the server cluster manager sees that a worker process has exceeded the RAM limit, the process is terminated, it is disabled, a new worker process is created, and connections and user requests are redistributed between worker processes.

The limit set was so small (300MB) that the worker process could not fully service even one intensive user - as a result, the server cluster manager was constantly restarting worker processes and reconnecting users. As soon as a new process was created and users connected to it, the RAM limit was almost instantly reached and caused the next restart. This took 100% of the processor load.

The "1C Server: Enterprise" component is not needed on client workstations, and will not be able to start there, since it requires the physical presence of a security key.

If the number of connected users is small (less than 50), the application server is usually installed on the same computer where MS SQL Server is running.

For systems with a large number of users and/or a large volume of information flows, separate installation is recommended, as well as the use of a server cluster.

The "1C: Enterprise Server Administration" component can also be useful on clients - for example, with its help you can see a list of infobases connected to a given "1C: Enterprise" server.

It is strongly recommended to install it on the server itself.

Access

Note:

To verify that access is provided, it is not enough to use the 1C: Enterprise server administration utility, and even more so, the presence of the server in the “Network Neighborhood” is not enough!

It is necessary for each client to log in to the database installed on the server - only this will give 100% confidence that access is provided.

1. Depending on the security policies, MS SQL Server uses Windows account authentication or MS SQL Server account authentication.




In the latter case, when creating a 1C: Enterprise database, the system will request the login and password of the MS SQL Server account (for example, sa), in the first case the login and password should be left empty:



and the system user on whose behalf the 1C: Enterprise server is running must be given rights in MS SQL Server, namely:

      full rights to the database in which the information base is located
      access to the master database (public role)
      recommended - rights to create a database, otherwise each new database will need to be first created using MS SQL Sever, and only then connected to the 1C: Enterprise server
      recommended - right to delete your database



For example, you can assign the user in question the fixed role processadmin or sysadmin.

Advice.

If all users simultaneously lose access to the working database, you need to double-check the user rights and roles in MS SQL Server, including those set for a specific database, that is, User mapping:




2. Server 1C: Enterprise accesses MS SQL Server through the Microsoft Data Access mechanism, so its components must be installed, and the user of the server 1C: Enterprise (see the previous paragraph) must have rights to run them.

3. Communication between clients and server is supported by TCP protocol, so it is necessary that this protocol is supported by both sides. There may be problems matching the server name and its IP address, for example, if a peer-to-peer network is used. In this case, you should record the correspondence in the file [C:\WINDOWS\] system32\drivers\etc\hosts .

Advice.

If the network is peer-to-peer, to ensure a permanent connection to the server, create a network drive that accesses any of the folders of this server.

4. If the Named Pipes protocol is used, and if MS SQL Server and the 1C: Enterprise server are installed on different computers, the user on whose behalf the 1C: Enterprise server is running must be registered in the list of users of the computer on which MS SQL Server is running.

5. In some cases, additional configuration of the Windows firewall may be required, that is, adding exceptions.

6. Some antivirus programs may block “unwanted” network traffic, so you may need to add to their exclusion lists.

7. The release of the 1C: Enterprise 8 platform must be absolutely identical on the client and on the server.

"Twins"

"One of the customers used two database servers, each of which housed one working database. Users worked - each simultaneously with both databases. The support service updated the 1C: Enterprise 8 platform on servers and clients.... And then complaints started pouring in for the impossibility of connecting to one or another database. An analysis of the situation showed that several people did the update on clients and servers, and the installing specialists did not double-check that they were installing the same release. Therefore, there was one release of the platform on one server. the second - another, on half of the clients - the first of these releases, on the other half - the other. It turned out that each user has access to only one of the databases.

To quickly solve the problem, each user had to install both releases of the platform and create separate shortcuts to log into each database.

Initial settings of MS SQL Server and database

“And that’s how it works”

MS SQL Server is distinguished by its simple initial installation, so not all administrators bother with its additional configuration - after performing the default installation, the database is working, users are logged in - the job is done. This approach almost always entails the emergence of problems after about a month or two - and, of course, suddenly and at the most inconvenient moment.

For example, if the database is intended for accounting purposes, before submitting tax reports, there is often a need to urgently recalculate certain data, and recalculate en masse, say, “all receipts of fixed assets from the beginning of the year.” Moreover, during the working day, without stopping the work of other database users.

And, of course, it is at this moment that it will be discovered that during such a recalculation the database “freezes”, or “crashes”, or does not allow other users to work.

This kind of “Murphy’s Law” applies to each of the following points.

Before starting to use MS SQL Server as a DBMS for 1C: Enterprise, it is recommended:

1. Set the value of the max degree of parallelism parameter to 1.

That is:

      After connecting to the server, enter the server properties through the context menu, Properties item
      then select the Advanced page and edit the max degree of parallelism parameter






Otherwise, some queries generated by the 1C: Enterprise server may cause the error "Intra-query parallelism caused your server command (process ID #XX) to deadlock. Rerun the query without intra-query parallelism by using the query hint option (maxdop 1 )". After this error, the client part often crashes.

The error will not appear consistently, since the query plan is formed differently depending on the accumulated statistics - it will manifest itself on large and complex queries, that is, at the most unfortunate moment.

2. Create a Maintenance Plan that shrinks the tempdb temporary table database on a nightly basis. The database of temporary tables is not always cleared automatically by the 1C: Enterprise server, and sometimes, as a result of an unsuccessfully written query, a temporary table of, for example, 50 GB in size may be created and not cleared. As a result, disk space may run out, as a result of which both the client and server parts may crash, and there is also a slight risk of data integrity violation.

That is, you need:

      go to MS SQL Management Studio
      after connecting to the server, expand the "Maintance plans" section
      create a new (or add to an existing) Service Plan,
      add the item "Execute T-SQL Statement task" to it (since you cannot select the tempdb database in the "Shrink database" task) with the code




1.USE
2.
3.GO
4.
5.DBCC SHRINKFILE (N"tempdev" , 0, TRUNCATEONLY)
6.
7.GO
8.
9.DBCC SHRINKFILE (N"templog" , 0, TRUNCATEONLY)
10.
11.GO

Note that the temporary table database file name may not be "tempdev". You can use a script to check this name

1.USE tempdb
2.
3.GO
4.
5.EXEC sp_helpfile
6.
7.GO




“Pot, don’t cook”

The most common way in practice to overfill tempdb and thereby crash the server is to forget to specify a condition when joining tables.

Namely, let’s say we have two tables in the database, each with 20 thousand records in size. Let's say there is a one-to-one correspondence between their records, and we write a query that creates a temporary table that contains 20 thousand records with fields from both source tables. But if we forget to specify the join condition, every record of the first table will be joined to every record of the second! That is, the resulting table will consist of 20’000 * 20’000 = 400 million records. And so on.

3. In order to reduce the load on the disk subsystem, it is recommended, if possible, to distribute the working database and tempdb, logs, and system swap file across different physical disks.

It is better to set the desired path for storing the working database files when creating it by editing the Path column:




To change the physical location of temporary table database files, use the ALTER DATABASE command, that is, in MS SQL Management Studio you need to run the following script ("New query" command)

1.USE master
2.
3.GO
4.
5.ALTER DATABASE tempdb
6.
7.MODIFY FILE (NAME = tempdev, FILENAME = "New_Disk:\New_Directory\tempdb.mdf")
8.
9.GO
10.
11.ALTER DATABASE tempdb

12.
13.MODIFY FILE (NAME = templog, FILENAME = "New_Disk:\New_Directory\templog.ldf")
14.
15.GO

4. The “growth” of the working database and its log should not be hampered – there should be no size limit, the “Autogrowth” property should be set as a percentage, the recommended value is 10%. Otherwise, adding data to the database, restoring from an archive, and other operations may take an unreasonably long time.

To set this property, you need to go to the database properties through the context menu, select the Files section, and open editing file properties:



5. It is recommended to enable support for the TCP/IP network protocol in MS SQL Server and disable all others, otherwise the joint operation of MS SQL Server and 1C: Enterprise server will be less stable.




6. In the same place - clear the Alias ​​section, because its installation leads to errors in the interaction between MS SQL Server and the 1C: Enterprise server.

Before starting to use the database, it is recommended:

1. When creating a database from "1C: Enterprise", set the "date offset" to 2000, otherwise an attempt to record a date earlier than 01/01/1753 (which is possible due to human factor) will cause failures in the database.

Attention! The date offset cannot be changed for an existing database!



2. Set the Recovery model to Simple, or create a Maintenance Plan, which will create a daily backup of the database and trim the transaction log (log file). Otherwise, during some operations the transaction log (log file) will grow very quickly: for example, when restructuring a database, the growth in the size of the log file can be several times larger than the size of the database itself.




3. Create a Maintenance Plan that performs the following routine tasks at least once a week:

      Creating a backup copy of the database.
      Update database statistics and clear the procedural cache (note that the autoupdate statistics property does not imply clearing the procedural cache).
      Clearing the procedural cache is not included in the standard operations of Maintenance Plans; this step must be defined as executing a script (Execute T-SQL Statement) with the following content:
      DBCC FREEPROCCACHE
      Reindexing database tables.






Of course, it makes sense to set up automatic sending of emails about successful/unsuccessful completion of tasks.




Conclusion

The issues that most often cause difficulties for system administrators and implementers of 1C: Enterprise 8 are considered in connection with the joint use of MS SQL Server and the client-server version of 1C: Enterprise 8.

The author hopes that he has covered “both sides of the coin” in a fairly consistent and accessible way.

P.S. Make backups often!

Client-server version of work- one of the options for operating the 1C:Enterprise 8 system.

The client-server version of the work is intended for use in workgroups or on an enterprise scale. It is implemented based on a three-tier client-server architecture.

The client-server architecture divides the entire working system into three different parts that interact with each other in a certain way:

The program running for the user (client application) interacts with the 1C:Enterprise 8 server cluster, and the cluster, if necessary, accesses the database server.

In this case, the physical cluster of 1C:Enterprise 8 servers and the database server can be located either on the same computer or on different ones. This allows the administrator to distribute the load between servers if necessary.

Using a 1C:Enterprise 8 server cluster allows you to concentrate the most extensive data processing operations on it. For example, when executing even very complex queries, the program running for the user will receive only the selection it needs, and all intermediate processing will be performed on the server. Typically, increasing the capacity of a server cluster is much easier than upgrading the entire fleet of client machines.

Another important aspect of using a 3-tier architecture is ease of administration and streamlining user access to the information base. In this option, the user does not need to be aware of the physical location of the configuration or database. All access is carried out through a cluster of 1C:Enterprise 8 servers. When accessing a particular infobase, the user must specify only the cluster name and the infobase name, and the system requests the user name and password, respectively.

1C:Enterprise 8 uses the capabilities of a database management system to efficiently retrieve information:

  • the query mechanism is focused on maximizing the use of the DBMS to perform calculations and generate reports,
  • viewing large dynamic lists is ensured without performing a large number of database calls; at the same time, the user is provided with effective search capabilities, as well as selection and sorting settings.

Deploying the client-server option and administering it is quite simple. For example, the creation of a database is carried out directly during the launch of the configurator (the same as for the file version).

Client Applications

Working in a client-server version is possible either directly with the cluster or through a web server. Moreover, in the case of direct connection to the cluster, the thick client and thin client use the TCP/IP protocol. When connecting via a web server, the thin client and web client use the HTTP or HTTPS protocol.

Server cluster

The 1C:Enterprise 8 server cluster is the main component of the platform, ensuring interaction between users and the database management system in a client-server mode. The presence of a cluster allows you to ensure uninterrupted, fault-tolerant, competitive operation of a large number of users with large information databases.

Database server

The following can be used as a database server:

Server cluster administration

The platform includes a set of various tools that allow the administrator to manage the composition of the cluster, information databases and user connections.

Performing basic functionality on the server

All work with application objects, reading and writing the database is performed only on the server. The functionality of forms and command interface is also implemented on the server.

The server prepares form data, arranges elements, and records form data after changes. The client displays a form already prepared on the server, enters data and calls the server to record the entered data and other necessary actions.

Similarly, the command interface is formed on the server and displayed on the client. Also, reports are generated entirely on the server and displayed on the client.

At the same time, the platform mechanisms are focused on minimizing the amount of data transferred to the client computer. For example, data from lists, tabular sections and reports is not transferred from the server immediately, but as the user views it.

The server runs:

  • Database queries
  • Data recording,
  • Carrying out documents,
  • Various calculations
  • Performing processing,
  • Generation of reports,
  • Preparing forms for display.

Runs on the client:

  • Receiving and opening forms,
  • Form display,
  • “Communication” with the user (warnings, questions...),
  • Small calculations in forms that require a quick response (for example, multiplying price by quantity),
  • Working with local files,
  • Working with commercial equipment.

Using the built-in language on the client

You can manage the functionality of forms not only on the server, but also on the client. The client supports the built-in language. It is used in cases where it is necessary to carry out calculations related to the form displayed on the screen, for example, to quickly (without contacting the server) calculate the amount of a document line based on price and quantity; ask the user a question and process the answer; read a file from the computer's file system and send it to the server.

However, the operation of the built-in language on the client is supported to a strictly limited extent. Client procedures in modules are explicitly separated from server procedures, and they use a limited composition of the object model of the built-in language.

Direct work with the database is not allowed on the client. Working directly with application objects is not allowed; for example, such types of built-in language as DirectoryObject.<имя> . Requests are not allowed. If you need to call actions with data in client code, you need to call server procedures that will already access the data.

Beta Menu

Setting up Server 1C:Enterprise 8.3 and PostgreSQL 9.4.2-1.1C. Full instructions

In this manual we will tell (and show) how to set up a connection 1C:Enterprise 8.3 And PostgreSQL 9.4.2 from the moment of installation of both services, right up to the creation of the information base. You can read about tuning this combination in.

The stages we have to go through:

  1. Installation Servers 1C:Enterprise(64-bit) for Windows
  2. Installation PostgreSQL 9.4.2-1.1C
  3. Creation of an Information Database.

More details under the cut!

Stage 0. Input data.

Server name - 1CServer
Server account name - Administrator
Account password - 123456Ab

1C account name on the server - USR1CV8
 1C account password on the server - 123456Cd

The PostgreSQL account name on the server is postgres
The password for the PostgreSQL account on the server is 123456Ef

PostgreSQL superuser name - postgres
PostgreSQL superuser password - 1234

Test Database Name − testdb

Stage 1. Installation Servers 1C:Enterprise(64-bit) for Windows


Stage 2. Install PostgreSQL and pgAdmin.


Stage 3. Creation of a 1C information base.

  1. Before performing the following operations, disable IPv6 on your network interface: Network and Sharing Center > LAN connection > Properties> Uncheck Internet Protocol Version 6 (TCP/IPv6).
  2. We launch the client 1C:Enterprise and add a new database.
  3. Creating a new infobase > Creating an infobase without configuration (for example, you can have any configuration) > On the 1C:Enterprise server >
  4. We fill in all the fields in accordance with our example (Stage 0):
    Server cluster 1C:Enterprise: 1CServer
    Name of the infobase in the cluster: testbd
    Secure connection: Turned off
    DBMS type: PostgreSQL
    Database server: 1CServer
    Database name: testbd
    Database user: postgres
    User password: 1234
  5. Further, further. We launch the created database in enterprise mode - everything works!

We remind you once again that PostgreSQL can be overclocked quite well. Details in.
And don't forget about 1C!
If there are any problems with the database, perhaps this will help you.
1C databases are possible!

Still have questions?

Something went wrong? Our company’s specialists will help you resolve any problems that may arise! Contact us! →

People constantly ask on forums on the Internet:
-How to install 1C:Enterprise 8.1 (8.2) on PostgreSQL on the Windows operating system
-Is it possible to install simultaneously two processes “1C:Enterprise Server Agent 8.1” and “1C:Enterprise Server Server Agent 8.2”. If yes, how to do it?
-After installing 1C and PostgreSQL the following errors are displayed:
A)



b) Error when creating an infobase:
An error occurred while performing an operation with the infobase.

c) Error when creating an infobase:
An error occurred while performing an operation with the infobase.


If you are faced with such problems and are looking for an answer, then this article is just for you.

1. Installation 1C:Enterprise 8.2

Starting installation of 1C:Enterprise 8.2.

Selecting components for installation; in addition, we will need:
-Server 1C:Enterprise
-Server administration 1C:Enterprise

Choosing Russian language

Next, you need to create a user on whose behalf the Server Agent service will be launched. So, check the box to install 1C server: Enterprise as a Windows service (Recommended). And we create a user. Even if you specify an existing one and accidentally set the wrong password, the service simply will not start after installation. And the installation itself will complete correctly, you need to click continue in the service startup window.

Let's start the installation

Installing the security driver: The security driver must be installed on the computer into which the physical key is inserted. Most often, the key is installed on the server, we will do the same and leave the default driver installation checkbox - install.

We complete the installation.

Find "1C: Enterprise Server Agent", go to properties:

And on the “login” tab, switch the login type to “with a system account”.

2. Installing PostgreSQL for 1C:Enterprise.

We start the PostgreSQL installation. The distribution kit itself must be downloaded from the official 1C website, because there is already a version configured to work with 1C.

We start the installation, select the Russian language, and click next.

The choice of components to install is left as default.

Cluster initialization.

If it is necessary for the server agent to be shared with the SQL server, check the box to support connections from any IP, not just localhost.
ATTENTION! We set the encoding: UTF-8 for both the server and the client.

Here we leave everything as default and click next again.

Additional modules can be left by default, if necessary, you can mark the required ones.

Starting the PostgreSQL installation

We complete the installation.

For the PostgreSQL service, I also usually switch the startup type - with a system account.

3. Creating an information base on the pgsql server and common errors that occur at this stage.

We go to 1C, add a new database, say that the database will be on the server and indicate the server parameters:

Attention. The database must be created specifically from 1C, otherwise if you create it from PgMyAdmin, then when you try to enter the database you will receive errors:

Error when creating an infobase:
An error occurred while performing an operation with the infobase.
DBMS error: ERROR: language "plpgsql" does not exist.
HINT: use CREATE LANGUAGE to load the language into the database.

Even if you go to pgMyAdmin and add the language "plpgsql" for the database, the following error will appear.

Error when creating an infobase:
An error occurred while performing an operation with the infobase.
DBMS error: ERROR: type "mvarchar" does not exist at character 31

SOLUTION: You need to create a database from 1C.

If, when creating a database from 1C: Enterprise, you received the following message:

An error occurred while performing an operation with the infobase.
DBMS error: ERROR: new encoding (UTF8) is incompatible with the encoding of the template database (WIN1251).
HINT: Use the same encoding as in the template database, or use template0 as template.

SOLUTION: When installing PostgreSQL, you left the Windows-1251 language for the server or client, which is installed by default, but you needed to change it to UTF-8

Reinstall PostgreSQL and specify UTF-8 language during installation.

4. Launching the 1C:Enterprise server agent 8.1 and

1C:Enterprise 8.2 server agent simultaneously on one computer.

With the advent of the 8.2 platform, many people began to have a problem when they needed the 8.2 platform to run on the server, but at the same time they would like to keep 8.1.

I propose one way to install two server agents, through port splitting.

First, install Platform 8.1, as described above. Next, install Platform 8.2. Now we should have 2 server agent services appear in the list of services. But when starting the second service, the following message will be displayed:

The "1C: Enterprise 8 Server Agent" service on the "local computer" was started and then stopped. Some services stop automatically if they are not in use by other services or programs.

This occurs due to the fact that the system is trying to run 2 worker processes simultaneously that operate on the same port.

To get out of this situation, let’s change the port, for example, Server Agent 8.1.
To do this, you need to stop the 1C:Enterprise 8.2 Server Agent service.
Launch 1C:Enterprise Server Agent 8.1.

Registry branch: HKLM\SYSTEM\ControlSet001\services\1C:Enterprise 8.1 Server Agent
Parmert: ImagePath
Value: "C:\Program Files\1cv81\bin\ragent.exe" -srvc -agent -regport 1641 -port 1640 -range 1660:1691 -d "C:\Program Files\1cv81\server"

After this change, we should have 2 1C Server Agents running simultaneously: Enterprise 8. We launch them accordingly.

Now we need to add a new server cluster. We do everything in working order:

If we don’t have a central 8.1 server, add it, but with port 1640:

Add a server cluster, change the port to 1641:

Now our 8.1 cluster will work via port 1640, and for 8.2 via 1540 (default port):

Adding workflow 8.1. ports accordingly also change to 16xx

We start the process, do not forget to check the box - the process is enabled.

Now everything should work for us.

The last thing that needs to be said: when creating a database on 8.1, you need to specify the port, because otherwise, the default port (1541) will be used and the program will write about platform version incompatibility.

RELATED ARTICLES