Our recommendations

• For compilers that perform multiple checkouts/checkins in a second, we recommend 25%-50% stronger hardware.
• VM Administrators should make sure that the hosting server is capable of accommodating the required resources.
• When seeing a low performance in DB queries, please check the disk queue.
• We strongly recommend placing the DB in the same Data Center as OpenLM Server.
• See recommendations for MS SQL Server below.
• For MySQL, we provide a sample configuration file for Windows (my.ini) & Linux (my.cnf) that should be revised by DBA.
• VM network controller should be available for each network card.
• In case of a big DB (from 25GB and up) and big load, each DB should have 3 files and 3 VM disk controllers for the: db file, log file and tmp file.

Best practices for using MySQL

1. Use the latest 5.7/8 MySQL release.
2. In order to utilize the system’s resources MySQL requires its configuration file (my.cnf/my.ini) to be set with correct values. Otherwise MySQL will not take advantage of its hosting machine’s resources. We recommend some settings – please see our suggestions for configuration files archived in .zip format according to your system size:

4GB_2Cores_Windows 

8GB_4Cores_Windows

16GB_8Cores_Linux

16GB_8Cores_Windows

24GB_8Cores_Windows

Best practices for using MS SQL Server

1. Customers need to apply a maintenance plan consisting of:
a) Periodic Statistics Update
b) Periodic Rebuild or Reorganization of Indexes
DBAs need to apply company maintenance policy also for OpenLM DB. In case such does not exist, a public package can be applied (Here is one).

2. Recommended memory allocation to MSSQL Server running (almost) exclusively on a Windows machine is no more than 80% of total machine memory.

3. OpenLM database should have the is_read_committed_snapshot_on parameter set.

To check if it is set:

SELECT is_read_committed_snapshot_on FROM sys.databases
WHERE name= 'YourDatabase'

To set:

DECLARE @sqlCommand varchar(1000)
DECLARE @db_name varchar(50)

SET @db_name = 'YourDatabase'

SET @sqlCommand = 'ALTER DATABASE ' + @db_name + ' SET ALLOW_SNAPSHOT_ISOLATION ON '
EXEC (@sqlCommand)

SET @sqlCommand = 'ALTER DATABASE ' + @db_name + ' SET SINGLE_USER WITH ROLLBACK IMMEDIATE '
EXEC (@sqlCommand)

SET @sqlCommand = 'ALTER DATABASE ' + @db_name + ' SET READ_COMMITTED_SNAPSHOT ON '
EXEC (@sqlCommand)

SET @sqlCommand = 'ALTER DATABASE ' + @db_name + ' SET MULTI_USER '
EXEC (@sqlCommand)

4. For better performance, we recommend installing tempdb, databases and log files on separate logical (and in some cases – even physical) disks.
A solid installation would have:

a) 1- disk for tempdb data (ssd configuration is recommended)
b) 1- disk for system DBs (msdb, model, master)
c) 1- disk for all logs (including tempdb logs)
d) 1- disk for all DBs Data

5. tempdb has a critical role, having all parameters, temporary tables and executing sorts and aggregations. A number of tempdb data files are recommended to be as a number of processors – up to 8 (more will have no effect or bad influence on performance).
6. Autogrowth units of database files is set by default to percent, which is dangerous. A good practice would be to use MB units, based on a predicted growth multiplied by record size. In any case, setting alerts on disk size is recommended.
7. It is recommended to set log size upfront.
8. A regular backup program is recommended in order to be able to resume crashes and control log files growth. Shrinking a database is a bad practice and is not recommended.9. In the case when data loss of 1 day is acceptable – it is OK to use a daily full backup and Simple recovery model. In case when a point in time restore is needed, use a full recovery model and proper schedule for log backups (every 1 hour or every 30 or even 15 minutes depending on the volume of inserted data to avoid log file growth).

9.In the case when data loss of 1 day is acceptable – it is OK to use a daily full backup and Simple recovery model. In case when the point in time restore is needed use a full recovery model and proper schedule for log backups (every 1 hour or every 30 or even 15 minutes depending on the volume of inserted data to avoid log file growth).