epdm高级培训系统管理维护solidworks hardware and network best practices_第1页
epdm高级培训系统管理维护solidworks hardware and network best practices_第2页
epdm高级培训系统管理维护solidworks hardware and network best practices_第3页
epdm高级培训系统管理维护solidworks hardware and network best practices_第4页
epdm高级培训系统管理维护solidworks hardware and network best practices_第5页
已阅读5页,还剩22页未读 继续免费阅读

付费下载

下载本文档

版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领

文档简介

1、SolidWorks Enterprise PDMHardware and Network Best PracticesJustin Webster, Product Manager, InFlow TechnologyTim Kwong, SolidWorks CorporationInternational Call in numbersAustralia: +61 (0) 2 8014 9491Austria: +43 (0) 7 2088 1053Belgium: +32 (0) 28 08 4374Canada: Denmark: +45 69 91 89 34Finland: Fr

2、ance: +33 (0) 182 880 178Germany: +49 (0) 898 7806 6470Ireland: +353 (0) 19 030 016Italy: +39 0 521 18 57 113Netherlands: +31 (0) 208 080 225Access Code: Webinar ID: New Zealand: +64 (0) 4 974 7248Norway: +47 21 04 30 58Spain: 81 6722Sweden: +46 (0) 852 500 192Switzerland: +41 (0) 225 3311 27United

3、Kingdom: +44 (0) 203 318 6075United States: AgendaIntroductionImplementation PlanningHardwareNetworkSQL Best PracticesCustomer “Frequently Asked Questions”TroubleshootingSQL ConnectionNetwork ConnectivityOther considerationsWorking with ISPsWAN AccelerationIntroductionPurposeProvide basic informatio

4、n to help plan and troubleshoot Enterprise PDM implementations. Should be used as a guideline to start a dialog with the customerPresentersJustin Webster InFlow TechnologyTim Kwong SolidWorks CorporationImplementation Planning - HardwareHardware Selection Services that are installed on a physical se

5、rver:SolidWorks Enterprise PDM Archive ServerSolidWorks Enterprise PDM Database ServerSolidWorks Enterprise PDM Web ServerMS SQL ServerImplementation Planning - HardwareArchive Server ServiceMaintains file vaults and provides client access to SolidWorks Enterprise PDM. Sends the files from archive s

6、erver to the clientSends files from archive server to archive server during replicationDatabase Server ServiceHelper service for SolidWorks Enterprise PDM file vault databases on the local SQL Server. The service connects to SQL, but can be installed on another server.This service processes jobs, su

7、ch as email notifications and replication. The service logs into SQL and queries for new jobs to process. This service can be restarted while users are using the system. When the service is restarted, it will begin processing jobs in the SQL queue.SQL ServerProvides storage, processing and controlle

8、d access of data, and rapid transaction processing. In short, this service controls how database information is delivered to the EPDM Clients, the database server process, the archive server process, and the web server.This service contributes the most to EPDM performance and utilizes the most proce

9、ssor and RAM. Generally speaking, you want to dedicate most of your hardware resources to this service.Implementation Planning - HardwareArchive Server ServiceMaintains file vaults and provides client access to SolidWorks Enterprise PDM. Sends the files from archive server to the clientSends files f

10、rom archive server to archive server during replicationDatabase Server ServiceHelper service for SolidWorks Enterprise PDM file vault databases on the local SQL Server. The service connects to SQL, but can be installed on another server.This service processes jobs, such as email notifications and re

11、plication. The service logs into SQL and queries for new jobs to process. This service can be restarted while users are using the system. When the service is restarted, it will begin processing jobs in the SQL queue.SQL ServerProvides storage, processing and controlled access of data, and rapid tran

12、saction processing. In short, this service controls how database information is delivered to the EPDM Clients, the database server process, the archive server process, and the web server.This service contributes the most to EPDM performance and utilizes the most processor and RAM. Generally speaking

13、, you want to dedicate most of your hardware resources to this service.Implementation Planning - HardwareRAM and Processor ConsiderationsSQL Server demands the most of RAM and ProcessorRAMRAM is consumed as SQL processes Enterprise PDM transactions or queriesSQL Server is most efficient with the ent

14、ire database loaded into RAMIf your database is 4GB in file size, SQL will perform best if 4GB RAM can be dedicated to SQL.32-bit machines are limited to 1.7GB RAMOnce RAM reaches limit, additional requests are processed by reading/writing to disk (affecting performance)Processor and OSThe Processor

15、 with execute the SQL requests once cached in RAM32 bit limits the amount of RAM that can be allocated to SQLWindows Standard Edition 64 bit or Windows 2008 Server R2 supports 32 GB of RAMArchive server and database server services are not largely affected by RAM and processor. It is always mended t

16、o put SQL Server on a dedicated machine.Implementation Planning - HardwareDisk Space ConsiderationsThe archive server retrieves files from the archive foldersDisk SpaceInitial Disk size should be based on their current data set times a revision factorAlways size the disk for the future growthTo calc

17、ulate the vault size growth over a range, see Solution S-046620To estimate Disk size and processor, use Customer SamplesImplementation Planning - HardwareWill hardware make a difference?Customer BenchmarkCurrent Server:32-bit Windows Server 2003 Standard SP24GB RAMSQL Server 2005 Standard SP2 Propos

18、ed Server:64-bit Windows Server 2008 Standard SP224 GB RAMSQL Server 2008 Standard R2Hardware can be the difference between a good and bad implementation.Search by File Name Search by Part NumberSearch by Doc Number55%57%66.0%Customer Hardware BenchmarkTime reductions with new hardwareImplementation

19、 Planning - NetworkNetwork ConsiderationsLatency and BandwidthLatencyLatency refers to the length of time to deliver information to a remote site. Latency is measured inmilliseconds using the “ping” function. A ping sends 32 bytes of information to the remote site and then 32 bytes are returned. The

20、 total time of the round trip is called the Latency.ping servername t (the “-t” switch runs a repeated ping to check packet loss)BandwidthBandwidth is the total amount of data that can be sent over a network. It is measured in Megabits per second (Mbps) or in Megabytes per second (Mbps, equal to MBp

21、s x 8). The following chart shows the bandwidth for different devices.DeviceBandwidth (Mbps)Bandwidth (MBps)T11.540.193T26.310.789T344.745.592T4274.1834.272T5400.3550.04410 Gigabit Ethernet9953.001244.125Mbps = MBps x 8Implementation Planning - NetworkNetwork ConsiderationsThroughputThroughput is no

22、rmally considered the “actual” bandwidth. Since no network is 100% efficient, Bandwidth can be less useful than throughput. Throughput can be calculated using the following formula:Throughput = TCP Buffer Size (KB)* / Latency(ms) x 8(constant)Example:Throughput = 64KB / 30ms * 8 = 17 Mbps = 2 MB/sSe

23、e Solution S-017952 to review options for measuring RAW BandwidthImplementation Planning - NetworkNetwork ConsiderationsWorking with the customerCustomers will often buy hardware but cannot change their network. Questions:What is the latency of each WAN connection?Do they see packet loss across the

24、connection?What type of connection exists between sites? (i.e. T1, shared T1)What other applications share the bandwidth?Potential ConcernsLatency over 250msPacket loss and service interruptionLow ThroughputListen to the customerif they are concerned that their network will not support PDM, make sur

25、e you perform the necessary tests. Do not just assume that it will be fine.If the customer has some network issues, make sure to set expectations.Implementation Planning - NetworkNetwork ConsiderationsBuild a network diagram SQL Configuration Best PracticesMemory AllocationAddress Windows Extension

26、(AWE) for 32-bit WindowsFor Windows Server 2003 and 2008 Enterprise and Datacenter version, SQL Server has the capacity to handle more than 4 GB of memory. You need to enable the AWE feature to take advantage of greater memory handling capacity. AWE is not required to be set on 64-bit OSs.SQL Config

27、uration Best PracticesSetting and Maintaining the Database and Transaction Log filesThe MDF and LDF can e fragmented if they are not set and managed properly.Dont depend on Autogrowth to manage the MDF and LDF files, because it can lead to fragmentation. Dont restrict growth of LDF fileSQL Configura

28、tion Best PracticesHard Drive ConfigurationLocal DisksConsider separating the Windows OS, database files (MDF), and log files (LDF) on its own RAID setup. For Example: Windows OS in RAID 1, database files in RAID 5 or 10, and log files in RAID 1.SANConsider separating the Windows OS, database files

29、(MDF), and log files (LDF) on its own LUN Logical Unit Number (LUN) setup.For large database, also try to put the Tempdb database file in a separate RAID or LUN setup.Hard Drive SpeedWhen considering hard drive for a server, try to select a configuration that has the most drive bays. Also try to cho

30、ose hard drives having 15K RPM if possible, because faster drives (spindles) result in better performance. Processor AffinitySQL Server supports processor affinity using the CPU affinity mass and affinity I/O mask. This feature can be used to dedicate the processors for SQL use only. Setting the mas

31、k manually can enhance performance for server, but it is usually done for large SQL deployment running multiple instances of SQL Server. For most Enterprise PDM deployments, use the automatic settings for Affinity mask for processors and I/O. Dont adjust the affinity on a production server, because

32、it can also degrade performance.SQL Configuration Best PracticesMaintenanceShrinking DatabaseShrinking a database is just removing unused pages. There is really no need to shrink a database unless large amounts of data (files) are deleted. Make sure to run a rebuild of indexes should after shrinking

33、, because the shrinking process can fragment the indexes.DailyIt is critical to perform daily backup of the database and log files. Make sure to offload the backup files from the SQL server if the backup files are saved locally on the SQL Server.For Full Recovery mode, backup both SQL database and t

34、ransaction log file. The transaction log can grow substantially when left unattended and possibly consume all the hard drive space when there is no restriction set for Autogrow. A nightly backup of transaction log will free up (truncate) log space for future need. The backup process will not shrink

35、a log file size, but there is really no need to shrink the log file if it is backed up regularly unless the initial size was setup too large.For Simple Recovery mode, just backup the SQL database. Backup cant be performed on the transaction log, because SQL Server truncates it regularly. Weekly or M

36、onthlySchedule a reorganize of indexes weekly and rebuild of indexes monthly using SQL Maintenance Plan Wizard. There is no need to do a reorganize before a rebuild process, because the rebuild process will drop all the indexes before rebuilding. See Solution S-042552.Troubleshooting Hardware and Ne

37、tworkClient Connection IssuesCould be SQL Server or network relatedBenchmark operations from local and remote sitesClient PerformanceCould be SQL Server or network relatedBenchmark operations from local and remote sitesBenchmarks should be done at different points during the dayInclude SQL Profiler

38、and Network Monitoring in your testingTroubleshooting Hardware and NetworkSQL ProfilerUsing profiler you can see how much time it takes to perform SQL operations and output an entire trace to a log for further reviewTroubleshooting Hardware and NetworkNetwork MonitoringThere are many tools that you can install to measure actual throughput on a workstationMake sure you measure your

温馨提示

  • 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
  • 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
  • 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
  • 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
  • 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
  • 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
  • 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

评论

0/150

提交评论