|
请使用QQ关联注册PLM之家,学习更多关于内容,更多精彩原创视频供你学习!
您需要 登录 才可以下载或查看,没有账号?注册
x
http://www-cad.fnal.gov/PLMWorld2008/Teamcenter%20Unified/The%2520Network%2520is%2520Slow!.pdf . g% r8 d5 r- E1 p& H+ D7 I
http://www.plmworld.org/p/bl/et/blogid=43&blogaid=218
4 h4 g7 a" H. f+ E3 A8 \9 `( b, n/ _: Zhttps://support.industrysoftware.automation.siemens.com/docs/teamcenter/Network_Performance_Tuning_V6.6.pdf7 a# Q3 |: C' U m, e7 i+ f
p4 l4 A# I8 e/ m0 m3 [
Teamcenter is Slow! Teamcenter Performance Bottlenecks & Mitigation. E& U: f$ K* T- E! \5 {
# T: \8 K0 h% c/ L Y
Larry Carpenter
) B- `6 e0 K/ o; c; }5 SAt the Chicago-Wisconsin RUG Fall Conference in October 2012, I gave a presentation on common Teamcenter performance issues and solutions along with listing, in one place, many other past Teamcenter performance presentations and white papers. It was very well received by a standing room crowd, so I thought I'd share it with the rest of the PLM World community. Here is a link to the presentation (you must be a TCUA SIG member to view it): http://www.plmworld.org/p/do/sd/sid=3758&type=0. C* l# x7 A T5 n3 `/ K2 Q) Z: R6 f
For those unwilling to join the TcUA SIG to view it, I wrote an abbreviated version here:2 @4 w- o" E# m( b' j$ l2 I$ b
Why Performance Matters, @ @: R1 _6 {- I* B
• Productivity – doing more work with less, n6 H* X4 [0 t9 F0 U% M% ?" u% l
o Improve end user productivity/ c, n1 C/ _0 s2 _" U. s5 z9 a k$ S
o Improves administrator productivity: ] |; S8 ]8 v' V0 i7 C8 p
fewer help desk calls responding to or fixing preventable problems) I# e- J5 P! r9 j
• Reliability
+ c1 g* |) I4 a2 s* Y1 T4 S& r7 Lo A well tuned Teamcenter environment also improves system stability and uptime.
; B1 }6 |+ V, }o Less opportunity for data corruption
( I* A. a! ]& C+ Wo Fewer operational errors/problems
' h; ~5 t( W3 A- g3 c' M• Cost Savings# N0 Z; P8 o8 N+ G* ]0 ?. K
o Less waiting means less time wasted.' p: n" J, K' I) C! S; [
Quick Case Study: Company ‘S’- X9 z% m* t) c( K/ a+ z; H
Performance was so bad that something which used to take 10 minutes was taking 2 hours to do.Did pretty much everything wrong, performance-wise, at first but ultimately fixed every major problem. What took 2 hours now only takes 2 minutes.: Z, ~2 H( Q5 f& W
See my March 2012 article on PLM World’s website for details:! @. S$ I I' a5 e; n# E: |0 f" i N+ r
Teamcenter Performance – Hard-Earned Lessons: http://www.plmworld.org/p/bl/ar/blogaid=152
6 v/ {: ?8 v6 v' \Common Performance Bottleneck Causes7 [3 M2 l* Y, b2 p4 X1 O3 d
Using OOTB settings — OOTB settings are intended for development environments; not production environments.% s; Z! N7 B8 c+ `; p
Overloaded servers — Putting too much on a single server.1 i$ R* g1 i" O+ C# s0 K
Operating system settings — Network settings most common (e.g. TCP parameters).- A8 ?0 O# A8 L. E5 e; b; J
Lack of performance monitoring & tuning — Performance doesn’t improve all by itself. It requires human intervention to determine root cause and address it.& o# B$ b* i8 u$ T
Databases – The Most Likely Culprit+ o9 e7 o& \9 Z$ \
• Database performance is highly sensitive to hardware, software, and DB configuration.
5 O" B3 ]' f8 g, | _4 D' h2 f• DBAs at companies generally don’t do performance monitoring and tuning of your Teamcenter DB., i3 e4 E+ x; |. |' S
• It’s typically the last place that Teamcenter administrators look for performance issues rather than the first.
: e6 p0 m; }8 v+ B) G/ cMitigating Database Server Bottlenecks. a! r) C6 S0 }
• Must use dedicated DB server- t' H0 x( s! @( D- E# j0 E( [$ t
o Do not use your DB server for anything other than your Teamcenter production database. This includes not serving additional databases from the same server.
# L3 C5 t: n( c: a+ D% p• Must have fast dedicated storage for DB files! G. q# g. n$ O
o Avoid using a NAS device; especially a shared NAS. Use DAS (Direct Attached Storage) instead.
- D# f6 d z$ N5 s7 f2 \8 qo Use multiple fast disk spindles partitioned with proper RAID levels according to the DB vendor & SPLM recommendations.
$ o9 W7 [+ k# B8 L( So Split DB data, temp, and log files across those RAID partitions according to DB vendor & SPLM recommendations.
/ V$ P- P+ G. Qo Use disk controllers with a battery/flash backed cache.
) s8 q# T4 L$ i6 i* Y [o Use multiple disk controllers if possible.
8 m8 V8 Z7 M$ i& R• Cram the RAM
5 [. [: x5 e$ f1 oo Ideally should be greater than the database size or in-memory footprint. Otherwise disk swapping/paging will occur.
. G; z8 O c$ X/ X. U: N, i• Use 64-bit OS & DB software
) F& S: P' x: |. Y. W2 ^) Po 32-bit software has severe limitations regardless of whether you use “/3G”, AWE, or PAE settings to access memory beyond 3-4G. It’s still a bottleneck.
5 m% B1 z# {8 D; ?/ p0 G& w• Use a good quality network adapter(s)
$ v$ C, ]3 b! }" a. lo Often overlooked as potential source of bottleneck (e.g. packets/sec limitations). Consider multiple NICs “trunked” for better throughput." H" M( \# Q. y4 ?
• DB maintenance tasks
/ Q9 E! s1 O' W6 Wo Update statistics and rebuild indexes regularly.6 j# c, n, M, b0 x. o! A' \9 t9 x
Common Teamcenter Server Bottleneck Causes9 |9 |) t/ v9 v1 r. _- l
• Overloaded Tc Servers
/ k' d' l. V3 N) s+ j• Poor Web Tier Configuration
- y `& ~5 G% E1 X- L: j; c" M• Poor FMS Configuration
5 g! ~- E) R8 }. K+ t% @• Debugging Turned ON
0 b. W, Z7 `2 q9 T& n3 @• Rich Client using OOTB settings
9 r. m: U/ Z5 z. L2 g‘Unload’ Overloaded Teamcenter Servers9 f- Y2 k2 O) j+ W$ F$ ?
• Put Dispatcher modules on separate computers away from other Tc Servers.
9 e" D5 n' ]# N' ~# b• Separate the Web & Enterprise tiers from the Tc Corporate and Volume servers.7 X/ [5 J' l3 K2 A
• Use load balancing for Middle Tier & FMS:
4 F$ D) S+ i% m& j, M' n/ b; J, Y Bo Use multiple Web/Enterprise Tier servers to open up potential hardware resource bottlenecks (e.g. CPU, RAM, network adapter I/O, disk I/O, MB bus, etc.).
9 k- {9 P4 r8 Io Set up multiple FSC cache servers to take load off busy volume servers/storage.4 c: h1 y, B: ?$ `& \3 w9 d: \" f
Web Tier Configuration
6 s" T9 k% a: \) H5 Y) i% `0 x• Do not use port 80 or 8080
2 \3 ~; b5 p7 A/ s5 |; f/ No HTTP traffic on those ports is considered web browsing traffic and is therefore given lowest priority on any network. Can also cause randomly dropped connections.
( K( F5 X" R; R, V$ M5 [) h• Enable http compression if not using WAN acceleration
! A3 {/ r/ ]7 i& |7 Uo Must be done on both web server and on clients to take effect.5 |. K+ E$ g$ C3 h+ Y' Q1 a
• Change/tune OOTB settings for Web Server/App, Y' ]4 L- v+ O5 \0 E
o E.g. timeout values, max # of threads, Java memory, etc.
8 }1 g: E) _0 ?3 ho Read tuning guides specific to your chosen web server/app
& _$ X, }+ i8 H' C1 k+ G X• Scale it up or sideways1 ^% A7 \; z, J1 e5 _) D8 M: v
o Add another web tier server, or increase # of work processes (e.g. Web Garden mode for IIS).
9 u; K; A' [4 { @) MCommon FMS Bottleneck Causes1 z6 `" K1 e& i# v {% G# t. i4 c
• Data improperly routed
4 o! }3 }% M- X) co E.g. Forcing data to go through a remote FSC server over a WAN and back again over the same WAN instead of simply pulling it directly from a nearby Volume/cache server.
3 [* F e2 e7 q• Using OOTB settings
3 Y6 C$ l9 C0 e. v( k9 {5 U, |* To For development purposes only, remember?2 q: T, ]1 s. D- R2 j
• Missing client IP address subnets
: u6 \/ u% L8 X, t' H0 _3 P( T• No load balancing
4 ]6 a- K; G/ k& v/ N# ]! Z• Not using remote cache servers for WAN users9 M$ b* u0 O; u0 N$ R7 W3 V
• Not using remote volume servers for WAN users
2 s# H- b" ?5 Q# AFMS Configuration; W. [% Q8 P, H" p1 L/ e4 z
• Ensure routing is correct
5 P9 N/ ]0 L$ [+ \* Lo Between multiple FSC groups via entry/exit/link parameters- L! ]8 B8 H! {9 h# {
o Between clients and their assigned FSC servers by using complete and accurate ‘clientmap’ parameters for ALL clients.
8 I9 C" s: Q& x4 l2 G3 F6 ?8 A5 Po Use correct transport algorithms for LAN and WAN users.% ]) S9 S2 t! y4 Q1 r5 J% C
• Tune FSC/FCC cache settings. y2 d+ G" `+ i8 N: B7 E; d4 o
o Simply picking a total read/write cache size is not enough.3 V9 x( q% W0 a. R5 u; C% i8 d
o Read “Sizing the FMS fast cache” in the Teamcenter help documentation. There is also a FMS cache sizing tool available from the GTAC web site. Link to FMS Calculator
5 N' d3 j1 F3 k/ m$ Z$ a• Ensure correct client maps& k. f! H% z9 D7 p" T- F" S
o Determine ALL potential client IP address ranges and where they are located. Don’t forget to account for VPN IP addresses.
! v* G' {0 `' ^/ F9 n/ h _! b0 D8 Xo Place client maps in the closest FSC group to their location.
, n: O1 l: I ]* T' }. e0 x• Use multiple FSCs for load balancing
' h7 l- y& v1 l; w% C5 e* Mo Use “front-end” FSC cache servers to reduce load on Volume servers. (Requires disabling of FSC direct routing)8 r9 p6 y4 |0 j8 u/ l, ~) D
o Use redundant FSC cache servers to load balance each other. Also provides fail-over.
2 n- W1 C2 N( Q: k5 D6 X+ ?' W• Place FSC cache servers close to users4 W: ^0 Z: G$ o3 @
o Greatly improves read performance (load time) over a WAN for groups consuming shared data.6 h( J# ?. o& n4 d K6 X
• Place Volume servers close to users/ ?# ~8 o {/ o3 R, w4 k
o Use Store & Forward or at least place remote volumes near remote users. (Backups of remote volumes are critical)
' L7 Z7 E1 N2 q5 bo Greatly improve write performance (save time) over a WAN especially for CAD data.
% m; ^1 l( F& C* z4 _1 s• Prepopulate FSC caches. N O$ \8 N+ Z4 ~
o Run a nightly script to prepopulate shared data across FSC caches.
( w$ _! w- v" zMisc Teamcenter Changes z6 w. F! [9 N# D s: ~! U
• Disable logging and other debugging tools. Turn them on only when needed. Examples,9 c* P+ K' J+ D
o TC_SECURITY_LOGGING=OFF
. X. ?0 Y; u! m' Z( R2 p+ \6 Io TC_APPLICATION_LOGGING=OFF% K! X: u3 s& U& C9 I1 O( S, b
o TC_SLOW_SQL=-1; Z C- t \9 |1 @% Z( w; T
• Rich Client ‘Teamcenter.ini’ configuration:# a1 Y* g6 v% u0 E
o Modify Teamcenter.ini file to increase Java Memory and other settings. Will improve performance/stability when perform large operations (CAD, PSE expansion, large Workflows).
; j2 O) n- c( n* S( k• Enable FCC File Warming
0 K R0 L# O$ i/ o& i" co Improve startup time by having FCC and Tc files pre-loaded at OS login.
3 I2 @6 w: M) YNetwork Performance# w s2 u: u7 l" G3 F; X( T
• Use a hardware-based WAN accelerator if using Teamcenter between multiple facilities over a WAN.& Z9 O& K. q/ ^: `% D2 N
o If none then make best use of software-based Web server and FMS WAN acceleration settings. (i.e. http compression & FMS compression)' s. Y6 C3 q {- `* O: Q
• Optimize OS network settings* U: q" J% w6 _; q Z" T
o OOTB settings are insufficient J2 W5 A l) d; t8 n9 o% w
o Applies to both servers and clients
' {0 u6 ]$ `& a- eo Examples, increasing # of tcp ports, disabling NAGLE algorithm, increasing buffer sizes, trunking multiple NICs, enabling Compound TCP, Large Windows Scaling
9 `7 W2 J, C- r. f6 Z. RRecommended Server Changes (Windows); a7 z$ d( @, q0 |
See Presentation.7 C0 b1 ^+ \0 u- g" x
Recommended Client Changes (Windows)& I# E1 O. |8 V5 @, U
See Presentation./ @4 m( X9 w8 k/ H( O' }
Performance Monitoring Tools
3 T- ^0 f5 r6 E6 K: s# d• Some useful performance monitoring tools:9 G. Y" q; y7 p
• Performance Monitor (comes with Windows)0 @1 n( U% {9 c3 x2 W! w' B
• Used to monitor OS, Teamcenter, and SQL use of resources
3 U3 z/ ]( l& v8 r; T( _% P P• Wireshark (freeware network packet sniffer)
" y! H2 \/ F6 Q$ J9 A1 Q• Used to help identify network TCP performance between computers." O4 ^3 g5 ]" t" s& D$ v
• Available at http://www.wireshark.org/
0 X' w( {0 H& q: O% _- v• Sysinternals Suite (free from Microsoft web site): u/ w' |* Y4 b* @- a( H1 A, ^ S+ [
• http://technet.microsoft.com/en-us/sysinternals/bb842062# f2 X0 }2 k- ^8 M {
Reference Materials Y& j5 L5 }) H5 O, _" D
• Available on GTAC Web Site @ http://support.industrysoftware.automation.siemens.com/docs/teamcenter/- N) j4 P$ a4 g* }+ w7 X8 B
o Teamcenter Deployment Guide
8 I5 D; x1 `1 ?o Teamcenter Network Performance Tuning# x& x" [# I1 @( p$ n
o JBoss 4.2.2 Installation & Tuning Guide
( R/ G9 g, V5 y2 ?5 H: {. ]! `& wo Optimizing Teamcenter Client Performance
( d% a& Z1 g& C& ho Teamcenter System Performance Analysis& c- I1 U+ R& Z# B L
• Oracle documentation & web sites
0 U; c' c4 Q3 f6 L• MS SQL Server documentation in addition to:
/ X: j, Y- F0 f- a7 }o Best Practices for Running Siemens Teamcenter on SQL Server* r) `6 N; O9 w# i5 N* Z
http://download.microsoft.com/download/7/3/6/7365D2BB-BB34-4D28-A128-F2C8FBA6E995/Siemens-Teamcenter-and-SQL-Server-Best-Practices.pdf8 }, G( k: j+ A( C0 ^) C
o Siemens Blog on Technet
; s8 z( j: |7 D1 s1 N http://blogs.technet.com/b/sql_server_isv/archive/tags/siemens/
0 y' l7 p3 H( }+ |3 ?/ G+ eo Siemens-Teamcenter-SQL-Resource-Page
& S4 P. d; ?: ^9 J$ |' ^ http://blogs.technet.com/b/sql_server_isv/p/siemens-teamcenter-sql-server-resource-page.aspx( Q4 Q8 _+ u' a. r
• Past PLM World Conference Presentations available at www.plmworld.org,
. Z3 I+ k% f& Wo Teamcenter 4-Tier Deployment – Best Practices4 T0 u) @5 L, ^& k; x7 ^' d4 F
o Teamcenter – Database Performance
( B' p% H8 Z* G( To JBoss Optimization for Teamcenter
; V* q6 p1 ?1 p0 [3 t9 e0 Yo Optimizing Teamcenter Client Performance
# H; _- u+ K A8 k% n7 \o Teamcenter Database Server Configuration & Tuning2 A, n! J* r8 B' a/ U% m/ G) L
Contact Information! t. p R/ q) J* A
Larry Carpenter P.E., PLM World Teamcenter SIG Chair, tcua@plmworld.org. I! d8 s6 A) y3 f
Teamcenter UA SIG: http://www.plmworld.org/TC_UA! l/ B9 b9 D% D& C3 s% v
LinkedIn: http://www.linkedin.com/pub/larry-carpenter-pe/44/5b8/aaa
" X2 ^- U* T3 K5 s dAlternate Contact Info: ideas2nx@plmworld.org, larry.carpenter@siemens.com
% T8 G5 R& e& W5 o# x3 g6 [! Q4 j- P# [# H. Z
|
|