1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
|
<?xml version="1.0" encoding="ISO-8859-1"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"><head><!--
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
This file is generated from xml source: DO NOT EDIT
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
-->
<title>Performance Scaling
- Apache HTTP Server</title>
<link href="../style/css/manual.css" rel="stylesheet" media="all" type="text/css" title="Main stylesheet" />
<link href="../style/css/manual-loose-100pc.css" rel="alternate stylesheet" media="all" type="text/css" title="No Sidebar - Default font size" />
<link href="../style/css/manual-print.css" rel="stylesheet" media="print" type="text/css" /><link rel="stylesheet" type="text/css" href="../style/css/prettify.css" />
<script src="../style/scripts/prettify.js" type="text/javascript">
</script>
<link href="../images/favicon.ico" rel="shortcut icon" /></head>
<body id="manual-page"><div id="page-header">
<p class="menu"><a href="../mod/">Modules</a> | <a href="../mod/quickreference.html">Directives</a> | <a href="http://wiki.apache.org/httpd/FAQ">FAQ</a> | <a href="../glossary.html">Glossary</a> | <a href="../sitemap.html">Sitemap</a></p>
<p class="apache">Apache HTTP Server Version 2.5</p>
<img alt="" src="../images/feather.gif" /></div>
<div class="up"><a href="./"><img title="<-" alt="<-" src="../images/left.gif" /></a></div>
<div id="path">
<a href="http://www.apache.org/">Apache</a> > <a href="http://httpd.apache.org/">HTTP Server</a> > <a href="http://httpd.apache.org/docs/">Documentation</a> > <a href="../">Version 2.5</a></div><div id="page-content"><div id="preamble"><h1>Performance Scaling
</h1>
<div class="toplang">
<p><span>Available Languages: </span><a href="../en/misc/perf-scaling.html" title="English"> en </a></p>
</div>
<p>The Performance Tuning page in the Apache 1.3 documentation says:
</p>
<blockquote><p>
“Apache is a general webserver, which is designed to be
correct first, and fast
second. Even so, its performance is quite satisfactory. Most
sites have less than 10Mbits of outgoing bandwidth, which
Apache can fill using only a low end Pentium-based
webserver.”</p>
</blockquote>
<p>However, this sentence was written a few years ago, and in the
meantime several things have happened. On one hand, web server
hardware has become much faster. On the other hand, many sites now
are allowed much more than ten megabits per second of outgoing
bandwidth. In addition, web applications have become more complex.
The classic brochureware site is alive and well, but the web has
grown up substantially as a computing application platform and
webmasters may find themselves running dynamic content in Perl, PHP
or Java, all of which take a toll on performance.
</p>
<p>Therefore, in spite of strides forward in machine speed and
bandwidth allowances, web server performance and web application
performance remain areas of concern. In this documentation several
aspects of web server performance will be discussed.
</p>
</div>
<div id="quickview"><ul id="toc"><li><img alt="" src="../images/down.gif" /> <a href="#what-will-and-will-not-be-discussed">What Will and Will Not Be Discussed
</a></li>
<li><img alt="" src="../images/down.gif" /> <a href="#monitoring-your-server">Monitoring Your Server
</a></li>
<li><img alt="" src="../images/down.gif" /> <a href="#configuring-for-performance">Configuring for Performance
</a></li>
<li><img alt="" src="../images/down.gif" /> <a href="#caching-content">Caching Content
</a></li>
<li><img alt="" src="../images/down.gif" /> <a href="#further-considerations">Further Considerations
</a></li>
</ul><ul class="seealso"><li><a href="#comments_section">Comments</a></li></ul></div>
<div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
<div class="section">
<h2><a name="what-will-and-will-not-be-discussed" id="what-will-and-will-not-be-discussed">What Will and Will Not Be Discussed
</a></h2>
<p>The session will focus on easily accessible configuration and tuning
options for Apache httpd 2.2 and 2.4 as well as monitoring tools.
Monitoring tools will allow you to observe your web server to
gather information about its performance, or lack thereof.
We'll assume that you don't have an unlimited budget for
server hardware, so the existing infrastructure will have to do the
job. You have no desire to compile your own Apache, or to recompile
the operating system kernel. We do assume, though, that you have
some familiarity with the Apache httpd configuration file.
</p>
</div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
<div class="section">
<h2><a name="monitoring-your-server" id="monitoring-your-server">Monitoring Your Server
</a></h2>
<p>The first task when sizing or performance-tuning your server is to
find out how your system is currently performing. By monitoring
your server under real-world load, or artificially generated load,
you can extrapolate its behavior under stress, such as when your
site is mentioned on Slashdot.
</p>
<h3><a name="monitoring-tools" id="monitoring-tools">Monitoring Tools
</a></h3>
<h4><a name="top" id="top">top
</a></h4>
<p>The top tool ships with Linux and FreeBSD. Solaris offers
<code>prstat(1)</code>. It collects a number of statistics for the
system and for each running process, then displays them
interactively on your terminal. The data displayed is
refreshed every second and varies by platform, but
typically includes system load average, number of processes
and their current states, the percent CPU(s) time spent
executing user and system code, and the state of the
virtual memory system. The data displayed for each process
is typically configurable and includes its process name and
ID, priority and nice values, memory footprint, and
percentage CPU usage. The following example shows multiple
httpd processes (with MPM worker and event) running on an
Linux (Xen) system:
</p>
<div class="example"><pre>
top - 23:10:58 up 71 days, 6:14, 4 users, load average: 0.25, 0.53, 0.47
Tasks: 163 total, 1 running, 162 sleeping, 0 stopped, 0 zombie
Cpu(s): 11.6%us, 0.7%sy, 0.0%ni, 87.3%id, 0.4%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 2621656k total, 2178684k used, 442972k free, 100500k buffers
Swap: 4194296k total, 860584k used, 3333712k free, 1157552k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
16687 example_ 20 0 1200m 547m 179m S 45 21.4 1:09.59 httpd-worker
15195 www 20 0 441m 33m 2468 S 0 1.3 0:41.41 httpd-worker
1 root 20 0 10312 328 308 S 0 0.0 0:33.17 init
2 root 15 -5 0 0 0 S 0 0.0 0:00.00 kthreadd
3 root RT -5 0 0 0 S 0 0.0 0:00.14 migration/0
4 root 15 -5 0 0 0 S 0 0.0 0:04.58 ksoftirqd/0
5 root RT -5 0 0 0 S 0 0.0 4:45.89 watchdog/0
6 root 15 -5 0 0 0 S 0 0.0 1:42.52 events/0
7 root 15 -5 0 0 0 S 0 0.0 0:00.00 khelper
19 root 15 -5 0 0 0 S 0 0.0 0:00.00 xenwatch
20 root 15 -5 0 0 0 S 0 0.0 0:00.00 xenbus
28 root RT -5 0 0 0 S 0 0.0 0:00.14 migration/1
29 root 15 -5 0 0 0 S 0 0.0 0:00.20 ksoftirqd/1
30 root RT -5 0 0 0 S 0 0.0 0:05.96 watchdog/1
31 root 15 -5 0 0 0 S 0 0.0 1:18.35 events/1
32 root RT -5 0 0 0 S 0 0.0 0:00.08 migration/2
33 root 15 -5 0 0 0 S 0 0.0 0:00.18 ksoftirqd/2
34 root RT -5 0 0 0 S 0 0.0 0:06.00 watchdog/2
35 root 15 -5 0 0 0 S 0 0.0 1:08.39 events/2
36 root RT -5 0 0 0 S 0 0.0 0:00.10 migration/3
37 root 15 -5 0 0 0 S 0 0.0 0:00.16 ksoftirqd/3
38 root RT -5 0 0 0 S 0 0.0 0:06.08 watchdog/3
39 root 15 -5 0 0 0 S 0 0.0 1:22.81 events/3
68 root 15 -5 0 0 0 S 0 0.0 0:06.28 kblockd/0
69 root 15 -5 0 0 0 S 0 0.0 0:00.04 kblockd/1
70 root 15 -5 0 0 0 S 0 0.0 0:00.04 kblockd/2</pre></div>
<p>Top is a wonderful tool even though it’s slightly resource
intensive (when running, its own process is usually in the
top ten CPU gluttons). It is indispensable in determining
the size of a running process, which comes in handy when
determining how many server processes you can run on your
machine. How to do this is described in <a href="#sizing-maxClients">sizing MaxClients</a>.
Top is, however, an interactive tool and running it
continuously has few if any advantages.
</p>
<h4><a name="free" id="free">free
</a></h4>
<p>This command is only available on Linux. It shows how much
memory and swap space is in use. Linux allocates unused
memory as file system cache. The free command shows usage
both with and without this cache. The free command can be
used to find out how much memory the operating system is
using, as described in the paragraph <a href="#sizing-maxClients">sizing MaxClients</a>.
The output of free looks like this:
</p>
<div class="example"><pre>
sctemme@brutus:~$ free
total used free shared buffers cached
Mem: 4026028 3901892 124136 0 253144 841044
-/+ buffers/cache: 2807704 1218324
Swap: 3903784 12540 3891244
</pre></div>
<h4><a name="vmstat" id="vmstat">vmstat
</a></h4>
<p>This command is available on many unix platforms. It
displays a large number of operating system metrics. Run
without argument, it displays a status line for that
moment. When a numeric argument is added, the status is
redisplayed at designated intervals. For example,
<code>vmstat 5</code>
causes the information to reappear every five seconds.
Vmstat displays the amount of virtual memory in use, how
much memory is swapped in and out each second, the number
of processes currently running and sleeping, the number of
interrupts and context switches per second and the usage
percentages of the CPU.
</p>
<p>
The following is <code>vmstat</code> output of an idle server:
</p>
<div class="example"><pre>
[sctemme@GayDeceiver sctemme]$ vmstat 5 3
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
0 0 0 0 186252 6688 37516 0 0 12 5 47 311 0 1 99
0 0 0 0 186244 6696 37516 0 0 0 16 41 314 0 0 100
0 0 0 0 186236 6704 37516 0 0 0 9 44 314 0 0 100
</pre></div>
<p>And this is output of a server that is under a load of one
hundred simultaneous connections fetching static content:
</p>
<div class="example"><pre>
[sctemme@GayDeceiver sctemme]$ vmstat 5 3
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
1 0 1 0 162580 6848 40056 0 0 11 5 150 324 1 1 98
6 0 1 0 163280 6856 40248 0 0 0 66 6384 1117 42 25 32
11 0 0 0 162780 6864 40436 0 0 0 61 6309 1165 33 28 40
</pre></div>
<p>The first line gives averages since the last reboot. The
subsequent lines give information for five second
intervals. The second argument tells vmstat to generate
three reports and then exit.
</p>
<h4><a name="se-toolkit" id="se-toolkit">SE Toolkit
</a></h4>
<p>The SE Toolkit is a system monitoring toolkit for Solaris.
Its programming language is based on the C preprocessor and
comes with a number of sample scripts. It can use both the
command line and the GUI to display information. It can
also be programmed to apply rules to the system data. The
example script shown in Figure 2, Zoom.se, shows green,
orange or red indicators when utilization of various parts
of the system rises above certain thresholds. Another
included script, Virtual Adrian, applies performance tuning
metrics according to.
</p>
<p>The SE Toolkit has drifted around for a while and has had
several owners since its inception. It seems that it has
now found a final home at Sunfreeware.com, where it can be
downloaded at no charge. There is a single package for
Solaris 8, 9 and 10 on SPARC and x86, and includes source
code. SE Toolkit author Richard Pettit has started a new
company, Captive Metrics4 that plans to bring to market a
multiplatform monitoring tool built on the same principles
as SE Toolkit, written in Java.
</p>
<h4><a name="dtrace" id="dtrace">DTrace
</a></h4>
<p>Given that DTrace is available for Solaris, FreeBSD and OS
X, it might be worth exploring it. There's also
mod_dtrace available for httpd.
</p>
<h4><a name="mod_status" id="mod_status">mod_status
</a></h4>
<p>The mod_status module gives an overview of the server
performance at a given moment. It generates an HTML page
with, among others, the number of Apache processes running
and how many bytes each has served, and the CPU load caused
by httpd and the rest of the system. The Apache Software
Foundation uses <code class="module"><a href="../mod/mod_status.html">mod_status</a></code> on its own
<a href="http://apache.org/server-status">web site</a>.
If you put the <code>ExtendedStatus On</code>
directive in your <code>httpd.conf</code>,
the <code class="module"><a href="../mod/mod_status.html">mod_status</a></code>
page will give you more information at the cost of a little
extra work per request.
</p>
<h3><a name="web-server-log-files" id="web-server-log-files">Web Server Log Files
</a></h3>
<p>Monitoring and analyzing the log files httpd writes is one of
the most effective ways to keep track of your server health and
performance. Monitoring the error log allows you to detect
error conditions, discover attacks and find performance issues.
Analyzing the access logs tells you how busy your server is,
which resources are the most popular and where your users come
from. Historical log file data can give you invaluable insight
into trends in access to your server, which allows you to
predict when your performance needs will overtake your server
capacity.
</p>
<h4><a name="ErrorLog" id="ErrorLog">Error Log
</a></h4>
<p>The error log will contain messages if the server has
reached the maximum number of active processes or the
maximum number of concurrently open files. The error log
also reflects when processes are being spawned at a
higher-than-usual rate in response to a sudden increase in
load. When the server starts, the stderr file descriptor is
redirected to the error logfile, so any error encountered
by httpd after it opens its logfiles will appear in this
log. This makes it good practice to review the error log
frequently.
</p>
<p>Before Apache httpd opens its logfiles, any errors will be
written to the stderr stream. If you start httpd manually,
this error information will appear on your terminal and you
can use it directly to troubleshoot your server. If your
httpd is started by a startup script, the destination of
early error messages depends on their design. The
<code>/var/log/messages</code>
file is usually a good bet. On Windows, early error
messages are written to the Applications Event Log, which
can be viewed through the Event Viewer in Administrative
Tools.
</p>
<p>
The Error Log is configured through the <code class="directive"><a href="../mod/core.html#errorlog">ErrorLog</a></code>
and <code class="directive"><a href="../mod/core.html#loglevel">LogLevel</a></code>
configuration directives. The error log of httpd’s main
server configuration receives the log messages that pertain
to the entire server: startup, shutdown, crashes, excessive
process spawns, etc. The <code class="directive"><a href="../mod/core.html#errorlog">ErrorLog</a></code>
directive can also be used in virtual host containers. The
error log of a virtual host receives only log messages
specific to that virtual host, such as authentication
failures and 'File not Found' errors.
</p>
<p>On a server that is visible to the Internet, expect to see a
lot of exploit attempt and worm attacks in the error log. A
lot of these will be targeted at other server platforms
instead of Apache, but the current state of affairs is that
attack scripts just throw everything they have at any open
port, regardless of which server is actually running or
what applications might be installed. You could block these
attempts using a firewall or <a href="http://www.modsecurity.org/">mod_security</a>,
but this falls outside the scope of this discussion.
</p>
<p>
The <code class="directive"><a href="../mod/core.html#loglevel">LogLevel</a></code>
directive determines the level of detail included in the
logs. There are eight log levels as described here:
</p>
<table>
<tr>
<td>
<p><strong>Level</strong></p>
</td>
<td>
<p><strong>Description</strong></p>
</td>
</tr>
<tr>
<td>
<p>emerg</p>
</td>
<td>
<p>Emergencies - system is unusable.</p>
</td>
</tr>
<tr>
<td>
<p>alert</p>
</td>
<td>
<p>Action must be taken immediately.</p>
</td>
</tr>
<tr>
<td>
<p>crit</p>
</td>
<td>
<p>Critical Conditions.</p>
</td>
</tr>
<tr>
<td>
<p>error</p>
</td>
<td>
<p>Error conditions.</p>
</td>
</tr>
<tr>
<td>
<p>warn</p>
</td>
<td>
<p>Warning conditions.</p>
</td>
</tr>
<tr>
<td>
<p>notice</p>
</td>
<td>
<p>Normal but significant condition.</p>
</td>
</tr>
<tr>
<td>
<p>info</p>
</td>
<td>
<p>Informational.</p>
</td>
</tr>
<tr>
<td>
<p>debug</p>
</td>
<td>
<p>Debug-level messages</p>
</td>
</tr>
</table>
<p>The default log level is warn. A production server should
not be run on debug, but increasing the level of detail in
the error log can be useful during troubleshooting.
Starting with 2.3.8 <code class="directive"><a href="../mod/core.html#loglevel">LogLevel</a></code>
can be specified on a per module basis:
</p>
<pre class="prettyprint lang-config">
LogLevel debug mod_ssl:warn
</pre>
<p>
This puts all of the server in debug mode, except for
<code class="module"><a href="../mod/mod_ssl.html">mod_ssl</a></code>, which tends to be very noisy.
</p>
<h4><a name="AccessLog" id="AccessLog">Access Log
</a></h4>
<p>Apache httpd keeps track of every request it services in its
access log file. In addition to the time and nature of a
request, httpd can log the client IP address, date and time
of the request, the result and a host of other information.
The various logging format features are documented in the
manual. This file exists by default for the main server and can be
configured per virtual host by using the <code class="directive"><a href="../mod/mod_log_config.html#transferlog">TransferLog</a></code>
or <code class="directive"><a href="../mod/mod_log_config.html#customlog">CustomLog</a></code>
configuration directive.
</p>
<p>The access logs can be analyzed with any of several free and
commercially available programs. Popular free analysis
packages include Analog and Webalizer. Log analysis should
be done offline so the web server machine is not burdened
by processing the log files. Most log analysis packages
understand the Common Log Format. The fields in the log
lines are explained in in the following:
</p>
<div class="example"><pre>
195.54.228.42 - - [24/Mar/2007:23:05:11 -0400] "GET /sander/feed/ HTTP/1.1" 200 9747
64.34.165.214 - - [24/Mar/2007:23:10:11 -0400] "GET /sander/feed/atom HTTP/1.1" 200 9068
60.28.164.72 - - [24/Mar/2007:23:11:41 -0400] "GET / HTTP/1.0" 200 618
85.140.155.56 - - [24/Mar/2007:23:14:12 -0400] "GET /sander/2006/09/27/44/ HTTP/1.1" 200 14172
85.140.155.56 - - [24/Mar/2007:23:14:15 -0400] "GET /sander/2006/09/21/gore-tax-pollution/ HTTP/1.1" 200 15147
74.6.72.187 - - [24/Mar/2007:23:18:11 -0400] "GET /sander/2006/09/27/44/ HTTP/1.0" 200 14172
74.6.72.229 - - [24/Mar/2007:23:24:22 -0400] "GET /sander/2006/11/21/os-java/ HTTP/1.0" 200 13457
</pre></div>
<table>
<tr>
<td>
<p><strong>Field</strong></p>
</td>
<td>
<p><strong>Content</strong></p>
</td>
<td>
<p><strong>Explanation</strong></p>
</td>
</tr>
<tr>
<td>
<p>Client IP</p>
</td>
<td>
<p>195.54.228.42</p>
</td>
<td>
<p>IP address where the request originated</p>
</td>
</tr>
<tr>
<td>
<p>RFC 1413 ident</p>
</td>
<td>
<p>-</p>
</td>
<td>
<p>Remote user identity as reported by their identd</p>
</td>
</tr>
<tr>
<td>
<p>username</p>
</td>
<td>
<p>-</p>
</td>
<td>
<p>Remote username as authenticated by Apache</p>
</td>
</tr>
<tr>
<td>
<p>timestamp</p>
</td>
<td>
<p>[24/Mar/2007:23:05:11 -0400]</p>
</td>
<td>
<p>Date and time of request</p>
</td>
</tr>
<tr>
<td>
<p>Request</p>
</td>
<td>
<p>"GET /sander/feed/ HTTP/1.1"</p>
</td>
<td>
<p>Request line</p>
</td>
</tr>
<tr>
<td>
<p>Status Code</p>
</td>
<td>
<p>200</p>
</td>
<td>
<p>Response code</p>
</td>
</tr>
<tr>
<td>
<p>Content Bytes</p>
</td>
<td>
<p>9747</p>
</td>
<td>
<p>Bytes transferred w/o headers</p>
</td>
</tr>
</table>
<h4><a name="rotating-log-files" id="rotating-log-files">Rotating Log Files
</a></h4>
<p>There are several reasons to rotate logfiles. Even though
almost no operating systems out there have a hard file size
limit of two Gigabytes anymore, log files simply become too
large to handle over time. Additionally, any periodic log
file analysis should not be performed on files to which the
server is actively writing. Periodic logfile rotation helps
keep the analysis job manageable, and allows you to keep a
closer eye on usage trends.
</p>
<p>On unix systems, you can simply rotate logfiles by giving
the old file a new name using mv. The server will keep
writing to the open file even though it has a new name.
When you send a graceful restart signal to the server, it
will open a new logfile with the configured name. For
example, you could run a script from cron like this:
</p>
<div class="example"><p><code>
APACHE=/usr/local/apache2<br />
HTTPD=$APACHE/bin/httpd<br />
mv $APACHE/logs/access_log
$APACHE/logarchive/access_log-‘date +%F‘<br />
$HTTPD -k graceful
</code></p></div>
<p>This approach also works on Windows, just not as smoothly.
While the httpd process on your Windows server will keep
writing to the log file after it has been renamed, the
Windows Service that runs Apache can not do a graceful
restart. Restarting a Service on Windows means stopping it
and then starting it again. The advantage of a graceful
restart is that the httpd child processes get to complete
responding to their current requests before they exit.
Meanwhile, the httpd server becomes immediately available
again to serve new requests. The stop-start that the
Windows Service has to perform will interrupt any requests
currently in progress, and the server is unavailable until
it is started again. Plan for this when you decide the
timing of your restarts.
</p>
<p>
A second approach is to use piped logs. From the
<code class="directive"><a href="../mod/mod_log_config.html#customlog">CustomLog</a></code>,
<code class="directive"><a href="../mod/mod_log_config.html#transferlog">TransferLog</a></code>
or <code class="directive"><a href="../mod/core.html#errorlog ">ErrorLog
</a></code>
directives you can send the log data into any program using
a pipe character (<code>|</code>). For instance:
</p>
<div class="example"><p><code>CustomLog "|/usr/local/apache2/bin/rotatelogs
/var/log/access_log 86400" common
</code></p></div>
<p>The program on the other end of the pipe will receive the
Apache log data on its stdin stream, and can do with this
data whatever it wants. The rotatelogs program that comes
with Apache seamlessly turns over the log file based on
time elapsed or the amount of data written, and leaves the
old log files with a timestamp suffix to its name. This
method for rotating logfiles works well on unix platforms,
but is currently broken on Windows.
</p>
<h4><a name="logging-and-performance" id="logging-and-performance">Logging and Performance
</a></h4>
<p>Writing entries to the Apache log files obviously takes some
effort, but the information gathered from the logs is so
valuable that under normal circumstances logging should not
be turned off. For optimal performance, you should put your
disk-based site content on a different physical disk than
the server log files: the access patterns are very
different. Retrieving content from disk is a read operation
in a fairly random pattern, and log files are written to
disk sequentially.
</p>
<p>
Do not run a production server with your error <code class="directive"><a href="../mod/core.html#loglevel">LogLevel</a></code>
set to debug. This log level causes a vast amount of
information to be written to the error log, including, in
the case of SSL access, complete dumps of BIO read and
write operations. The performance implications are
significant: use the default warn level instead.
</p>
<p>If your server has more than one virtual host, you may give
each virtual host a separate access logfile. This makes it
easier to analyze the logfile later. However, if your
server has many virtual hosts, all the open logfiles put a
resource burden on your system, and it may be preferable to
log to a single file. Use the <code>%v</code>
format character at the start of your <code class="directive"><a href="../mod/mod_log_config.html#logformat">LogFormat</a></code>
and starting 2.3.8 of your <code class="directive"><a href="../mod/core.html#errorlog">ErrorLog</a></code>
to make httpd print the hostname of the virtual host that
received the request or the error at the beginning of each
log line. A simple Perl script can split out the log file
after it rotates: one is included with the Apache source
under <code>support/split-logfile</code>.
</p>
<p>
You can use the <code class="directive"><a href="../mod/mod_log_config.html#bufferedlogs">BufferedLogs</a></code>
directive to have Apache collect several log lines in
memory before writing them to disk. This might yield better
performance, but could affect the order in which the
server's log is written.
</p>
<h3><a name="generating-a-test-load" id="generating-a-test-load">Generating A Test Load
</a></h3>
<p>It is useful to generate a test load to monitor system
performance under realistic operating circumstances. Besides
commercial packages such as <a href="http://learnloadrunner.com/">LoadRunner</a>
,there are a number of freely available tools to generate a
test load against your web server.
</p>
<ul>
<li>Apache ships with a test program called ab, short for
Apache Bench. It can generate a web server load by
repeatedly asking for the same file in rapid succession.
You can specify a number of concurrent connections and have
the program run for either a given amount of time or a
specified number of requests.
</li>
<li>Another freely available load generator is http load11 .
This program works with a URL file and can be compiled with
SSL support.
</li>
<li>The Apache Software Foundation offers a tool named flood12
. Flood is a fairly sophisticated program that is
configured through an XML file.
</li>
<li>Finally, JMeter13 , a Jakarta subproject, is an all-Java
load-testing tool. While early versions of this application
were slow and difficult to use, the current version 2.1.1
seems to be versatile and useful.
</li>
<li>
<p>ASF external projects, that have proven to be quite
good: grinder, httperf, tsung, <a href="http://funkload.nuxeo.org/">FunkLoad</a>
</p>
</li>
</ul>
<p>When you load-test your web server, please keep in mind that if
that server is in production, the test load may negatively
affect the server’s response. Also, any data traffic you
generate may be charged against your monthly traffic allowance.
</p>
</div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
<div class="section">
<h2><a name="configuring-for-performance" id="configuring-for-performance">Configuring for Performance
</a></h2>
<h3><a name="apache-configuration" id="apache-configuration">Httpd Configuration
</a></h3>
<p>The Apache 2.2 httpd is by default a pre-forking web server.
When the server starts, the parent process spawns a number of
child processes that do the actual work of servicing requests.
But Apache httpd 2.0 introduced the concept of the
Multi-Processing Module (MPM). Developers can write MPMs to
suit the process- or threadingarchitecture of their specific
operating system. Apache 2 comes with special MPMs for Windows,
OS/2, Netware and BeOS. On unix-like platforms, the two most
popular MPMs are Prefork and Worker. The Prefork MPM offers the
same pre-forking process model that Apache 1.3 uses. The Worker
MPM runs a smaller number of child processes, and spawns
multiple request handling threads within each child process. In
2.4 MPMs are no longer hard-wired. They too can be exchanged
via <code class="directive"><a href="../mod/core.html#loadmodule">LoadModule</a></code>.
The default MPM in 2.4 is the event MPM.
</p>
<p>The maximum number of workers, be they pre-forked child
processes or threads within a process, is an indication of how
many requests your server can manage concurrently. It is merely
a rough estimate because the kernel can queue connection
attempts for your web server. When your site becomes busy and
the maximum number of workers is running, the machine
doesn't hit a hard limit beyond which clients will be
denied access. However, once requests start backing up, system
performance is likely to degrade.
</p>
<p>Finally, if the httpd server in question is not executing any third-party
code, via <code>mod_php</code>, <code>mod_perl</code> or similar,
we recommend the use of <code class="module"><a href="../mod/mpm_event.html">mpm_event</a></code>. This MPM is ideal
for situations where httpd serves as a thin layer between clients and
backend servers doing the real job, such as a proxy or cache.
</p>
<h4><a name="MaxClients" id="MaxClients">MaxClients
</a></h4>
<p>
The <code>MaxClients
</code>
directive in your Apache httpd configuration file specifies
the maximum number of workers your server can create. It
has two related directives, <code>MinSpareServers
</code>
and <code>MaxSpareServers
</code>
,which specify the number of workers Apache keeps waiting
in the wings ready to serve requests. The absolute maximum
number of processes is configurable through the <code>
ServerLimit
</code>
directive.
</p>
<h4><a name="spinning-threads" id="spinning-threads">Spinning Threads
</a></h4>
<p>For the prefork MPM of the above directives are all there is
to determining the process limit. However, if you are
running a threaded MPM the situation is a little more
complicated. Threaded MPMs support the <code>
ThreadsPerChild
</code>
directive1 . Apache requires that <code>MaxClients
</code>
is evenly divisible by <code>ThreadsPerChild
</code>
.If you set either directive to a number that doesn’t
meet this requirement, Apache will send a message of
complaint to the error log and adjust the <code>
ThreadsPerChild
</code>
value downwards until it is an even factor of <code>
MaxClients
</code>
.
</p>
<h4><a name="sizing-maxClients" id="sizing-maxClients">Sizing MaxClients
</a></h4>
<p>Optimally, the maximum number of processes should be set so
that all the memory on your system is used, but no more. If
your system gets so overloaded that it needs to heavily
swap core memory out to disk, performance will degrade
quickly. The formula for determining <code class="directive"><a href="../mod/mpm_common.html#maxclients">MaxClients</a></code>
is fairly simple:
</p>
<div class="example"><p><code>
total RAM − RAM for OS − RAM for external programs<br />
MaxClients =
-------------------------------------------------------<br />
RAM per httpd process
</code></p></div>
<p>The various amounts of memory allocated for the OS, external
programs and the httpd processes is best determined by
observation: use the top and free commands described above
to determine the memory footprint of the OS without the web
server running. You can also determine the footprint of a
typical web server process from top: most top
implementations have a Resident Size (RSS) column and a
Shared Memory column.
</p>
<p>The difference between these two is the amount of memory
per-process. The shared segment really exists only once and
is used for the code and libraries loaded and the dynamic
inter-process tally, or 'scoreboard,' that Apache
keeps. How much memory each process takes for itself
depends heavily on the number and kind of modules you use.
The best approach to use in determining this need is to
generate a typical test load against your web site and see
how large the httpd processes become.
</p>
<p>The RAM for external programs parameter is intended mostly
for CGI programs and scripts that run outside the web
server process. However, if you have a Java virtual machine
running Tomcat on the same box it will need a significant
amount of memory as well. The above assessment should give
you an idea how far you can push <code>MaxClients
</code>
,but it is not an exact science. When in doubt, be
conservative and use a low <code>MaxClients
</code>
value. The Linux kernel will put extra memory to good use
for caching disk access. On Solaris you need enough
available real RAM memory to create any process. If no real
memory is available, httpd will start writing ‘No space
left on device’ messages to the error log and be unable
to fork additional child processes, so a higher <code>
MaxClients
</code>
value may actually be a disadvantage.
</p>
<h4><a name="selecting-your-mpm" id="selecting-your-mpm">Selecting your MPM
</a></h4>
<p>The prime reason for selecting a threaded MPM is that
threads consume fewer system resources than processes, and
it takes less effort for the system to switch between
threads. This is more true for some operating systems than
for others. On systems like Solaris and AIX, manipulating
processes is relatively expensive in terms of system
resources. On these systems, running a threaded MPM makes
sense. On Linux, the threading implementation actually uses
one process for each thread. Linux processes are relatively
lightweight, but it means that a threaded MPM offers less
of a performance advantage than in other environments.
</p>
<p>Running a threaded MPM can cause stability problems in some
situations For instance, should a child process of a
preforked MPM crash, at most one client connection is
affected. However, if a threaded child crashes, all the
threads in that process disappear, which means all the
clients currently being served by that process will see
their connection aborted. Additionally, there may be
so-called "thread-safety" issues, especially with
third-party libraries. In threaded applications, threads
may access the same variables indiscriminately, not knowing
whether a variable may have been changed by another thread.
</p>
<p>This has been a sore point within the PHP community. The PHP
processor heavily relies on third-party libraries and
cannot guarantee that all of these are thread-safe. The
good news is that if you are running Apache on Linux, you
can run PHP in the preforked MPM without fear of losing too
much performance relative to the threaded option.
</p>
<h4><a name="spinning-locks" id="spinning-locks">Spinning Locks
</a></h4>
<p>Apache httpd maintains an inter-process lock around its
network listener. For all practical purposes, this means
that only one httpd child process can receive a request at
any given time. The other processes are either servicing
requests already received or are 'camping out' on
the lock, waiting for the network listener to become
available. This process is best visualized as a revolving
door, with only one process allowed in the door at any
time. On a heavily loaded web server with requests arriving
constantly, the door spins quickly and requests are
accepted at a steady rate. On a lightly loaded web server,
the process that currently "holds" the lock may
have to stay in the door for a while, during which all the
other processes sit idle, waiting to acquire the lock. At
this time, the parent process may decide to terminate some
children based on its <code>MaxSpareServers
</code>
directive.
</p>
<h4><a name="the-thundering-herd" id="the-thundering-herd">The Thundering Herd
</a></h4>
<p>The function of the 'accept mutex' (as this
inter-process lock is called) is to keep request reception
moving along in an orderly fashion. If the lock is absent,
the server may exhibit the Thundering Herd syndrome.
</p>
<p>Consider an American Football team poised on the line of
scrimmage. If the football players were Apache processes
all team members would go for the ball simultaneously at
the snap. One process would get it, and all the others
would have to lumber back to the line for the next snap. In
this metaphor, the accept mutex acts as the quarterback,
delivering the connection "ball" to the
appropriate player process.
</p>
<p>Moving this much information around is obviously a lot of
work, and, like a smart person, a smart web server tries to
avoid it whenever possible. Hence the revolving door
construction. In recent years, many operating systems,
including Linux and Solaris, have put code in place to
prevent the Thundering Herd syndrome. Apache recognizes
this and if you run with just one network listener, meaning
one virtual host or just the main server, Apache will
refrain from using an accept mutex. If you run with
multiple listeners (for instance because you have a virtual
host serving SSL requests), it will activate the accept
mutex to avoid internal conflicts.
</p>
<p>
You can manipulate the accept mutex with the <code>
AcceptMutex
</code>
directive. Besides turning the accept mutex off, you can
select the locking mechanism. Common locking mechanisms
include fcntl, System V Semaphores and pthread locking. Not
all are available on every platform, and their availability
also depends on compile-time settings. The various locking
mechanisms may place specific demands on system resources:
manipulate them with care.
</p>
<p>There is no compelling reason to disable the accept mutex.
Apache automatically recognizes the single listener
situation described above and knows if it is safe to run
without mutex on your platform.
</p>
<h3><a name="tuning-the-operating-system" id="tuning-the-operating-system">Tuning the Operating System
</a></h3>
<p>People often look for the 'magic tune-up' that will
make their system perform four times as fast by tweaking just
one little setting. The truth is, present-day UNIX derivatives
are pretty well adjusted straight out of the box and there is
not a lot that needs to be done to make them perform optimally.
However, there are a few things that an administrator can do to
improve performance.
</p>
<h4><a name="ram-and-swap-space" id="ram-and-swap-space">RAM and Swap Space
</a></h4>
<p>The usual mantra regarding RAM is "more is
better". As discussed above, unused RAM is put to good
use as file system cache. The Apache processes get bigger
if you load more modules, especially if you use modules
that generate dynamic page content within the processes,
like PHP and mod_perl. A large configuration file-with many
virtual hosts-also tends to inflate the process footprint.
Having ample RAM allows you to run Apache with more child
processes, which allows the server to process more
concurrent requests.
</p>
<p>While the various platforms treat their virtual memory in
different ways, it is never a good idea to run with less
disk-based swap space than RAM. The virtual memory system
is designed to provide a fallback for RAM, but when you
don't have disk space available and run out of
swappable memory, your machine grinds to a halt. This can
crash your box, requiring a physical reboot for which your
hosting facility may charge you.
</p>
<p>Also, such an outage naturally occurs when you least want
it: when the world has found your website and is beating a
path to your door. If you have enough disk-based swap space
available and the machine gets overloaded, it may get very,
very slow as the system needs to swap memory pages to disk
and back, but when the load decreases the system should
recover. Remember, you still have <code>MaxClients
</code>
to keep things in hand.
</p>
<p>Most unix-like operating systems use designated disk
partitions for swap space. When a system starts up it finds
all swap partitions on the disk(s), by partition type or
because they are listed in the file <code>/etc/fstab
</code>
,and automatically enables them. When adding a disk or
installing the operating system, be sure to allocate enough
swap space to accommodate eventual RAM upgrades.
Reassigning disk space on a running system is a cumbersome
process.
</p>
<p>Plan for available hard drive swap space of at least twice
your amount of RAM, perhaps up to four times in situations
with frequent peaking loads. Remember to adjust this
configuration whenever you upgrade RAM on your system. In a
pinch, you can use a regular file as swap space. For
instructions on how to do this, see the manual pages for
the <code>mkswap
</code>
and <code>swapon
</code>
or <code>swap
</code>
programs.
</p>
<h4><a name="ulimit-files-and-processes" id="ulimit-files-and-processes">ulimit: Files and Processes
</a></h4>
<p>Given a machine with plenty of RAM and processor capacity,
you can run hundreds of Apache processes if necessary. . .
and if your kernel allows it.
</p>
<p>Consider a situation in which several hundred web servers
are running; if some of these need to spawn CGI processes,
the maximum number of processes would occur quickly.
</p>
<p>However, you can change this limit with the command
</p>
<div class="example"><p><code>
ulimit [-H|-S] -u [newvalue]
</code></p></div>
<p>This must be changed before starting the server, since the
new value will only be available to the current shell and
programs started from it. In newer Linux kernels the
default has been raised to 2048. On FreeBSD, the number
seems to be the rather unusual 513. In the default user
shell on this system, <code>csh
</code>
the equivalent is <code>limit
</code>
and works analogous the the Bourne-like <code>ulimit
</code>
:
</p>
<div class="example"><p><code>
limit [-h] maxproc [newvalue]
</code></p></div>
<p>Similarly, the kernel may limit the number of open files per
process. This is generally not a problem for pre-forked
servers, which just handle one request at a time per
process. Threaded servers, however, serve many requests per
process and much more easily run out of available file
descriptors. You can increase the maximum number of open
files per process by running the
</p>
<div class="example"><p><code>ulimit -n [newvalue]
</code></p></div>
<p>command. Once again, this must be done prior to starting
Apache.
</p>
<h4><a name="setting-user-limits-on-system-startup" id="setting-user-limits-on-system-startup">Setting User Limits on System Startup
</a></h4>
<p>Under Linux, you can set the ulimit parameters on bootup by
editing the <code>/etc/security/limits.conf
</code>
file. This file allows you to set soft and hard limits on a
per-user or per-group basis; the file contains commentary
explaining the options. To enable this, make sure that the
file <code>/etc/pam.d/login
</code>
contains the line
</p>
<div class="example"><p><code>session required /lib/security/pam_limits.so
</code></p></div>
<p>All items can have a 'soft' and a 'hard'
limit: the first is the default setting and the second the
maximum value for that item.
</p>
<p>
In FreeBSD's <code>/etc/login.conf
</code>
these resources can be limited or extended system wide,
analogously to <code>limits.conf</code>.
'Soft' limits can be specified with <code>-cur</code>
and 'hard' limits with <code>-max</code>.
</p>
<p>Solaris has a similar mechanism for manipulating limit
values at boot time: In <code>/etc/system</code>
you can set kernel tunables valid for the entire system at
boot time. These are the same tunables that can be set with
the <code>mdb</code>
kernel debugger during run time. The soft and hard limit
corresponding to ulimit -u can be set via:
</p>
<div class="example"><p><code>
set rlim_fd_max=65536<br />
set rlim_fd_cur=2048
</code></p></div>
<p>Solaris calculates the maximum number of allowed processes
per user (<code>maxuprc</code>) based on the total amount
available memory on the system (<code>maxusers</code>).
You can review the numbers with
</p>
<div class="example"><p><code>sysdef -i | grep maximum
</code></p></div>
<p>but it is not recommended to change them.
</p>
<h4><a name="turn-off-unused-services-and-modules" id="turn-off-unused-services-and-modules">Turn Off Unused Services and Modules
</a></h4>
<p>Many UNIX and Linux distributions come with a slew of
services turned on by default. You probably need few of
them. For example, your web server does not need to be
running sendmail, nor is it likely to be an NFS server,
etc. Turn them off.
</p>
<p>On Red Hat Linux, the chkconfig tool will help you do this
from the command line. On Solaris systems <code>svcs</code>
and <code>svcadm</code>
will show which services are enabled and disable them
respectively.
</p>
<p>In a similar fashion, cast a critical eye on the Apache
modules you load. Most binary distributions of Apache
httpd, and pre-installed versions that come with Linux
distributions, have their modules enabled through the
<code class="directive">LoadModule</code> directive.
</p>
<p>Unused modules may be culled: if you don't rely on
their functionality and configuration directives, you can
turn them off by commenting out the corresponding
<code class="directive">LoadModule</code>
lines. Read the documentation on each module’s
functionality before deciding whether to keep it enabled.
While the performance overhead of an unused module is
small, it's also unnecessary.
</p>
</div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
<div class="section">
<h2><a name="caching-content" id="caching-content">Caching Content
</a></h2>
<p>Requests for dynamically generated content usually take
significantly more resources than requests for static content.
Static content consists of simple filespages, images, etc.-on disk
that are very efficiently served. Many operating systems also
automatically cache the contents of frequently accessed files in
memory.
</p>
<p>Processing dynamic requests, on the contrary, can be much more
involved. Running CGI scripts, handing off requests to an external
application server and accessing database content can introduce
significant latency and processing load to a busy web server. Under
many circumstances, performance can be improved by turning popular
dynamic requests into static requests. In this section, two
approaches to this will be discussed.
</p>
<h3><a name="making-popular-pages-static" id="making-popular-pages-static">Making Popular Pages Static
</a></h3>
<p>By pre-rendering the response pages for the most popular queries
in your application, you can gain a significant performance
improvement without giving up the flexibility of dynamically
generated content. For instance, if your application is a
flower delivery service, you would probably want to pre-render
your catalog pages for red roses during the weeks leading up to
Valentine's Day. When the user searches for red roses,
they are served the pre-rendered page. Queries for, say, yellow
roses will be generated directly from the database. The
mod_rewrite module included with Apache is a great tool to
implement these substitutions.
</p>
<h4><a name="example-a-statically-rendered-blog" id="example-a-statically-rendered-blog">Example: A Statically Rendered Blog
</a></h4>
<p>Blosxom is a lightweight web log package that runs as a CGI.
It is written in Perl and uses plain text files for entry
input. Besides running as CGI, Blosxom can be run from the
command line to pre-render blog pages. Pre-rendering pages
to static HTML can yield a significant performance boost in
the event that large numbers of people actually start
reading your blog.
</p>
<p>To run blosxom for static page generation, edit the CGI
script according to the documentation. Set the $static dir
variable to the <code class="directive">DocumentRoot</code>
of the web server, and run the script from the command line
as follows:
</p>
<div class="example"><p><code>$ perl blosxom.cgi -password='whateveryourpassword'
</code></p></div>
<p>This can be run periodically from Cron, after you upload
content, etc. To make Apache substitute the statically
rendered pages for the dynamic content, we’ll use
mod_rewrite. This module is included with the Apache source
code, but is not compiled by default. It can be built with
the server by passing the option <code>--enable-rewrite[=shared]</code>
to the configure command. Many binary distributions of
Apache come with <code class="module"><a href="../mod/mod_rewrite .html">mod_rewrite </a></code> included. The following is an
example of an Apache virtual host that takes advantage of
pre-rendered blog pages:
</p>
<pre class="prettyprint lang-config">
Listen *:8001
<VirtualHost *:8001>
ServerName blog.sandla.org:8001
ServerAdmin sander@temme.net
DocumentRoot "/home/sctemme/inst/blog/httpd/htdocs"
<Directory
"/home/sctemme/inst/blog/httpd/htdocs">
Options +Indexes
Require all granted
RewriteEngine on
RewriteCond %{REQUEST_FILENAME} !-f
RewriteCond %{REQUEST_FILENAME} !-d
RewriteRule ^(.*)$ /cgi-bin/blosxom.cgi/$1 [L,QSA]
</Directory>
RewriteLog /home/sctemme/inst/blog/httpd/logs/rewrite_log
RewriteLogLevel 9
ErrorLog /home/sctemme/inst/blog/httpd/logs/error_log
LogLevel debug
CustomLog /home/sctemme/inst/blog/httpd/logs/access_log common
ScriptAlias /cgi-bin/ /home/sctemme/inst/blog/bin/
<Directory "/home/sctemme/inst/blog/bin">
Options +ExecCGI
Require all granted
</Directory>
</VirtualHost>
</pre>
<p>
The <code class="directive">RewriteCond</code>
and <code class="directive">RewriteRule</code>
directives say that, if the requested resource does not
exist as a file or a directory, its path is passed to the
Blosxom CGI for rendering. Blosxom uses Path Info to
specify blog entries and index pages, so this means that if
a particular path under Blosxom exists as a static file in
the file system, the file is served instead. Any request
that isn't pre- rendered is served by the CGI. This
means that individual entries, which show the comments, are
always served by the CGI which in turn means that your
comment spam is always visible. This configuration also
hides the Blosxom CGI from the user-visible URL in their
Location bar. mod_rewrite is a fantastically powerful and
versatile module: investigate it to arrive at a
configuration that is best for your situation.
</p>
<h3><a name="caching-content-with-mod_cache" id="caching-content-with-mod_cache">Caching Content With mod_cache
</a></h3>
<p>The mod_cache module provides intelligent caching of HTTP
responses: it is aware of the expiration timing and content
requirements that are part of the HTTP specification. The
mod_cache module caches URL response content. If content sent
to the client is considered cacheable, it is saved to disk.
Subsequent requests for that URL will be served directly from
the cache. The provider module for mod_cache, mod_disk_cache,
determines how the cached content is stored on disk. Most
server systems will have more disk available than memory, and
it's good to note that some operating system kernels cache
frequently accessed disk content transparently in memory, so
replicating this in the server is not very useful.
</p>
<p>To enable efficient content caching and avoid presenting the
user with stale or invalid content, the application that
generates the actual content has to send the correct response
headers. Without headers like <code>Etag:</code>,
<code>Last-Modified:</code> or <code>Expires:</code>,
<code class="module"><a href="../mod/mod_cache.html">mod_cache</a></code> can not make the right decision on whether to cache
the content, serve it from cache or leave it alone. When
testing content caching, you may find that you need to modify
your application or, if this is impossible, selectively disable
caching for URLs that cause problems. The mod_cache modules are
not compiled by default, but can be enabled by passing the
option <code>--enable-cache[=shared]</code>
to the configure script. If you use a binary distribution of
Apache httpd, or it came with your port or package collection,
it may have <code class="module"><a href="../mod/mod_cache.html">mod_cache</a></code> already included.
</p>
<h4><a name="example-wiki" id="example-wiki">Example: wiki.apache.org
</a></h4>
<p>
The Apache Software Foundation Wiki is served by MoinMoin.
MoinMoin is written in Python and runs as a CGI. To date, any
attempts to run it under mod_python has been unsuccessful.
The CGI proved to place an untenably high load on the
server machine, especially when the Wiki was being indexed
by search engines like Google. To lighten the load on the
server machine, the Apache Infrastructure team turned to
mod_cache. It turned out <a href="/httpd/MoinMoin">MoinMoin
</a>
needed a small patch to ensure proper behavior behind the
caching server: certain requests can never be cached and
the corresponding Python modules were patched to send the
proper HTTP response headers. After this modification, the
cache in front of the Wiki was enabled with the following
configuration snippet in <code>httpd.conf</code>:
</p>
<pre class="prettyprint lang-config">
CacheRoot /raid1/cacheroot
CacheEnable disk /
# A page modified 100 minutes ago will expire in 10 minutes
CacheLastModifiedFactor .1
# Always check again after 6 hours
CacheMaxExpire 21600
</pre>
<p>This configuration will try to cache any and all content
within its virtual host. It will never cache content for
more than six hours (the <code class="directive"><a href="../mod/mod_cace.html#cachemaxexpire">CacheMaxExpire</a></code>
directive). If no <code>Expires:</code>
header is present in the response, <code class="module"><a href="../mod/mod_cache.html">mod_cache</a></code> will compute
an expiration period from the <code>Last-Modified:</code>
header. The computation using <code class="directive"><a href="../mod/mod_cache.html#cachelastmodifiedfactor">CacheLastModifiedFactor</a></code>
is based on the assumption that if a page was recently
modified, it is likely to change again in the near future
and will have to be re-cached.
</p>
<p>
Do note that it can pay off to <em>disable</em>
the <code>ETag:</code>
header: For files smaller than 1k the server has to
calculate the checksum (usually MD5) and then send out a
<code>304 Not Modified</code>
response, which will use up some CPU and still saturate
the same amount of network resources for the transfer (one
TCP packet). For resources larger than 1k it might prove
CPU expensive to calculate the header for each request.
Unfortunately there does currently not exist a way to cache
these headers.
</p>
<pre class="prettyprint lang-config">
<FilesMatch \.(jpe?g|png|gif|js|css|x?html|xml)>
FileETag None
</FilesMatch>
</pre>
<p>
This will disable the generation of the <code>ETag:</code>
header for most static resources. The server does not
calculate these headers for dynamic resources.
</p>
</div><div class="top"><a href="#page-header"><img alt="top" src="../images/up.gif" /></a></div>
<div class="section">
<h2><a name="further-considerations" id="further-considerations">Further Considerations
</a></h2>
<p>Armed with the knowledge of how to tune a sytem to deliver the
desired the performance, we will soon discover that <em>one</em>
system might prove a bottleneck. How to make a system fit for
growth, or how to put a number of systems into tune will be
discussed in <a href="http://wiki.apache.org/httpd/PerformanceScalingOut">PerformanceScalingOut</a>.
</p>
</div></div>
<div class="bottomlang">
<p><span>Available Languages: </span><a href="../en/misc/perf-scaling.html" title="English"> en </a></p>
</div><div class="top"><a href="#page-header"><img src="../images/up.gif" alt="top" /></a></div><div class="section"><h2><a id="comments_section" name="comments_section">Comments</a></h2><div class="warning"><strong>Notice:</strong><br />This is not a Q&A section. Comments placed here should be pointed towards suggestions on improving the documentation or server, and may be removed again by our moderators if they are either implemented or considered invalid/off-topic. Questions on how to manage the Apache HTTP Server should be directed at either our IRC channel, #httpd, on Freenode, or sent to our <a href="http://httpd.apache.org/lists.html">mailing lists</a>.</div>
<script type="text/javascript"><!--//--><![CDATA[//><!--
var comments_shortname = 'httpd';
var comments_identifier = 'http://httpd.apache.org/docs/trunk/misc/perf-scaling.html';
(function(w, d) {
if (w.location.hostname.toLowerCase() == "httpd.apache.org") {
d.write('<div id="comments_thread"><\/div>');
var s = d.createElement('script');
s.type = 'text/javascript';
s.async = true;
s.src = 'https://comments.apache.org/show_comments.lua?site=' + comments_shortname + '&page=' + comments_identifier;
(d.getElementsByTagName('head')[0] || d.getElementsByTagName('body')[0]).appendChild(s);
}
else {
d.write('<div id="comments_thread">Comments are disabled for this page at the moment.<\/div>');
}
})(window, document);
//--><!]]></script></div><div id="footer">
<p class="apache">Copyright 2013 The Apache Software Foundation.<br />Licensed under the <a href="http://www.apache.org/licenses/LICENSE-2.0">Apache License, Version 2.0</a>.</p>
<p class="menu"><a href="../mod/">Modules</a> | <a href="../mod/quickreference.html">Directives</a> | <a href="http://wiki.apache.org/httpd/FAQ">FAQ</a> | <a href="../glossary.html">Glossary</a> | <a href="../sitemap.html">Sitemap</a></p></div><script type="text/javascript"><!--//--><![CDATA[//><!--
if (typeof(prettyPrint) !== 'undefined') {
prettyPrint();
}
//--><!]]></script>
</body></html>
|
