Performance test on Jetty/HSQLDB with a single wiki between 16.10.16 and 17.10.2
Environment
- Windows 11 64bit version 25H2
- Quad core i7-9750H
- Oracle JDK 21.0.9 (build 21.0.9+7-LTS-338)
- Dumbbench 0.505
- VisualVM 2.2
- permdir in a SSD
- standard Jetty/HSQLDB package
- only main wiki
- configuration:
- Admin user:
- hidden documents enabled
- Admin user:
- import
Most of response time statistics are made with Dumbbench, you can install it using cpan.
For example on Debian/Ubuntu:
$ sudo apt-get install cpanminus
$ sudo cpanm Module::Install
$ sudo cpanm DumbbenchOn Windows 11:
- Install perl from https://strawberryperl.com/
- Go to Windows Terminal and install Dumbbench:
PS C:\> perl
PS C:\> cpanm Dumbbench (If needed) Download wget.exe (64 bit) from https://eternallybored.org/misc/wget/
Make a folder for testing the instances and put wget.exe there.
Unless specifically indicated otherwise all the tests are executed with a fully initialized instance (core extensions don't run anymore, SOLR index is fully up to date, etc.) and logged in with Admin user. Basically except for a few special tests we are using an instance that already been started and run until the proc is flat for some time.
Login with Admin user (using port 1115):
.\wget.exe --save-cookies cookies.txt --post-data 'j_username=Admin&j_password=admin&j_rememberme=true' "http://127.0.0.1:1115/xwiki/bin/loginsubmit/XWiki/XWikiLogin?j_username=Admin&j_password=admin&j_rememberme=true"Tests
Summary
"similar": difference is lower than 10%
"slightly": difference is lower than 20%
Speed
| Actions | Difference | |
|---|---|---|
| Jetty startup | similar | |
| First access | not existing page without UI | similar |
| not existing page with UI | similar | |
| Reload | not existing page without UI | +25.77% faster |
| not existing page with UI | +30.62% faster | |
| empty page without UI | similar | |
| empty page with UI | +37.15% faster | |
| Main.WebHome without UI | similar | |
| Main.WebHome with UI | +40.75% faster | |
| SOLR | Full SOLR reindex | +33.33% faster |
| SOLR sync when index is empty | - 80.13% slower 1 | |
| SOLR sync when there is nothing to do | - 50.59% slower 1 | |
| Result of search finding lots of results | +34.37% faster | |
| Result of search finding one result | +39.70% faster | |
| Rendering | Page with 1000 macros without UI | similar |
| Page with 1000 html macros without UI | similar | |
| Wiki creation | From flavor | similar |
| From template | similar | |
Memory
| Actions | Difference |
|---|---|
| Heap Memory after jetty startup | +5 MB |
| Heap Memory after full SOLR index | +14 MB |
1 For the SOLR Sync tests lower times was reported https://jira.xwiki.org/browse/XWIKI-23920, but according to some tests made on other workstation, the time were much better than the ones obtained here.
Jetty startup
Unit: milliseconds
This information is obtained from the log with the line looking like 2026-01-15 14:53:57.343:INFO :oejs.Server:main: Started oejs.Server@74a6f9c1{STARTING}[12.1.5,sto=0] @31804ms.
- 16.10.16: 28731
- 17.10.2: 31804
First access
First HTTP request after restarting Jetty.
First access to not existing page without UI
(/xwiki/bin/get/NoSpace/NoPage)
Unit: milliseconds
Firefox network statistics to load the root resource (image/js/css/etc. resources are not taken into account).
- 16.10.16: 36341
- 17.10.2: 37474
First access to not existing page with UI
(/xwiki/bin/view/NoSpace/NoPage)
Unit: seconds
Done by hand with a chronometer (image/js/css/etc. resources are not taken into account).
- 16.10.16: 45
- 17.10.2: 43
Reload
XWiki is doing nothing on background and everything is both in server and client cache already. Values are calculated using dumbbench (image/js/css/etc. resources are not taken into account).
Reload of not existing page without UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/get/NoSpace/NoPage"Unit: milliseconds
- 16.10.16: 208
- 17.10.2: 154
Reload of not existing page with UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/view/NoSpace/NoPage"Unit: milliseconds
- 16.10.16: 219
- 17.10.2: 152
Reload of empty page without UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/get/EmptySpace/EmptyPage"Unit: milliseconds
- 16.10.16: 17
- 17.10.2: 18
Reload of empty page with UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/view/EmptySpace/EmptyPage"Unit: milliseconds
- 16.10.16: 270
- 17.10.2: 170
Reload of Main.WebHome without UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/get/Main/WebHome"Unit: milliseconds
- 16.10.16: 24
- 17.10.2: 23
Reload of Main.WebHome with UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/view/Main/WebHome"Unit: milliseconds
- 16.10.16: 285
- 17.10.2: 169
SOLR
Full SOLR reindex
Unit: seconds
Note: On XWiki versions > XWiki 17.8.0+, delete the data\cache\solr\search_9 folder, restart, load home page and after INFO o.x.s.s.i.j.IndexerJob - Finished job of type [solr.indexer] with identifier [[solr, indexer]] log, take the last elapsedTime from <permdir>/data/jobs/status/3/solr/indexer/status.xml from status.xml.zip.
- 16.10.16: 54
- 17.10.2: 36
SOLR sync
Only the thread determining what need to be indexed and not the actual indexing.
Note: On XWiki versions > XWiki 17.8.0+, calculate the elapsed time from the beginning of the Indexing job until the synchronization is done (after the log INFO o.x.s.s.i.j.IndexerJob - {} documents added, {} deleted and {} updated during the synchronization of the Solr index.
The timestamps are found in <permdir>/data/jobs/status/3/solr/indexer/log.xml.
when index is empty
Unit: milliseconds
- 16.10.16: 1509 (1.51 s)
- 17.10.2: 2728 (2.72 s)
Test results for XWiki 17.10.2:
Test 1: 2961 (2.96 s)
Test 2: 2728 (2.72 s)
Test 3: 2824 (2.82 s)
Test 4: 2737 (2.73 s)
Test 5: 2790 (2.79 s)
when there is nothing to do
Unit: milliseconds
- 16.10.16: 502 ms
- 17.10.2: 756 ms
Test results for XWiki 17.10.2:
Test 1: 756 ms
Test 2: 795 ms
Test 3: 840 ms
Test 4: 787 ms
Test 5: 820 ms
Search
XWiki is doing nothing on background and everything is both in server and client cache already. Values are calculated using dumbbench (image/js/css/etc. resources are not taken into account).
Unit: milliseconds
Result of search finding lots of results
Login with Admin user and make sure hidden document display is enabled.
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt """http://127.0.0.1:1115/xwiki/bin/view/Main/Search?sort=score&sortOrder=desc&r=1&f_type=DOCUMENT&text=*"""- 16.10.16: 448
- 17.10.2: 294
Result of search finding one result
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt """http://127.0.0.1:1115/xwiki/bin/view/Main/Search?sort=score&sortOrder=desc&r=1&f_type=DOCUMENT&text=name%3A1000Macros"""- 16.10.16: 336
- 17.10.2: 203
Rendering
Page with 1000 macros without UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/get/Test/1000Macros"Unit: milliseconds
- 16.10.16: 26
- 17.10.2: 28
Page with 1000 html macros without UI
dumbbench --float -p 0.1 -i 100 -- wget -q -o NUL -O NUL --load-cookies cookies.txt "http://127.0.0.1:1115/xwiki/bin/get/Test/1000HTMLMacros"Unit: milliseconds
- 16.10.16: 460
- 17.10.2: 428
Wiki creation
Unit: nanoseconds
From flavor
Create the wiki test1 from default flavor and get the time spent from the job (</rootStep> in <permdir>/jobs/status/wikicreation/createandinstall/test1/status.xml).
- 16.10.16: 60324437600 (60 s)
- 17.10.2: 60419739900 (60 s)
From template
Make test1 a template wiki.
Create the wiki test2 from wiki test1 and get the time spend from the job (</rootStep> in <permdir>/jobs/status/wikicreation/createandinstall/test2/status.xml).
- 16.10.16: 48008664300 (48 s)
- 17.10.2: 48373255500 (48 s)
Heap Memory
Attach VisualVM to get the information.
Unit: MB
Heap Memory after jetty startup
Start jetty and wait until Java is not doing anything on both (wait for potential initialization of various stuff like core extension cache, etc). Force garbage collector before looking at "used".
- 16.10.16: 124 used, 273 allocated
- 17.10.2: 129 used, 250 allocated
Heap Memory after full SOLR index
Delete the data/solr folder, restart, load home page then wait until nothing is happening in background. Force garbage collector before looking at "used".
Unit: MB
- 16.10.16: 312 used, 1059 allocated
- 17.10.2: 326 used, 1073 allocated