SLIDE 1
Improving Performance of Directory Operations in Gluster Poornima - - PowerPoint PPT Presentation
Improving Performance of Directory Operations in Gluster Poornima - - PowerPoint PPT Presentation
Improving Performance of Directory Operations in Gluster Poornima Gurusiddaiah, Raghavendra Gowdappa, Manoj Pillai Red Hat March 2017
SLIDE 2
SLIDE 3
SLIDE 4
SLIDE 5
SLIDE 6
SLIDE 7
SLIDE 8
SLIDE 9
- ○
- ○
- ○
SLIDE 10
- ○
- ○
SLIDE 11
- ○
- ○
○
- ○
○
SLIDE 12
- ○
SLIDE 13
Performance Evaluation: Operations Used
find-name
- find <mount-point> -name <file-name>
find-size
- find <mount-point> -size <file-size>
untar
- cd <mount-point>; tar -xf <dataset.tar>
SLIDE 14
Testbed
Servers
- 4 servers, 36 drives per server
○ Upto 144 bricks, each brick on single physical disk ○ Each brick is an XFS filesystem on a thin LV (RHEL thin-provisioning)
- 10GbE link used for tests
Clients
- single client
- 10GbE links used for tests
SLIDE 15
Software Versions and Tuning
OS
- RHEL 7.3
Gluster
- Baseline: 3.10
- Compared to 3.10+enhancements
Gluster Tuning
- gluster volume set <volname> cluster.lookup-optimize on
- gluster volume set <volname> client.event-threads 4
- gluster volume set <volname> performance.client-io-threads on
SLIDE 16
SLIDE 17
SLIDE 18
SLIDE 19
SLIDE 20
SLIDE 21
- ○
- ○
SLIDE 22
- ○
○
- ○
SLIDE 23
- gluster volume set <volname> group metadata-cache
SLIDE 24
- ○
performance.parallel-readdir on ○ cluster.readdir-hashed on
SLIDE 25
SLIDE 26
SLIDE 27
SLIDE 28
SLIDE 29
- ○
○
- ○
○ ○
SLIDE 30
SLIDE 31
SLIDE 32