Felix Kölzow
2020-May-27 07:07 UTC
[Gluster-users] Readdirp (ls -l) Performance Improvement
Dear Rafi, thanks for your effort. I think this is of great interest of many gluster users. Thus, I would really encourage you to test and to further improve this feature. Maybe it is beneficial to create a certain guideline which things should be tested to make this feature really ready for productive use. Thanks in advance. Felix On 27/05/2020 07:56, RAFI KC wrote:> > Hi All, > > I have been working on POC to improve readdirp performance > improvement. At the end of the experiment, The results are showing > promising result in performance, overall there is a 104% improvement > for full filesystem crawl compared to the existing solution. Here is > the short test numbers. The tests were carried out in 16*3 setup with > 1.5 Million dentries (Both files and dir). The system also contains > some empty directories. *In the result the proposed solution is 287% > faster than the plane volume and 104% faster than the parallel-readdir > based solution.* > > * > * > > Configuration > > > > Plain volume > > > > Parallel-readdir > > > > Proposed Solution > > FS Crawl Time in Seconds > > > > 16497.523 > > > > 8717.872 > > > > 4261.401 > > In short, the basic idea behind the proposal is the efficient managing > of readdir buffer in gluster along with prefetching the dentries for > intelligent switch-over to the next buffer. The detailed problem > description, deign description and results are available in the > doc.https://docs.google.com/document/d/10z4T5Sd_-wCFrmDrzyQtlWOGLang1_g17wO8VUxSiJ8/edit > > > > If anybody can help with the testing on a different kind of workloads, > I would be very happy to assist. If wanted to test the patch and run a > performance test on your setup, I could help with back-porting the > patch to the version of your choice. > > > https://review.gluster.org/24469 > > https://review.gluster.org/24470 > > > Regards > > Rafi KC > > > ________ > > > > Community Meeting Calendar: > > Schedule - > Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC > Bridge: https://bluejeans.com/441850968 > > Gluster-users mailing list > Gluster-users at gluster.org > https://lists.gluster.org/mailman/listinfo/gluster-users-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20200527/6cb248b2/attachment.html>
Hi Felix, Thanks for your mail. I will test it more to make sure that it doesn't break anything. Also I have added a configuration key for easier switching to the older code in case if there is any problem. If you can help me in any manner in testing or performance numbers, please let me know. Regards Rafi KC On 27/05/20 12:37 pm, Felix K?lzow wrote:> > Dear Rafi, > > > thanks for your effort. I think this is of great interest of many > gluster users. Thus, I would really encourage you to > > test and to further improve this feature. Maybe it is beneficial to > create a certain guideline which things should be tested > > to make this feature really ready for productive use. > > > Thanks in advance. > > Felix > > On 27/05/2020 07:56, RAFI KC wrote: >> >> Hi All, >> >> I have been working on POC to improve readdirp performance >> improvement. At the end of the experiment, The results are showing >> promising result in performance, overall there is a 104% improvement >> for full filesystem crawl compared to the existing solution. Here is >> the short test numbers. The tests were carried out in 16*3 setup with >> 1.5 Million dentries (Both files and dir). The system also contains >> some empty directories. *In the result the proposed solution is 287% >> faster than the plane volume and 104% faster than the >> parallel-readdir based solution.* >> >> * >> * >> >> Configuration >> >> >> >> Plain volume >> >> >> >> Parallel-readdir >> >> >> >> Proposed Solution >> >> FS Crawl Time in Seconds >> >> >> >> 16497.523 >> >> >> >> 8717.872 >> >> >> >> 4261.401 >> >> In short, the basic idea behind the proposal is the efficient >> managing of readdir buffer in gluster along with prefetching the >> dentries for intelligent switch-over to the next buffer. The detailed >> problem description, deign description and results are available in >> the >> doc.https://docs.google.com/document/d/10z4T5Sd_-wCFrmDrzyQtlWOGLang1_g17wO8VUxSiJ8/edit >> >> >> >> If anybody can help with the testing on a different kind of >> workloads, I would be very happy to assist. If wanted to test the >> patch and run a performance test on your setup, I could help with >> back-porting the patch to the version of your choice. >> >> >> https://review.gluster.org/24469 >> >> https://review.gluster.org/24470 >> >> >> Regards >> >> Rafi KC >> >> >> ________ >> >> >> >> Community Meeting Calendar: >> >> Schedule - >> Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC >> Bridge:https://bluejeans.com/441850968 >> >> Gluster-users mailing list >> Gluster-users at gluster.org >> https://lists.gluster.org/mailman/listinfo/gluster-users > > ________ > > > > Community Meeting Calendar: > > Schedule - > Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC > Bridge: https://bluejeans.com/441850968 > > Gluster-users mailing list > Gluster-users at gluster.org > https://lists.gluster.org/mailman/listinfo/gluster-users-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20200527/1554bb12/attachment.html>