[Gluster-devel] 3.7.9 update

Kaushal M kshlmster at gmail.com
Mon Mar 14 03:10:06 UTC 2016


Hey Vijay,

Awesome work on the tests!

Could you please list down the tests you run for releases?  This would
help the other release-maintainers could run them as well. We could
also try automating these tests, so that it becomes even easier.

~kaushal

On Sun, Mar 13, 2016 at 10:50 PM, Vijay Bellur <vbellur at redhat.com> wrote:
> Hey All,
>
> I have been running tests with the latest HEAD of release-3.7  on a 2x2
> distributed replicated volume. Here are some updates:
>
> - Write Performance has seen an improvement as seen by running perf-test.sh
> [1]
>
>
> v3.7.9 with FUSE client
>
> Testname                Time
> emptyfiles_create       961.83
> emptyfiles_delete       600.08
> smallfiles_create       1508.38
> smallfiles_rewrite      1325.60
> smallfiles_read         598.50
> smallfiles_reread       384.65
> smallfiles_delete       623.66
> largefile_create        18.33
> largefile_rewrite       19.17
> largefile_read          11.44
> largefile_reread        0.31
> largefile_delete        0.66
> directory_crawl_create  981.21
> directory_crawl         30.64
> directory_recrawl       28.01
> metadata_modify         1117.92
> directory_crawl_delete  423.08
>
> v3.7.8 with FUSE client
>
> Testname                Time
> emptyfiles_create       953.87
> emptyfiles_delete       577.46
> smallfiles_create       1837.33
> smallfiles_rewrite      2349.37
> smallfiles_read         604.22
> smallfiles_reread       394.48
> smallfiles_delete       629.74
> largefile_create        73.86
> largefile_rewrite       76.23
> largefile_read          11.36
> largefile_reread        0.31
> largefile_delete        0.65
> directory_crawl_create  985.16
> directory_crawl         31.10
> directory_recrawl       26.94
> metadata_modify         1422.60
> directory_crawl_delete  382.57
>
> Hopefully this addresses the write performance drop we observed with 3.7.8.
>
> - Regular file system test tools like iozone, dbench etc. are running fine
> with the fuse client.
>
> - Rolling upgrade from 3.7.8 to the latest release-3.7 HEAD worked fine with
> I/O happening from a fuse client.
>
> - There is a memory leak in FUSE client that I observed while running
> perf-test.sh. A statedump revealed that there was a ref leak on several
> inodes. I have sent a possible patch [2] which addressed problems in my test
> setup. This does need careful review and more testing. Given the memory
> leaks we have been observing with fuse, I feel that it would be good to
> review mount/fuse for possible leaks and run more tests before releasing
> 3.7.9. I am looking at pushing out tagging by 2-3 days to mid week to
> accomplish this. Niels, Raghavendra - can you provide additional help with
> reviewing here?
>
> - Tiering has seen a lot of patches in 3.7.9. Dan, Nithya - can you please
> assist in preparation of release notes by summarizing the changes and
> providing inputs on the general readiness of tiering?
>
> Thanks,
> Vijay
>
> [1] https://github.com/avati/perf-test/blob/master/perf-test.sh
>
> [2] http://review.gluster.org/#/c/13689/
>
>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at gluster.org
> http://www.gluster.org/mailman/listinfo/gluster-devel


More information about the Gluster-devel mailing list