[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-10-05 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-934692217 Squash and rebase. Ready for merge, pending another round of integration testing for this final version. Will merge tomorrow. As @Apache9 suggested we can commit this and polish

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-10-04 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-933695763 I am making a final pass now to address remaining review comments and checkstyle. -- This is an automated message from the Apache Git Service. To respond to the message,

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-10-03 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-933062764 Let's try this again... Pushed an update for the spotbugs warnings, after rebasing on current master. -- This is an automated message from the Apache Git Service. To respond

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-30 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-931826452 Support Compressor#reinit. We will need this for supporting per table or per CF overrides of compression codec configuration parameters. This will be

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-30 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-931532069 **ITLCC Compression Test** s3n://commoncrawl/crawl-data/CC-MAIN-2021-10/segments/1614178347293.1/warc/CC-MAIN-20210224165708-20210224195708-0.warc.gz 

(310,856

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-30 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-931517322 @virajjasani > > Unlike the other two wrapper implementations (for aircompressor and lz4) this one is slightly different because the Snappy API doesn't adjust the

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-29 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-929758227 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-29 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-930458359 Rebased on HEAD of master branch. Similar to the _hbase-compression-snappy_ module, implemented with the Xerial snappy JNI wrapper, for those who would prefer that over

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-28 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-929758227 I have completed scale load testing with the codecs I am most interested in -- lz4/lz4-java, lz4/aircompressor, and snappy/aircompressor -- and the system is stable with them in

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-24 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-926991545 Backup module flakes continue to pollute the precommit test results for this patch. Actually all tests that are relevant are passing. There is a new javac warning in

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-24 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-926852423 Rebase and squash, except for current round of review feedback and checkstyle/spotbugs findings. -- This is an automated message from the Apache Git Service. To respond to the

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-24 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-926813280 > A bit busy recently and soon it will be a long holiday in China, which is 10.1 - 10.7. So if you guys think the PR is good to go, just go ahead, do not need to wait for me.

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-23 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-926269774 I spent some time today optimizing the LZMA codec. Yes, it is a super slow compressor (by design), but actually decompression is faster than both gzip and bzip2, if... XZ for

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-23 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-926069898 Squash and rebase. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-22 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-925454258 Some updates: Couldn't get Maven to do what I want with shading in the compression modules, not critical, so dropped this. Optimized the case where in compress() if

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-22 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-925144417 Of particular interest is the _hbase-compression-lz4_ module. lz4-java is up to 52% faster than Hadoop native lz4 codec for decompression as measured in microbenchmarks (see

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-22 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-925142647 Rebased and updated with javadoc fix. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-22 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-925078144 All failed tests are unrelated. The test result is actually clean. There are some javadoc issues, I will fix them. There are 8 failures in

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-21 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-924410033 Latest push adds a missing file. There might be a precommit that fails in between. I was able to improve performance by recycling the byte buffers. I will post updated

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-21 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-924113630 The spotbugs results are all like: Could not find resource

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-21 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-924112077 Let me look at the checkstyle and findbugs results. Any idea what "root in the patch failed" means? @busbey @ndimiduk -- This is an automated message from the Apache

[GitHub] [hbase] apurtell commented on pull request #3691: HBASE-26259 Fallback support to pure Java compression

2021-09-20 Thread GitBox
apurtell commented on pull request #3691: URL: https://github.com/apache/hbase/pull/3691#issuecomment-923485224 Including unit test execution time the new modules add approximately 1 minute and 20 seconds to the build. $ ( cd hbase-compression && mvn clean install )