Uh oh!
There was an error while loading. Please reload this page.
HBASE-22072 High read/write intensive regions may cause long crash - #214
Conversation
Apache-HBase
commented
May 3, 2019
💔 -1 overall
This message was automatically generated. |
Uh oh!
There was an error while loading. Please reload this page.
carp84
commented
May 4, 2019
Please check and fix the checkstyle issue as reported. The change itself looks good but I need some time to read and fully understand the issue and solution as discussed in HBASE-22072. Will be back once done. Thanks. |
Uh oh!
There was an error while loading. Please reload this page.
anoopsjohn
left a comment
There was a problem hiding this comment.
Pls address the volatile comment and then we are good to go.. I think this will solve many strange issues that we were seeing with the new way of delayed compacted files cleanup. Good one.
ramkrish86
commented
May 6, 2019
Made the 'closing' variable to volatile. fixed checkstyle issue and moved the test to a new class. |
Apache-HBase
commented
May 6, 2019
🎊 +1 overall
This message was automatically generated. |
ramkrish86
commented
May 7, 2019
Seems all good now in terms of QA. |
) * HBASE-22072 High read/write intensive regions may cause long crash recovery * Make the 'closing' variable as volatile and move the test case to standlone class
) * HBASE-22072 High read/write intensive regions may cause long crash recovery * Make the 'closing' variable as volatile and move the test case to standlone class
) * HBASE-22072 High read/write intensive regions may cause long crash recovery * Make the 'closing' variable as volatile and move the test case to standlone class
) * HBASE-22072 High read/write intensive regions may cause long crash recovery * Make the 'closing' variable as volatile and move the test case to standlone class
…pache#214) * HBASE-22072 High read/write intensive regions may cause long crash recovery * Make the 'closing' variable as volatile and move the test case to standlone class
…pache#214) * HBASE-22072 High read/write intensive regions may cause long crash recovery * Make the 'closing' variable as volatile and move the test case to standlone class (cherry picked from commit 8d56693) Change-Id: I3bd6e4f030458516aebe3c7adb96712cbef16b20
Allows either updateReaders() or close() to happen at the same time. The JIRA reporter has tested the fix provided.