Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

HPCC-21091 Fix smart join race condition causing crash #11981

Merged
merged 1 commit into from Dec 14, 2018
Merged

HPCC-21091 Fix smart join race condition causing crash #11981

merged 1 commit into from Dec 14, 2018

Conversation

jakesmith
Copy link
Member

@jakesmith jakesmith commented Dec 6, 2018 •
edited

  1. Fix problems when row arrays spilt under memory pressure.
    Code was thread unsafe – rows could be added and row array expanded
    whilst saving, and subsequent flushes of any remaining rows was not using
    spillable array correctly and could hit null rows and cause the crash
    reported by this JIRA.

NB: this should resolve an old issue (HPCC_17331) that has some debugging
code (still left in for now).

  1. Improve choice of row array to spill. Used to check if any had any rows
    to process and spill in order, which meant if rows dripping in from
    different sources and memory pressure was trying to flush, would spill lots
    of small sets. Change to round robin through row arrays.

  2. Move some of the code that cleans, compacts and prepares the row arrays
    after oom event trying to prepare global lookup, to an earlier common point.
    And before compacting ensure row array is force flushed to ensure relocated,
    in the event it has spilt.

  3. Add/change a few comments

Signed-off-by: Jake Smith jake.smith@lexisnexisrisk.com

Type of change:

  • This change is a bug fix (non-breaking change which fixes an issue).
  • This change is a new feature (non-breaking change which adds functionality).
  • This change improves the code (refactor or other change that does not change the functionality)
  • This change fixes warnings (the fix does not alter the functionality or the generated code)
  • This change is a breaking change (fix or feature that will cause existing behavior to change).
  • This change alters the query API (existing queries will have to be recompiled)

Checklist:

  • My code follows the code style of this project.
    • My code does not create any new warnings from compiler, build system, or lint.
  • The commit message is properly formatted and free of typos.
    • The commit message title makes sense in a changelog, by itself.
    • The commit is signed.
  • My change requires a change to the documentation.
    • I have updated the documentation accordingly, or...
    • I have created a JIRA ticket to update the documentation.
    • Any new interfaces or exported functions are appropriately commented.
  • I have read the CONTRIBUTORS document.
  • The change has been fully tested:
    • I have added tests to cover my changes.
    • All new and existing tests passed.
    • I have checked that this change does not introduce memory leaks.
    • I have used Valgrind or similar tools to check for potential issues.
  • I have given due consideration to all of the following potential concerns:
    • Scalability
    • Performance
    • Security
    • Thread-safety
    • Premature optimization
    • Existing deployed queries will not be broken
    • This change fixes the problem, not just the symptom
    • The target branch of this pull request is appropriate for such a change.
  • There are no similar instances of the same problem that should be addressed
    • I have addressed them here
    • I have raised JIRA issues to address them separately
  • This is a user interface / front-end modification
    • I have tested my changes in multiple modern browsers
    • The component(s) render as expected

Testing:

@hpcc-jirabot
Copy link

https://track.hpccsystems.com/browse/HPCC-21091
Jira updated

@jakesmith jakesmith force-pushed the hpcc-21091 branch 2 times, most recently from 0d11b38 to e90c2fc Compare December 6, 2018 10:52
@jakesmith
Copy link
Member Author

@ghalliday - please review

@ghalliday
Copy link
Member

Need to update the check boxes.

ghalliday
ghalliday reviewed Dec 7, 2018
View reviewed changes
Copy link
Member

@ghalliday ghalliday left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@jakesmith code looks generally good - but found one problem

thorlcr/activities/lookupjoin/thlookupjoinslave.cpp Show resolved Hide resolved
@jakesmith
Copy link
Member Author

@ghalliday - please see fix for nextRhsToSpill issue.

@ghalliday
Copy link
Member

@jakesmith looks good. Please squash and fill in the checkboxes.

@jakesmith
Copy link
Member Author

@ghalliday - squashed and filled checkboxes.

@ghalliday
Copy link
Member

@richardkchapman should this be considered for 7.0.6?

@richardkchapman
Copy link
Member

@ghalliday Yes I should think so, if we are confident in the fix

@ghalliday
Copy link
Member

@jakesmith I have checked through again, and I think it would be worth targeting 7.0.6

@jakesmith
HPCC-21091 Fix smart join race condition causing crash
ea5e17f 
1) Fix problems when row arrays spilt under memory pressure.
Code was thread unsafe – rows could be added and row array expanded
whilst saving, and subsequent flushes of any remaining rows was not using
spillable array correctly and could hit null rows and cause the crash
reported by this JIRA.

NB: this should resolve an old issue (HPCC_17331) that has some debugging
code (still left in for now).

2) Improve choice of row array to spill. Used to check if any had any rows
to process and spill in order, which meant if rows dripping in from
different sources and memory pressure was trying to flush, would spill lots
of small sets. Change to round robin through row arrays.

3) Move some of the code that cleans, compacts and prepares the row arrays
after oom event trying to prepare global lookup, to an earlier common point.
And before compacting ensure row array is force flushed to ensure relocated,
in the event it has spilt.

4) Add/change a few comments

Signed-off-by: Jake Smith <jake.smith@lexisnexisrisk.com>
@jakesmith jakesmith changed the base branch from candidate-7.2.0 to candidate-7.0.6 December 12, 2018 09:09
@jakesmith
Copy link
Member Author

@ghalliday - retargeted to candidate-7.0.6

@HPCCSmoketest
Copy link
Contributor

Automated Smoketest: ✅
OS: centos 7.4.1708 (Linux 3.10.0-327.28.3.el7.x86_64)
Sha: ea5e17f
Build: success
Build: success
HPCC Start: OK

Unit tests result:

Test total passed failed errors timeout
unittest 95 95 0 0 0
wutoolTest(Dali) 19 19 0 0 0
wutoolTest(Cassandra) 19 19 0 0 0

Regression test result:

phase total pass fail
setup (hthor) 11 11 0
setup (thor) 11 11 0
setup (roxie) 11 11 0
test (hthor) 808 808 0
test (thor) 732 732 0
test (roxie) 881 881 0

HPCC Stop: OK
Time stats:

Prep time Build time Package time Install time Start time Test time Stop time Summary
17 sec (00:00:17) 187 sec (00:03:07) 0 sec (00:00:00) 3 sec (00:00:03) 18 sec (00:00:18) 1495 sec (00:24:55) 19 sec (00:00:19) 1739 sec (00:28:59)

@ghalliday ghalliday merged commit c605352 into hpcc-systems:candidate-7.0.6 Dec 14, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@ghalliday ghalliday ghalliday left review comments

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
5 participants
@jakesmith @hpcc-jirabot @ghalliday @richardkchapman @HPCCSmoketest