daos icon indicating copy to clipboard operation
daos copied to clipboard
verified publisher icon daos-stack

DAOS-15256 object: coll ops against object with sparse ranks

Open Nasf-Fan opened this issue 1 year ago • 13 comments

We define the degree of rank# sparse for an object as: (the count of ranks on which the object shards reside) / (the count of the ranks in the pool map)

For example: the pool is built on the 10-ranks set: [1, 3, 5, 7, 9, 10, 11, 12, 13, 14]

Case1: If the object shards reside on the ranks set [1, 3, 9, 5], then its sparse degree is equal to 4 / 10 = 0.4

Case2: If the object shards reside on the ranks set [12, 10, 13], then its sparse degree is equal to 3 / 10 = 0.3

Based on such sparse degree, when handle collective operation on client, we will prepare related target bitmap as following:

If the sparse degree is not less than 0.35, then pre-allocate continuous array (that indexed with the rank#) to record related targets bitmap for the collective operation. It is the most efficient solution for locating the rank slot with the rank#. It may waste some DRAM space, but that is temporary before sending out the collective RPC. The percentage of waste DRAM will not be more than 65%.

Otherwise, if the sparse degree is less than 0.5, then use b+tree as the basic data structure to record related targets bitmap. It will introduce some overhead for data comparing and movement when locate the rank slot with the rank#, but it avoids occupying too much DRAM space, especially when the sparse degree is very small.

In practice, before completely scanning the object layout, we cannot exactly know on which ranks the object shards reside. So we need to calculate the sparse degree based on some estimation. For example, when generate the object layout, record the maximum and the minimum rank# on which related object shards reside, then use the width of the ranks range to estimate the ranks count for the given object. Because these ranks may be not continuous, then the count may be over-estimated. Consider that, we can adjust the boundary of sparse degree for above ranks array and b+tree solutions, such as 0.35 => 0.45, or similarly.

Features: ec rebuild

Before requesting gatekeeper:

  • [ ] Two review approvals and any prior change requests have been resolved.
  • [ ] Testing is complete and all tests passed or there is a reason documented in the PR why it should be force landed and forced-landing tag is set.
  • [ ] Features: (or Test-tag*) commit pragma was used or there is a reason documented that there are no appropriate tags for this PR.
  • [ ] Commit messages follows the guidelines outlined here.
  • [ ] Any tests skipped by the ticket being addressed have been run and passed in the PR.

Gatekeeper:

  • [ ] You are the appropriate gatekeeper to be landing the patch.
  • [ ] The PR has 2 reviews by people familiar with the code, including appropriate owners.
  • [ ] Githooks were used. If not, request that user install them and check copyright dates.
  • [ ] Checkpatch issues are resolved. Pay particular attention to ones that will show up on future PRs.
  • [ ] All builds have passed. Check non-required builds for any new compiler warnings.
  • [ ] Sufficient testing is done. Check feature pragmas and test tags and that tests skipped for the ticket are run and now pass with the changes.
  • [ ] If applicable, the PR has addressed any potential version compatibility issues.
  • [ ] Check the target branch. If it is master branch, should the PR go to a feature branch? If it is a release branch, does it have merge approval in the JIRA ticket.
  • [ ] Extra checks if forced landing is requested
    • [ ] Review comments are sufficiently resolved, particularly by prior reviewers that requested changes.
    • [ ] No new NLT or valgrind warnings. Check the classic view.
    • [ ] Quick-build or Quick-functional is not used.
  • [ ] Fix the commit message upon landing. Check the standard here. Edit it to create a single commit. If necessary, ask submitter for a new summary.

Nasf-Fan avatar Feb 27 '24 16:02 Nasf-Fan

Ticket title is 'Optimize object shards scanning for collective operation' Status is 'In Progress' https://daosio.atlassian.net/browse/DAOS-15256

github-actions[bot] avatar Feb 27 '24 16:02 github-actions[bot]

Test stage Unit Test on EL 8.8 completed with status UNSTABLE. https://build.hpdd.intel.com/job/daos-stack/job/daos//view/change-requests/job/PR-13881/1/testReport/

daosbuild1 avatar Feb 27 '24 17:02 daosbuild1

Test stage Functional on EL 8.8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/4/execution/node/1173/log

daosbuild1 avatar Feb 29 '24 20:02 daosbuild1

Test stage Functional on EL 8.8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/5/execution/node/1219/log

daosbuild1 avatar Mar 01 '24 13:03 daosbuild1

Test stage Functional on EL 8.8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/6/execution/node/1149/log

daosbuild1 avatar Mar 03 '24 12:03 daosbuild1

telemetry_list failed for DAOS-15337, to be retested.

Nasf-Fan avatar Mar 04 '24 00:03 Nasf-Fan

Test stage Functional Hardware Medium Verbs Provider completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/7/execution/node/1447/log

daosbuild1 avatar Mar 04 '24 17:03 daosbuild1

Test stage Functional Hardware Medium completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/7/execution/node/1431/log

daosbuild1 avatar Mar 04 '24 22:03 daosbuild1

Test stage Functional on EL 8.8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/12/execution/node/1182/log

daosbuild1 avatar Mar 07 '24 21:03 daosbuild1

Test stage NLT on EL 8.8 completed with status UNSTABLE. https://build.hpdd.intel.com/job/daos-stack/job/daos//view/change-requests/job/PR-13881/13/testReport/

daosbuild1 avatar Mar 08 '24 01:03 daosbuild1

Test stage Functional Hardware Large completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/14/execution/node/1406/log

daosbuild1 avatar Mar 09 '24 11:03 daosbuild1

Test stage Functional Hardware Large completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/18/execution/node/1401/log

daosbuild1 avatar Mar 13 '24 23:03 daosbuild1

Test stage Functional Hardware Medium completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/18/execution/node/1447/log

daosbuild1 avatar Mar 14 '24 00:03 daosbuild1

Test stage Functional on EL 8.8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/19/execution/node/1173/log

daosbuild1 avatar Mar 20 '24 16:03 daosbuild1

Test stage Functional on EL 8.8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/20/execution/node/1198/log

daosbuild1 avatar Mar 21 '24 15:03 daosbuild1

Test stage Build on EL 8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/21/execution/node/315/log

daosbuild1 avatar Mar 22 '24 02:03 daosbuild1

Test stage Build RPM on EL 9 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/21/execution/node/368/log

daosbuild1 avatar Mar 22 '24 02:03 daosbuild1

Test stage Build RPM on Leap 15.5 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/21/execution/node/300/log

daosbuild1 avatar Mar 22 '24 02:03 daosbuild1

Test stage Build RPM on EL 8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/21/execution/node/346/log

daosbuild1 avatar Mar 22 '24 02:03 daosbuild1

Test stage Build DEB on Ubuntu 20.04 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/21/execution/node/297/log

daosbuild1 avatar Mar 22 '24 02:03 daosbuild1

Test stage Build on Leap 15.5 with Intel-C and TARGET_PREFIX completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/21/execution/node/517/log

daosbuild1 avatar Mar 22 '24 02:03 daosbuild1

Test stage Functional on EL 8.8 completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/22/execution/node/1219/log

daosbuild1 avatar Mar 22 '24 22:03 daosbuild1

Test stage Functional Hardware Medium completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/23/execution/node/1405/log

daosbuild1 avatar Mar 26 '24 23:03 daosbuild1

Test stage Functional Hardware Medium Verbs Provider completed with status FAILURE. https://build.hpdd.intel.com//job/daos-stack/job/daos/view/change-requests/job/PR-13881/23/execution/node/1451/log

daosbuild1 avatar Mar 27 '24 04:03 daosbuild1