Uploaded image for project: 'Infinispan'
  1. Infinispan
  2. ISPN-5021

Nodes that finish the rebalance later can see outdated values

    XMLWordPrintable

Details

    Description

      Copied from ISPN-4444

      If the CH_UPDATE command is delayed on the old owner, the new owners might update the key without the old owner knowing, and a locality check on the old owner won't help.

      I remember one thing that struck me when reading the Raft algorithm was that they install configuration changes symmetrically, in 3 phases. We might need to do the same for our rebalance:
      1. T0: read_ch=old, write_ch=old
      2. start a rebalance
      3. T1: read_ch=old, write_ch=old+new
      4. new owners have all the data
      5. T2: read_ch=new, write_ch=old+new
      6. remove old cache entries and ignore further writes
      7. T3: read_ch=new, write_ch=new

      Attachments

        Issue Links

          causes

          Bug - A problem that impairs or prevents the functions of the product. JBEAP-8958 RpcException: No successful response

          • Critical - To be worked on immediately following BLOCKER issues.
          • Verified
          is blocked by

          Bug - A problem that impairs or prevents the functions of the product. ISPN-5046 PartitionHandling: split during commit can leave the cache inconsistent after merge

          • Critical - To be worked on immediately following BLOCKER issues.
          • Closed
          is incorporated by

          Bug - A problem that impairs or prevents the functions of the product. ISPN-7958 Backports for 8.2.8.Final

          • Major - To be worked after higher priority work items are addressed. When the severity is high and the effort to change it is low to moderate, this term is chosen. Issues in this category likely have an existing workaround but implementation or execution may be non-trivial.
          • Closed

          Task - Represents a small unit of work that is not end-user facing. ISPN-7327 Backports for 8.2.7.Final

          • Blocker - To be worked above all other priorities. This term is chosen when the severity of the issue is very high, has no workaround, or the effort for the change is comparatively low. Issues that may be very publicly visible and could generate significant media attention may also drive a higher priority.
          • Closed
          is related to

          Bug - A problem that impairs or prevents the functions of the product. ISPN-4444 After state transfer, a node is able to read keys it no longer owns from its data container

          • Critical - To be worked on immediately following BLOCKER issues.
          • Closed

          Task - Represents a small unit of work that is not end-user facing. ISPN-5016 Specify and document cache consistency guarantees

          • Critical - To be worked on immediately following BLOCKER issues.
          • Closed
          relates to

          Bug - A problem that impairs or prevents the functions of the product. ISPN-9028 Write-only segments should be invalidated during the READ_NEW phase

          • Major - To be worked after higher priority work items are addressed. When the severity is high and the effort to change it is low to moderate, this term is chosen. Issues in this category likely have an existing workaround but implementation or execution may be non-trivial.
          • Closed

          Activity

            People

              rvansa1@redhat.com Radim Vansa (Inactive)
              dberinde@redhat.com Dan Berindei (Inactive)
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: