coordination hash table gets additional member when using ZK, throwing off hashing
Bug #1535857 reported by
Adam Gandelman
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Astara |
Fix Released
|
Medium
|
Adam Gandelman |
Bug Description
When tooz is configured to use zookeeper, the membership list reported by the library always includes an additional 'leader' member. When using the member list to create the hash table, we end up with a partition in the table for a member that doesn't actually exist. We should account for this zookeeper-ism and just ignore that member.
Changed in astara: | |
assignee: | nobody → Adam Gandelman (gandelman-a) |
importance: | Undecided → Medium |
milestone: | none → mitaka-2 |
Changed in astara: | |
status: | New → In Progress |
To post a comment you must log in.
Reviewed: https:/ /review. openstack. org/261098 /git.openstack. org/cgit/ openstack/ astara/ commit/ ?id=8ad5cffc802 845f4a1f68440a9 63b99eac935b0f
Committed: https:/
Submitter: Jenkins
Branch: master
commit 8ad5cffc802845f 4a1f68440a963b9 9eac935b0f
Author: Adam Gandelman <email address hidden>
Date: Wed Dec 23 12:50:09 2015 -0800
Filter the 'leader' member from cluster membership list
When configured to use zookeeper, tooz reports an additional member
named 'leader' in its membership list. This can mess with construction
of the hash table. This removes it from the list of members we use for
the DHT.
Closes-Bug: #1535857
Change-Id: I1d3d7c27984c99 c46d6d26e0b9f21 c3b5a511617