[
https://issues.apache.org/jira/browse/HBASE-7709?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13738867#comment-13738867
]
Jean-Daniel Cryans commented on HBASE-7709:
-------------------------------------------
The logic here is getting big enough that it should be encapsulated, but I
can't think right now if a nice way to do it.
WALEdit will need more javadoc.
The reason WALEdit.scopes wasn't instantiated is that you wouldn't need to keep
an extra TreeMap around if replication wasn't enabled. It seems this patch
changes that assumption. If it makes more sense to always instantiate it then
it should be final, there are a bunch of "if (scopes == null)" that aren't
needed anymore, and "if (scopes != null)" would always be true.
bq. A -> B -> C -> A replication
I think you meant the last cluster to be B? It seems we should refactor
TestMasterReplication a bit because with this patch it would just look like the
same code is running 3 times (to the untrained eye).
> Infinite loop possible in Master/Master replication
> ---------------------------------------------------
>
> Key: HBASE-7709
> URL: https://issues.apache.org/jira/browse/HBASE-7709
> Project: HBase
> Issue Type: Bug
> Components: Replication
> Affects Versions: 0.94.6, 0.95.1
> Reporter: Lars Hofhansl
> Fix For: 0.98.0, 0.95.2, 0.94.12
>
> Attachments: HBASE-7709.patch, HBASE-7709-rev1.patch
>
>
> We just discovered the following scenario:
> # Cluster A and B are setup in master/master replication
> # By accident we had Cluster C replicate to Cluster A.
> Now all edit originating from C will be bouncing between A and B. Forever!
> The reason is that when the edit come in from C the cluster ID is already set
> and won't be reset.
> We have a couple of options here:
> # Optionally only support master/master (not cycles of more than two
> clusters). In that case we can always reset the cluster ID in the
> ReplicationSource. That means that now cycles > 2 will have the data cycle
> forever. This is the only option that requires no changes in the HLog format.
> # Instead of a single cluster id per edit maintain a (unordered) set of
> cluster id that have seen this edit. Then in ReplicationSource we drop any
> edit that the sink has seen already. The is the cleanest approach, but it
> might need a lot of data stored per edit if there are many clusters involved.
> # Maintain a configurable counter of the maximum cycle side we want to
> support. Could default to 10 (even maybe even just). Store a hop-count in the
> WAL and the ReplicationSource increases that hop-count on each hop. If we're
> over the max, just drop the edit.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
