[
https://issues.apache.org/jira/browse/CASSANDRA-1072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12964786#action_12964786
]
Jonathan Ellis edited comment on CASSANDRA-1072 at 11/29/10 11:06 AM:
----------------------------------------------------------------------
Is this an accurate summary of the current state of 1072 and CASSANDRA-1546?
1072-pro: tested in production. Supports counters in subcolumns.
1072-con: IP-based partitioning is prone to catastrophic data loss if node IPs
change (which is fairly common on an EC2/EBS deployment) or on
decommission-after-bootstrap-before-repair (which is fairly uncommon ime) or
possibly in other scenarios as well; changing 1072 to use UUIDs instead of IPs
to fix this would be a substantial undertaking. In the worst case, there is no
ConsistencyLevel guarantee higher than ONE even with repair-on-write.
1546-pro: ConsistencyLevel support is built in; support for optional "marker"
allows retry on failure as well. "LocalCounterColumn" approach limits the
damage that can be done by bugs compared to relying on post-transfer "cleaning"
of non-local data. Supports node IP changes.
1546-con: essentially untested. Some known corner cases involving node
movement need to be addressed. Would require using a custom comparator to
emulate counters in subcolumns.
was (Author: jbellis):
Is this an accurate summary of the current state of 1072 and CASSANDRA-1546?
1072-pro: tested in production. Supports counters in subcolumns.
1072-con: IP-based partitioning is prone to catastrophic data loss if node IPs
change (which is fairly common on an EC2/EBS deployment) or on
decommission-after-bootstrap-before-repair (which is fairly uncommon ime) or
possibly in other scenarios as well; changing 1072 to use UUIDs instead of IPs
to fix this would be a substantial undertaking. In the worst case, there is no
ConsistencyLevel guarantee higher than ONE even with repair-on-write.
1546-pro: ConsistencyLevel support is built in; support for optional "marker"
allows retry on failure as well. "LocalCounterColumn" approach is less scary
if there are bugs than relying on post-transfer "cleaning" of non-local data.
Supports node IP changes.
1546-con: essentially untested. Some known corner cases involving node
movement need to be addressed. Would require using a custom comparator to
emulate counters in subcolumns.
> Increment counters
> ------------------
>
> Key: CASSANDRA-1072
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1072
> Project: Cassandra
> Issue Type: Sub-task
> Components: Core
> Reporter: Johan Oskarsson
> Assignee: Kelvin Kakugawa
> Attachments: CASSANDRA-1072.112210.patch, CASSANDRA-1072.patch,
> increment_test.py, Partitionedcountersdesigndoc.pdf
>
>
> Break out the increment counters out of CASSANDRA-580. Classes are shared
> between the two features but without the plain version vector code the
> changeset becomes smaller and more manageable.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
