cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sam Tunnicliffe (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-9459) SecondaryIndex API redesign
Date Wed, 19 Aug 2015 16:39:47 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-9459?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14703314#comment-14703314
] 

Sam Tunnicliffe commented on CASSANDRA-9459:
--------------------------------------------

I've pushed some further commits which I think make this ready for another round of review
(pinging [~slebresne])

Outstanding issues (some of which may be split out to follow up tickets):

* I'd particularly like some feedback on the way that {{SIM.WriteTimeTransaction#onUpdated}}
and {{Index#updateRow}} handle Memtable updates. Rather than pass the existing, update and
reconciled versions of the row to every registered Index, and have each of them potentially
perform a very similar set of diff operations, the diffing is done once (in the transaction)
and only the deltas are passed to the {{Indexes}} - one row containing the subset of the existing
row that is now gone, another with the subset of the merged row that was added in the operation.
This feels a bit counter-intuitive and writing the javadoc on {{Index#updateRow}} was tough,
so I'm a bit concerned that this is going to be tricky for implementors to work with (but
I don't want to duplicate the diffing if we can avoid it).
* Provide a better API for (re)building indexes. The current approach assumes that indexes
should always be built from the merged view of data in SSTables, but this may not always be
the case. That said, this is true for most existing implementations, and so is optimised to
perform only a single pass through the data. I don't want to prohibit that optimisation, so
some further thought it required.
* Lookups in {{SecondaryIndexManager}} could certainly be improved with some better datastructures,
rather than always resorting to a scan through the entire collection of registered indexes
* There is a mismatch between the name of an index as stored in schema and in the value returned
from Index.getIndexName, which for the builtin index impls is the name of the underlying index
CFS. This leaks into a number of places, notably around (re)building indexes. I've opened
CASSANDRA-10127 for this.
* I'm not entirely happy with the way we validate restrictions using {{Index.supportsExpression}}.
It seems a bit blunt, but I haven't been able to come up with anything better yet.


I've avoided squashing the [wip branch|https://github.com/beobal/cassandra/tree/9459-wip]
since people have already commented on that, but I have had to rebase it several times so
although the commit history has been overwritten, it's remained more or less semantically
consistent.


> SecondaryIndex API redesign
> ---------------------------
>
>                 Key: CASSANDRA-9459
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9459
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: Sam Tunnicliffe
>            Assignee: Sam Tunnicliffe
>             Fix For: 3.0 beta 1
>
>
> For some time now the index subsystem has been a pain point and in large part this is
due to the way that the APIs and principal classes have grown organically over the years.
It would be a good idea to conduct a wholesale review of the area and see if we can come up
with something a bit more coherent.
> A few starting points:
> * There's a lot in AbstractPerColumnSecondaryIndex & its subclasses which could be
pulled up into SecondaryIndexSearcher (note that to an extent, this is done in CASSANDRA-8099).
> * SecondayIndexManager is overly complex and several of its functions should be simplified/re-examined.
The handling of which columns are indexed and index selection on both the read and write paths
are somewhat dense and unintuitive.
> * The SecondaryIndex class hierarchy is rather convoluted and could use some serious
rework.
> There are a number of outstanding tickets which we should be able to roll into this higher
level one as subtasks (but I'll defer doing that until getting into the details of the redesign):
> * CASSANDRA-7771
> * CASSANDRA-8103
> * CASSANDRA-9041
> * CASSANDRA-4458
> * CASSANDRA-8505
> Whilst they're not hard dependencies, I propose that this be done on top of both CASSANDRA-8099
and CASSANDRA-6717. The former largely because the storage engine changes may facilitate a
friendlier index API, but also because of the changes to SIS mentioned above. As for 6717,
the changes to schema tables there will help facilitate CASSANDRA-7771.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message