[
https://issues.apache.org/jira/browse/CASSANDRA-8897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14516345#comment-14516345
]
Stefania commented on CASSANDRA-8897:
-------------------------------------
bq. Hmm. RandomAccessReader the only instances of this? If so could we just
grab our buffer in reBuffer()?
RAR and CRAR. However this would not help the 'greedy' iterators that access
the first value in the constructor, like the merge iterator. It also leaves us
with a null reference in two classes which are somewhat big.
It turns out the biggest offender is building the secondary indexes, after I
changed {{KeyIterator}} and {{ReducingKeyIterator}} the errors disappeared from
the dtests.
The logging is back at ERROR level at the moment, so that I could spot any
occurrences more easily. I don't mind removing it or lowering to DEBUG however.
Test status:
These two fail on Jenkins but run on my machine:
{code}
consistency_test.TestConsistency.short_read_quorum_delete_test (runs OK on my
box, timed out on Jenkins)
repair_test.TestRepair.simple_sequential_repair_test 3 min 52 sec 1 (runs
OK on my box)
{code}
Remaining failures (both dtest and testall) happen also on trunk.
> Remove FileCacheService, instead pooling the buffers
> ----------------------------------------------------
>
> Key: CASSANDRA-8897
> URL: https://issues.apache.org/jira/browse/CASSANDRA-8897
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: Benedict
> Assignee: Stefania
> Fix For: 3.0
>
>
> After CASSANDRA-8893, a RAR will be a very lightweight object and will not
> need caching, so we can eliminate this cache entirely. Instead we should have
> a pool of buffers that are page-aligned.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
