[
https://issues.apache.org/jira/browse/HBASE-10191?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13918407#comment-13918407
]
stack commented on HBASE-10191:
-------------------------------
[~mcorgan]
bq. It's basically creating small in-memory HFiles that can be compacted
several times in memory without going to disk, and holding on to the WAL
entries until they do go to disk.
Pardon dumb questions, "creating small in-memory HFiles..." -- from a small
CSLM that does the sort for us? Or, I remember talking to Martin Thompson once
trying to ask how he'd go about the MemStore 'problem' and I'm sure he didn't
follow what I was on about (I was doing a crappy job explaining I'm sure),, but
other than his usual adage of try everything and measure, he suggested just
trying a sort on the fly... Are you thinking the same Matt? So we'd keep
around Cells and then once we had a batch or if after some nanos had elapsed,
we'd do a merge sort w/ current set of in-memory edits and then put in place
the new sorted 'in-memory-hfile' and up the mvcc read point so it was readable?
Once they got to a certain size we'd do like we do now with snapshot and start
up a new foreground set of edits to merge into?
bq. ...and holding on to the WAL entries until they do go to disk
What you thinking here? Would be good if the WAL system was not related to the
MemStore system (though chatting w/ [~liyin] recently, he had an idea that
would make the WAL sync more 'live' if WAL sync updated mvcc (mvcc and seqid
being tied).
bq. Anoop, Ram, and I were throwing around ideas of making in-memory HFiles out
of memstore snapshots....
Would be sweet if the value at least was not on heap.... Sounds like nice
experiment Andrew.
> Move large arena storage off heap
> ---------------------------------
>
> Key: HBASE-10191
> URL: https://issues.apache.org/jira/browse/HBASE-10191
> Project: HBase
> Issue Type: Umbrella
> Reporter: Andrew Purtell
>
> Even with the improved G1 GC in Java 7, Java processes that want to address
> large regions of memory while also providing low high-percentile latencies
> continue to be challenged. Fundamentally, a Java server process that has high
> data throughput and also tight latency SLAs will be stymied by the fact that
> the JVM does not provide a fully concurrent collector. There is simply not
> enough throughput to copy data during GC under safepoint (all application
> threads suspended) within available time bounds. This is increasingly an
> issue for HBase users operating under dual pressures: 1. tight response SLAs,
> 2. the increasing amount of RAM available in "commodity" server
> configurations, because GC load is roughly proportional to heap size.
> We can address this using parallel strategies. We should talk with the Java
> platform developer community about the possibility of a fully concurrent
> collector appearing in OpenJDK somehow. Set aside the question of if this is
> too little too late, if one becomes available the benefit will be immediate
> though subject to qualification for production, and transparent in terms of
> code changes. However in the meantime we need an answer for Java versions
> already in production. This requires we move the large arena allocations off
> heap, those being the blockcache and memstore. On other JIRAs recently there
> has been related discussion about combining the blockcache and memstore
> (HBASE-9399) and on flushing memstore into blockcache (HBASE-5311), which is
> related work. We should build off heap allocation for memstore and
> blockcache, perhaps a unified pool for both, and plumb through zero copy
> direct access to these allocations (via direct buffers) through the read and
> write I/O paths. This may require the construction of classes that provide
> object views over data contained within direct buffers. This is something
> else we could talk with the Java platform developer community about - it
> could be possible to provide language level object views over off heap
> memory, on heap objects could hold references to objects backed by off heap
> memory but not vice versa, maybe facilitated by new intrinsics in Unsafe.
> Again we need an answer for today also. We should investigate what existing
> libraries may be available in this regard. Key will be avoiding
> marshalling/unmarshalling costs. At most we should be copying primitives out
> of the direct buffers to register or stack locations until finally copying
> data to construct protobuf Messages. A related issue there is HBASE-9794,
> which proposes scatter-gather access to KeyValues when constructing RPC
> messages. We should see how far we can get with that and also zero copy
> construction of protobuf Messages backed by direct buffer allocations. Some
> amount of native code may be required.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
