[
https://issues.apache.org/jira/browse/JCR-1213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12543985
]
Ard Schrijvers commented on JCR-1213:
-------------------------------------
Follow up:
I have been trying to cache docNumbers with respect to their IndexSegments,
which obviously do change less frequently. Caching based on an entire
CachingMultiReader is trivial, but also, the performance gain is to small,
since the multiReader changes to frequently.
But, I am having difficulties somebody might be able to help me with
Bottom line is, we do not want to cache based on the entire CachingMultiReader,
but on its segments it consists of. Now, I build (hacked for the time being)
something that in DocId.UUIDDocId I keep track of the segment reference through
a WeakReference. But....as I didn't see behavior I was expecting, I found
another difficulty:
The multiReader which is created in MultiIndex is a CachingMultiReader
consisting of ReadOnlyIndexReader's. When something changed in one of the
indexes, a new multiReader is constructed....but instead of reusing the
non-changed ReadOnlyIndexReader instances, every ReadOnlyIndexReader is
re-constructed (not the shared/caching reader they consist of though ) , but,
since the instances our multiReader consists of are recreated, my
WeakReferences based on segment instances are useless.
So, instead of using a WeakReference on the multiReader segments, I could get
the sharedReader instance out of it, but this means casting and adding methods,
something we really do not want of course (and i am not sure if the BitSet
keeing track of deleted might have changed without the sharedReader being
changed ( to be honest, I cannot yet grasp the big picture about keeping track
of the deleted bitset ) ).
So, does anybody have an idea how we might be able to have in
MultiIndex.getIndexReader() only new instances of the ReadOnlyIndexReaders
which actually changed...Or is this not an option? I did try to add it to the
AbstractIndex to check wether there was already an instance of
ReadOnlyIndexReader but then I get AlreadyClosedException in lucene.
I can make another JIRA issue for it if others think it might be valuable and
not part of this issue. WDOT?
> UUIDDocId cache does not work properly because of weakReferences in
> combination with new instance for combined indexreader
> ---------------------------------------------------------------------------------------------------------------------------
>
> Key: JCR-1213
> URL: https://issues.apache.org/jira/browse/JCR-1213
> Project: Jackrabbit
> Issue Type: Improvement
> Components: query
> Affects Versions: 1.3.3
> Reporter: Ard Schrijvers
> Fix For: 1.4
>
>
> Queries that use ChildAxisQuery or DescendantSelfAxisQuery make use of
> getParent() functions to know wether the parents are correct and if the
> result is allowed. The getParent() is called recursively for every hit, and
> can become very expensive. Hence, in DocId.UUIDDocId, the parents are cached.
> Currently, docId.UUIDDocId's are cached by having a WeakRefence to the
> CombinedIndexReader, but, this CombinedIndexReader is recreated all the time,
> implying that a gc() is allowed to remove the 'expensive' cache.
> A much better solution is to not have a weakReference to the
> CombinedIndexReader, but to a reference of each indexreader segment. This
> means, that in getParent(int n) in SearchIndex the return
> return id.getDocumentNumber(this) needs to be replaced by return
> id.getDocumentNumber(subReaders[i]); and something similar in
> CachingMultiReader.
> That is all. Obviously, when a node/property is added/removed/changed, some
> parts of the cached DocId.UUIDDocId will be invalid, but mainly small indexes
> are updated frequently, which obviously are less expensive to recompute.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
