lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yonik Seeley (JIRA)" <j...@apache.org>
Subject [jira] Updated: (LUCENE-2779) Use ConcurrentHashMap in RAMDirectory
Date Mon, 29 Nov 2010 16:05:11 GMT

     [ https://issues.apache.org/jira/browse/LUCENE-2779?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Yonik Seeley updated LUCENE-2779:
---------------------------------

    Attachment: TestCHM.java

Note: there is an important change of semantics to listAll() - it can now "lie".
Previously, it always gave an accurate listing of files as they existed at some point in time.
 Now, you can get back a list of files that never really existed together at any point in
time (i.e. the lie).

I'm OK with these new semantics because we have the same limitation already on FSDirectory.
 This issue manifested in LUCENE-2585, so we should keep in mind that we could now also see
this with RAMDirectory.

Here's a test program that demonstrates the issue by having a single writer adding 5,-5 to
a map before removing the old 4,-4, etc.  At any point in time, there will always be a pair
of numbers.  The single reader looks for this and spits out an error if not found.

output from my box:
iterations=1000000 errors=4120 minLen=0 maxLen=10

> Use ConcurrentHashMap in RAMDirectory
> -------------------------------------
>
>                 Key: LUCENE-2779
>                 URL: https://issues.apache.org/jira/browse/LUCENE-2779
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Store
>            Reporter: Shai Erera
>            Assignee: Shai Erera
>            Priority: Minor
>             Fix For: 3.1, 4.0
>
>         Attachments: LUCENE-2779.patch, LUCENE-2779.patch, TestCHM.java
>
>
> RAMDirectory synchronizes on its instance in many places to protect access to map of
RAMFiles, in addition to updating the sizeInBytes member. In many places the sync is done
for 'read' purposes, while only in few places we need 'write' access. This looks like a perfect
use case for ConcurrentHashMap
> Also, syncing around sizeInBytes is unnecessary IMO, since it's an AtomicLong ...
> I'll post a patch shortly.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message