[
https://issues.apache.org/jira/browse/HBASE-6651?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hiroshi Ikeda updated HBASE-6651:
---------------------------------
Attachment: HBASE-6651-V14-trunk.patch
HBASE-6651-V14-0.98.patch
HBASE-6651-V14-0.96.patch
Added fixed patches for 0.96, 0.98 and trunk.
I reduced changing of existing codes, which have been quickly changed and too
difficult to aim. I also added new methods to ShardMap to refer its registered
objects, which introduced possibility to invalidly access thread unsafe
objects, but I compromised because of applicability to existing codes, and then
I excluded some unused classes that I had created.
This ticket was created about more than a year ago and I already forgot many
things, but HBase still uses PoolMap that violates the contract of Map,
inconsistently changes its methods' behaviors by pool-type, and fails to keep
thread-safety, so I hope the patches help.
> Improve thread safety of HTablePool
> -----------------------------------
>
> Key: HBASE-6651
> URL: https://issues.apache.org/jira/browse/HBASE-6651
> Project: HBase
> Issue Type: Bug
> Components: Client
> Affects Versions: 0.94.1
> Reporter: Hiroshi Ikeda
> Assignee: Hiroshi Ikeda
> Attachments: HBASE-6651-V10.patch, HBASE-6651-V11.patch,
> HBASE-6651-V12.patch, HBASE-6651-V13.patch, HBASE-6651-V14-0.96.patch,
> HBASE-6651-V14-0.98.patch, HBASE-6651-V14-trunk.patch, HBASE-6651-V2.patch,
> HBASE-6651-V3.patch, HBASE-6651-V4.patch, HBASE-6651-V5.patch,
> HBASE-6651-V6.patch, HBASE-6651-V7.patch, HBASE-6651-V8.patch,
> HBASE-6651-V9.patch, HBASE-6651.patch, sample.zip, sample.zip,
> sharedmap_for_hbaseclient.zip
>
>
> There are some operations in HTablePool accessing PoolMap in multiple places
> without any explicit synchronization.
> For example HTablePool.closeTablePool() calls PoolMap.values(), and calls
> PoolMap.remove(). If other threads add new instances to the pool in the
> middle of the calls, the newly added instances might be dropped.
> (HTablePool.closeTablePool() also has another problem that calling it by
> multiple threads causes accessing HTable by multiple threads.)
> Moreover, PoolMap is not thread safe for the same reason.
> For example PoolMap.put() calles ConcurrentMap.get() and calles
> ConcurrentMap.put(). If other threads add a new instance to the concurent map
> in the middle of the calls, the new instance might be dropped.
> And also implementations of Pool have the same problems.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
