[
https://issues.apache.org/jira/browse/JCR-314?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12525131
]
prios edited comment on JCR-314 at 9/5/07 11:01 AM:
---------------------------------------------------------
It seems that during the time a versioning operation is in progress (writer) it
is not possible to establish a new session to the repository. This seems to be
so because access to the root node of the version histories
(/jcr:system/jcr:versionStorage) is required to create a session and this node
is included on the active writer changelog. In other words the reader need to
access an item (root node of the version storage) that has a dependency on the
active writer changelog.
>From XAVersionManager constructor
try {
state = (NodeState) stateMgr.getItemState(vMgr.getHistoryRootId());
} catch (ItemStateException e) {
...
}
(An XAVersionManager is instantiated during the instantiation of an
XASessionImpl)
Note: I've observed this in an XA environment.
Am I wrong or right ?
What may be the consequences of allowing a reader to get a read lock in this
situation even if the writer's changelog includes the VERSION_STORAGE_NODE_ID ?
was (Author: prios):
It seems that when XA transactions are used the fine grained ISM locking
strategy *always* block readers when there is a writer, because the root node
of the version histories always has a dependency on the active writer changelog.
>From XAVersionManager constructor
try {
state = (NodeState) stateMgr.getItemState(vMgr.getHistoryRootId());
} catch (ItemStateException e) {
...
}
Am I wrong or right ?
> Fine grained locking in SharedItemStateManager
> ----------------------------------------------
>
> Key: JCR-314
> URL: https://issues.apache.org/jira/browse/JCR-314
> Project: Jackrabbit
> Issue Type: Improvement
> Components: core
> Affects Versions: 0.9, 1.0, 1.0.1, 1.1, 1.1.1, 1.2.1, 1.2.2, 1.2.3
> Reporter: Marcel Reutegger
> Attachments: FineGrainedISMLocking-v2.patch,
> FineGrainedISMLocking.patch, ISMLocking.patch, writes-per-second.jpg
>
>
> The SharedItemStateManager (SISM) currently uses a simple read-write lock to
> ensure data consistency. Store operations to the PersistenceManager (PM) are
> effectively serialized.
> We should think about more sophisticated locking to allow concurrent writes
> on the PM.
> One possible approach:
> If a transaction is currently storing data in a PM a second transaction may
> check if the set of changes does not intersect with the first transaction. If
> that is the case it can safely store its data in the PM.
> This fine grained locking must also be respected when reading from the SISM.
> A read request for an item that is currently being stored must be blocked
> until the store is finished.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
