Bug 51704

Bug: 51704

Impact:

Informational

Product:

GemStone/S 64 Bit

Versions:

3.7.4.3, 3.7.4.1, 3.7.4, 3.7.1, 3.7, 3.6.8, 3.6.7, 3.6.6, 3.6.5, 3.6.4, 3.6.3, 3.6.2, 3.6.1, 3.6, 3.5.8, 3.5.7, 3.5.6, 3.5.5, 3.5.4, 3.5.3, 3.5.2, 3.5.1, 3.5, 3.4.5, 3.4.4, 3.4.3, 3.4.2, 3.4.1, 3.4, 3.3.9, 3.3.8, 3.3.7, 3.3.6, 3.3.x, 3.2.17, 3.2.16, 3.2.15, 3.2.x, 3.1.0.6, 3.1.0.5, 3.1.x, 3.0.x, 2.4.8, 2.4.5.1, 2.4.4.8, 2.4.4.7, 2.4.x, 2.2.5.4, 2.x, 1.x

Platform:

Fixed In:

Inconsistent large object state after a failed commit, before abort

After a commit fails and before an abort, the view of the repository may be in an inconsistent state for large RC objects. This allows examination of the commit failures, but there is a risk if you perform some kinds of operations on any RC collections in the repository, before the next abort, that the data is incorrect and it may report corrupt object errors.

In a large collection, the object's nodes are on multiple pages. Nodes that are not in memory at the time of the commit that failed are faulted in using the view for the RC replay attempt that happened during the commit, and thus include changes from other transactions. Other nodes of the same collection that were in memory are not refreshed; these nodes will not include changes from other transactions.

Workaround

Use caution in executing code and performing operations after a failed commit until you have aborted, especially involving large RC collections.