On performance-limited systems where CPU cycles are strictly rationed, FusionReactor's Administration Manager (FRAM) and installed instances appear to start very slowly.
Two threads which participate in the startup of FusionReactor can in some circumstances begin to compete for CPU resources, if the first thread takes longer than normal to complete.
This occurs commonly in processor-limited environments, Amazon EC2 m1.small or t1.micro instances, for example.
We have confirmed this is not a deadlock – FusionReactor does complete the startup procedure, though this can be radically longer than usual.
We have confirmed this as a bug, and are supplying a hotfix for customers experiencing this issue.
The hotfix should applied only to FusionReactor 4.5.5. Please update if necessary.
Hotfixes are cumulative: any future hotfixes on the FusionReactor 4.5.x stream, up to but not including the next minor point release (4.5.6), will contain this hotfix (and all earlier hotfixes since the last point release).
We may decide, at our discretion, to release a full suite of installers on the 4.5.x maintenance stream at some future point. This will also incorporate hotfixes whose version numbers are the same or less than that of the installer. This means that this hotfix will be included in a potential future installer whose version is 4.5.6 or greater.
This hotfix does not change the master version number (4.5.5) of the software.
|Components:||Plugins + OSGi|
|Last Updated:||21/Aug/14 6:03 PM|
Comments are closed.