Version management system using plural control fields for synchronizing two versions of files in a multiprocessor system

In a data processing system including a first processing unit storing a first version of a pre-specified plurality of data sets, a second processing unit storing a second version of the pre-specified plurality of data sets, and a facility for transferring data sets between the first and second processing units, a method for synchronizing the first and second versions of the pre-specified plurality of data sets, comprising:

• maintaining a sync-complete control field and a sync-in-progress control field corresponding to each of the first and second versions of each data set in the pre-specified plurality, the sync-complete control field indicating when set that the corresponding version is synchronized and the sync-in-progress control field indicating when set that the corresponding version is in the process of being synchronized;
• clearing the sync-complete control field and the sync-in-progress control field in response to a change in the corresponding version;
• first detecting the sync-complete control field for the first and second versions of the data sets, and if either or both are cleared, then setting the sync-in-progress control field corresponding to a selected version of the first and second versions;
• transferring a copy of the selected version to a temporary file to establish a temporary version;
• second detecting the sync-in-progress control field in the selected version, and if it is set when the temporary version is complete, then setting the sync-complete control field and clearing the sync-in-progress control field corresponding to the selected version;
• setting the sync-complete control field in the temporary version;
• and
• renaming the temporary version to the one of the first and second versions not selected in the first detecting step, and removing the one not selected in the first detecting step.

In a data processing system storing a first version of a pre-specified plurality of data sets and storing a second version of the pre-sepcified plurality of data sets, an apparatus for synchronizing the first and second versions of the pre-specified plurality of data sets, comprising:

• processing means, associated with each of the first and second versions for reading and writing the versions of the pre-specified data sets, including
• means for storing an index node identifying the version with a sync-complete control field and a sync-in-progress control field, the sync-complete control field indicating when set that the identified version is synchronized and the sync-in-progress control field indicating when set that the identified version is in the process of being synchronized;
• means for clearing the sync-complete control field and the sync-in-progress control field in a given index node in response to a write to the version identified by the given index node;
• and
• synchronizing means, in communication with the processing means, for synchronizing the first and second versions, including
• first means for detecting the sync-complete control field for the first and second versions, and if either or both are cleared, then setting the sync-in-progress control field in the index node identifying a selected version of the first and second versions;
• means for transferring a copy of the selected version to a temporary file to establish a temporary version;
• second means for detecting the sync-in-progress control field for the selected version, and if it is set when the temporary version is complete, then setting the sync-complete control field and clearing the sync-in-progress control field in the index node identifying the selected version;
• means for setting the sync-complete control field in the temporary version;
• and
• means for updating the index node for the temporary version to rename the temporary version to the name of the one of the first and second versions not selected by the first means for detecting and removing the version not selected.