A:
|
The RUJ filenames
were *supposed* to be unique in Rdb and DBMS...but the algorithms
used to create the RUJ files names did not always guarantee that uniqueness.
In previous versions of Rdb and DBMS, the RUJ file names were
“made unique” by adding a date/time stamp to the filename.
As processors became faster, this was not always sufficient
to guarantee uniqueness and could have resulted in file access
conflicts. The next attempt
at uniqueness involved generating RUJ filenames based on the root
file name, the monitor ID and a unique “bind handle.”
However, long root file names would cause the RUJ file names
to be truncated…again, resulting in possible RUJ file access
conflicts. Most recently, the
algorithm was changed to truncate the database name portion of the
root file rather than the monitor-id/bind handle. Hopefully, this
has finally solved the “unique RUJ file name” issue.
In any case, SCI recommends that you do not place version
limits on your RUJ directories, and avoid purging RUJ files.
Because the duplicate RUJ file you purge may just be the one you
need! |