To what extent does Sun NFS deviate from one-copy file update semantics? Construct a scenario in which two user-level processes sharing a file would operate correctly in a single UNIX host but would observe inconsistencies when running in different hosts.

Question

What will be an ideal response?

Answer 1

After a write operation, the corresponding data cached in clients other than the one performing the write is invalid (since it does not reflect the current contents of the file on the NFS server), but the other clients will not discover the discrepancy and may use the stale data for a period of up to 3 seconds (the time for which cached blocks are assumed to be fresh). For directories, the corresponding period is 30 seconds, but the consequences are less serious because the only operations on directories are to insert and delete file names.
Scenario: any programs that depend on the use of file data for synchronization would have problems. For example, program A checks and sets two locks in a set of lock bits stored at the beginning of a file, protecting records within the file. Then program A updates the two locked records. One second later program B reads the same locks from its cache, finds them unset, sets them and updates the same records. The resulting values of the two records are undefined.