General discussion


OPENVMS File I/O Performance

By bclark ·
anyone from the VMS world out here?
i have a large indexed file with a
corrupt index. i can read/write a new
version of the file and recover the
data - BUT - the runtime exceeds 35
hours. AlphaServer 4100, 3-400 mhz
CPU's, diskspace not the problem.
i need to cut the runtime by a
significant margin.
thanks in advance for any help

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Comments

Collapse -

Have you looked at CONVERT

by Buckshot In reply to OPENVMS File I/O Performa ...

You may want to look at using the CONVERT utility to correct your corrupt index. First try running an ANALYZE/RMS/FDL on the file to create an FDL file. If this is successful, you should have a file with the same name as the indexed file, but a file type of FDL. Then run the convert utility as follows:
$ CONVERT/FDL={filename}.FDL {filename}.DAT -

If this completes without errors, then you can run SEARCH/LOG against both files to search for a bogus string. Check thenumber of lines searched to make sure the numbers match! I have had a corrupt next-record pointer cause the CONVERT utility to skip several thousand records, and the search was the fastest way to check this. If you experience problems with this, aread/write program with lots of error checking is probably your best solution.

Related Discussions

Related Forums