Re: Can the VMS DELETE statement "pend"?

From: David M. Desroches <dmd_at_wpi.WPI.EDU>
Date: 9 Jun 93 09:43:21
Message-ID: <DMD.93Jun9094321_at_wpi.WPI.EDU>


> In article Chris Little writes:
 

> This is an Oracle and a VMS related issue, so I'm posting to both
> groups.
>
> I have a DCL script which removes the file DBPROD.DMP from a disk
> and begins an Oracle export which generates the file again. I have
> to delete the file first because it occupies 90% of the disk. I run
> this script every morning.
>
> This weekend I was doing a database reorganisation, which involved
> the export, recreation of the database, and then a full import from
> the file DBPROD.DMP. The import took much longer than anticipated,
> so the next day's export began before it finished.
>
> I suspected the attempted deletion of the file DBPROD.DMP would have
> failed because it was still locked by the import. I also suspected
> the export would generate a new version of the file and would
> eventually abort due to insufficient space on the device.
>
> The actual outcome was as follows:
>
> The import continued and completed successfully. The export started
> but aborted due to insufficient space on the device.
>
> And the interesting thing:
>
> There was only one DBPROD.DMP left on the device and it was the one
> generated by the export which aborted. The file used by the import
> was gone, yet the import succeeded.
>
> Incidentally the import completed three hours after the export
> started. The export aborted after one hour. I have no idea how
> this happened; I never thought a VMS DELETE statement would pend.
> Is that possible?

Were you using the /LOG qualifier in your delete to ensure that the file was indeed deleted? Any error messages during the execution of the DCL script? Also, is your file version limit set to 1?

Without more information, there is not much we can do to help you.

--
**************************************************************
* David M. Desroches       * College Computing Center        *
* dmdesroches_at_jake.wpi.edu * Worcester Polytechnic Institute *
**************************************************************
Received on Wed Jun 09 1993 - 09:43:21 CEST

Original text of this message