ADSM-L

Querying API files

1996-10-17 15:51:57
Subject: Querying API files
From: Tom Denier <tom AT STAFF.UDC.UPENN DOT EDU>
Date: Thu, 17 Oct 1996 15:51:57 -0400
My site uses SQL-BackTrack with an API interface to ADSM on a number of
client nodes to back up Oracle databases. Our current disaster recovery
procedures send active backup files offsite. I am trying to find out
whether all the files needed to recover the Oracle databases are considered
active backups by ADSM. I attempted to do this by starting dsmc on one
of the clients and running 'query backup' commands. I initially got
responses indicating that no file matched my search criteria, partly
because the SQL-BackTrack documentation is unclear about the owner
USERID of the files it sends to the server, and partly because the ADSM
documentation is unclear about the semantics of the file specification
operand of the query backup command. I finally got a query that actually
found some files, and it failed with an 'ANS4245E Format unknown'
message. According to the messages manual, this message occurs only
when restoring backup files or retrieving archive files. ADSM Level
I has acknowledged that the message is also generated when a normal
ADSM client is used to query backup files stored by an API client.
I was using a Version 1 server and client, but ADSM Level 1 informed
me that the same restriction applies to Version 2. I don't understand
the reason for this restriction. All of the information the query is
supposed to return, with the possible exception of the file size, is
needed by the ADSM server to do its file management. This implies that
the server is capable of providing the information regardless of the
type of client used to create the backup files. ADSM Level I advised
me to contact the supplier of the API application. One of my co-workers
is working on this, but I am not optimistic. As I noted above, the
vendor doesn't even tell us clearly what owner they specify, and I
know that repeated attempts to get the vendor to tell us something
useful about the names of the files created have been completely
unsuccessful. Can anyone suggest any other approach to determining the
names of the files the API application produces and which of them are
active backups?
<Prev in Thread] Current Thread [Next in Thread>
  • Querying API files, Tom Denier <=