saw this in a oracle-l post.
From Mr Dave . Herring
In working with Oracle on this issue, they had me recreate my controlfiles, then add a second channel allocation as part of the DELETE OBSOLETE, this one to disk.
All backups perform the following as their last step:
SQL 'ALTER SESSION SET tracefile_identifier=CONTROLFILE';
SQL 'ALTER DATABASE BACKUP CONTROLFILE TO trace';
... and for some reason not being able to automatically remove obsolete versions of this controlfile backup was part of the problem
follows the situation
> I've got a regular job to delete obsolete pieces from our RMAN catalog,
> which now is failing with "Memory fault". The system is Tru64 5.1b
> running Oracle 9.2.0.6 and the RMAN catalog is 10.2.0.2. Backups are to
> tape using Legato Networker 7.1.1.
>
> The procedure is very simple, just connecting to the target and catalog,
> allocating a channel and deleting:
>
> CONNECT CATALOG ../..@..
> CONNECT TARGET ../..@..
> ALLOCATE CHANNEL FOR MAINTENANCE DEVICE TYPE sbt_tape;
> DELETE NOPROMPT OBSOLETE;
>
> The ONLY thing I get back is:
>
> RMAN retention policy will be applied to the command
> RMAN retention policy is set to redundancy 4
> Memory fault
>
> No Oracle errors show up in either database. I can't find messages or
> info anywhere as to why this is happening. I tried setting my retention
> to larger numbers, thinking there's too many pieces to delete, but that
> didn't work. Current retention is 4 copies. I also tried using RECOVERY
> WINDOW TO xxx DAYS, which failed too. I even truss'ed the session and it
> ends with:
>
> ...
> write(10, "\0BD\0\006\0\0\0\0\003 ^".., 189) = 189
> read(10, "01 m\0\006\0\0\0\0\00701".., 2064) = 365
> write(10, "\0BD\0\006\0\0\0\0\003 ^".., 189) = 189
> read(10, "01 t\0\006\0\0\0\0\00701".., 2064) = 372
> Incurred fault #32, FLTBOUNDS %pc = 0x000000012105DD84 addr =
> 0x000000011FFF4D00
> Received signal #11, SIGSEGV [default]
> siginfo: SIGSEGV SEGV_MAPERR addr=0x0000000000000058
> Err#11 Error 11 occurred.
> *** process killed ***
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment