Missing archived logs in v$archived_log
zigzagdna
running in archived log mode.
Yet when I go and look in v$archived_log, I see some archived logs
missing,
i.e, when I do select name from v$archived_log, I see:
arch1_891_XXX.arc
arch1_892_xxx.arc
arch1_895_xxx.arc
I see this happening every two three days,
I do run
alter system archive log current
during my RMAN backups but that should not be the cause.
I have seen this behavior since I started using Oracle 11.1.0.7,
never had this issue
with older releases.
I a not using any data guard. Any ideas will be greatly appreciated.
Mladen Gogala
Are you using RAC? I've had situations like that with RAC databases when
the archive was not written down by the node I expected. In that
situation, every instance had its own log_archive_dest.
zigzagdna
>
> - Show quoted text -
This is non RAC single instance.
bdbafh
what do you see in the alert log regarding the log sequence# 893 being
archived?
-bdbafh
zigzagdna
>
> - Show quoted text -
I looked in alert.log. I do not see anything which indicates that this
log was not archived. I do
see
ORA-19510: failed to set size of blocks for file "" (block size=)
Looking more in this error, it appears that archived logs are getting
deleted while archive switch is being done.
I have no clue who is deleting archived logs. My RMAN backup when they
do archived log backup do
not delete the archived log. Following configuration parameters for
RMAN
CONFIGURE RETENTION POLICY TO REDUNDANCY 30;
CONFIGURE ARCHIVELOG DELETION POLICY TO NONE; # default
Should retain archived logs.
Yet ARCHIVED LOGS are being depleted by some Oracle process. Clearly
there is some bug in Oracle 11.1.0.7.
ddf
>
> - Show quoted text -
Most likely your backup command. When archivelog deletion policy is
set to NONE then all archivelogs are eligible for removal after
backup:
"The archived redo log deletion policy is configured to NONE by
default. In this case, RMAN considers archived redo log files in the
recovery area as eligible for deletion if they meet both of the
following conditions:
The archived redo logs, whether in the flash recovery area or outside
of it, have been transferred to the required remote destinations
specified by LOG_ARCHIVE_DEST_n.
The archived redo logs have been backed up at least once to disk or
SBT or the logs are obsolete according to the backup retention policy.
The backup retention policy considers logs obsolete only if the logs
are not needed by a guaranteed restore point and the logs are not
needed by Oracle Flashback Database. Archived redo logs are needed by
Flashback Database if the logs were created later than
SYSDATE-'DB_FLASHBACK_RETENTION_TARGET'."
You've backed them up, so they're eligible for deletion. I expect
your backup command is deleting all input expecting that your non-
existent deletion policy is 'protecting' them; it isn't. If you do
want to keep them around a while you need to actually configure a
deletion policy:
CONFIGURE ARCHIVELOG DELETION POLICY BACKED UP 2 TIMES TO SBT
for example. Then this first backup would NOT delete the input.
I think the only 'bug' in this situation is you not reading the
manual:
http://download.oracle.com/docs/cd/B28359_01/backup.111/b28270/rcmconfb.htm#BRADV89441
David Fitzjarrell
Sybrand Bakker
>I think the only 'bug' in this situation is you not reading the
>manual:
>
>http://download.oracle.com/docs/cd/B28359_01/backup.111/b28270/rcmconfb.htm#BRADV89441
>
>
So nothing unusual with this particular Oracle user.
--
Sybrand Bakker
Senior Oracle DBA
zigzagdna
>
> - Show quoted text -
Thanks.
My recovery window retention policy is set to 30, so 30 backups have
to be reatined before deletion despite default deletion poilcy of
NONE.
You do not explain why v$archived_log has gaps for archive logs.
Same scripts have been working with Oracle9i and Oracle 10g. There is
clearly undocumented (bug) behavior in Oracle 11g.
ddf
>
> - Show quoted text -
The gaps are probably due to RMAN deleting them; any other process of
deleting archivelogs won't touch V$ARCHIVED_LOG entries as they are
read from the controlfile.
You have searched Metalink (MOS) for such bugs? You can't be the only
person on earth running 11.1.0.7 experiencing such behaviour.
David Fitzjarrell