We use the 'shadow mirror' process on a Hitachi SAN (with
Oracle 8.1.7 on Solaris 8 and Veritas VxFS) successfully to refresh
our development db without a suspend, or even hot backup mode.
After the hair that I didn't pull out turned gray (though
I am not willing to detail the incompetence on a semi-public
list), we got the hardware/file system set up correctly and
we have done this several times with no trouble.
One issue I had with Hitachi support was that they insisted I
had to suspend the database at the time of the split when we
were seeing data corruption that was clearly unrelated to Oracle
behavior. It turned out to be low level misconfiguration, I believe
at the Veritas file system level. I never got a satisfactory
explanation.
Here is my current (possibly inaccurate) understanding of the
process. I would be more than happy to be corrected on any of
the details.
The shadow mirror process is 'atomic' in the sense that there is
some kind of journaling so that when the mirror is split, it is
done so that the shadow is a copy of the disk at a particular
time, so it looks like the disk would look after a crash or a
shutdown abort. Oracle crash recovery has worked as advertised,
which is sufficient for our development db needs. If we do get a
bad mirror, we would be able to resync and resplit quickly. But
as I said, we've done this without incident at least 8 or 9
times. The only times we've had to resync and resplit were due
to human error.
If you are using Shadow mirror, and this is for backup purposes,
you may want the extra security of hot backup. But the mirror
split should not cause split blocks, so I'm not sure that it
would actually do much for you. To be honest, I haven't thought
through all the ramifications of using this as a backup method.
If you're not using Shadow mirror, but some other mirror method,
the above may not apply.
Hope this helps.
-Chris
-Original Message-
From: Hemant K Chitale [mailto:[EMAIL PROTECTED]
Sent: Thursday, March 27, 2003 9:44 AM
To: Multiple recipients of list ORACLE-L
Subject: may not be necessary -- was RE: Oracle DB Backups on SAN with
Jeremiah / Deborah,
My understanding is/was that the Snapshot creation wasn't
atomic -- it can
take a little bit of time
and, therefore, it becomes necessary to suspend I/O.
Now, I haven't had a chance yet to speak to the Sun/Hitachi
engineers and I
am going
by what management has understood and conveyed to me -- that
the database must
be quiesced. Hopefully, next week, I will be allowed to
speak to the
engineers before
they set up the SAN.
Reading Oracle's documentation in the Backup and Recovery guide
Using the Oracle8i SUSPEND/RESUME functionality, you can
suspend I/O to
the database, then split the mirror and make a backup of the
split mirror.
This feature, which complements the hot backup functionality,
allows you to
quiesce the database so that no new I/O can be performed. You
can then
access the suspended database to make backups without I/O
interference.
Note: Some RAID devices benefit from suspending writes while
the split
operation is occurring; your RAID vendor can advise you on
whether your
system would benefit from this feature.
and
After a successful database suspension, you can back up the
database to
disk or break the mirrors. Because suspending a database does
not guarantee
immediate termination of I/O, Oracle recommends that you precede the
SUSPEND statement with a BEGIN BACKUP statement to place the
tablespaces in
hot backup mode.
You must use conventional operating system backup methods to
back up split
mirrors. RMAN cannot make database backups or copies because these
operations require reading the datafile headers. After the
database backup
is finished or the mirrors are re-silvered, then you can
resume normal
database operations using the RESUME statement.
Backing up a suspended database without splitting mirrors can
cause an
extended database outage because the database is inaccessible
during this
time. If backups are taken by splitting mirrors, however,
then the outage
is nominal. The outage time depends on the size of cache to
flush, the
number of datafiles, and the time required to break the mirror
I did get the impression that a SUSPEND was necessary.
However, I have read a Hitachi document at
http://www.hds.com/pdf/ods.pdf
and I think that a SUSPEND is not mandatory.
I will come back to the list when I get more information from
the Sun/Hitachi
engineers and see the scripts/script-templates that they will
be providing.
Hemant
LEGAL NOTICE:
Unless expressly stated otherwise, this message is confidential and may be privileged.
It is intended for the addressee(s) only. Access to this e-mail by anyone else is
unauthorized. If you are not an addressee, any disclosure or copying of the contents
or any action taken (or not taken) in reliance