We did some testing today and Legato did as I expected (and hoped) it
would do.
Oracle (Full backup to DiskPool2), Level 1 to DiskPool2. Staged it all
to tape (orapool2).
Another Level 2 to DiskPool2.
Corrupted the DB and the restore did just (from both the Tape and Disk).
Another question is if the DBA had previously done a backup to orapool1
(tape) and then we switched to Disk/Staging. I'm assuming that one
restore with RMAN will need to go and get the data from orapool2 and
another step will be to get the data from DiskPool2 (from where ever it
now sits).
Also, our restores from tape/disk this afternoon didn't show much
improvement with disk - 460MB (1:09 (tape) versus 00:57 (disk)). I' told
that there should be some performance enhancements with 7.1-2 over
7.1-1. Has this been established by anyone?
Thanks,
Randy
> -----Original Message-----
> From: Legato NetWorker discussion
[mailto:NETWORKER AT LISTMAIL.TEMPLE DOT EDU]
> On Behalf Of Stephen Carter
> Sent: Wednesday, August 18, 2004 4:16 AM
> To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
> Subject: Re: [Networker] Adv_File + Staging + RMAN
>
> On Tue, 17 Aug 2004 13:04:00 -0400, Randy Doering
<rdoering AT COMCAST DOT NET>
> wrote:
>
> >Greetings,
> >
> > Our environment: Server - Tru64 Unix 5.1B (PK2) NW 7.1-1.
> > Client - Linux AS 3.0, Oracle 9i RAC Linux
Client
> >7.1-1.
> >
> > I'm in the midst of implementing Adv_File Devices and Staging.
> >
> > I think Networker has it all figured out if I'm doing file
system
> >backups using the Linux Client S/W because I tell Legato which Pool
to
> use
> >with the Group/Pool relationship.
> >
> > However, when the Oracle backups are kicked off by RMAN, our
DBA
> >needs to specify which Pool to go to. So if the following scenario
> happens,
> >would we be able to recover?
> >
> > Full Backups are done on Sunday at 05:00, Incr (Wednesday) and
> >another Incr on Friday.
> >
> > So, on Sunday the full was done to the "orapool2". On Tuesday,
I
> had
> >him point the backups to "DiskPool2", that sends the Incr to the
> DiskPool2
> >Adv_File device on Wednesday and Friday. On Friday afternoon we have
a DB
> >Crash and need to do a Full Recovery.
> >
> > I think the answer to this question is to put the Full + All
> >Incrementals out to the DiskPool2 device. When it comes time to
Stage,
> I'll
> >Stage from DiskPool2 to orapool2 (tape) and Legato will know to go to
> tape
> >for the data in the case of recovery. Is this correct?
> >
> > However, another question from the DBA is: What If I want to
> recover
> >to a Point In Time of say 2 weeks ago (when we were sending things
> directly
> >to tape)?
> >
> > Thanks for any advice - I use this Listmailer frequently and
> >appreciate the feedback.
> >
> >Randy
> >
> >--
> >Note: To sign off this list, send a "signoff networker" command via
email
> >to listserv AT listmail.temple DOT edu or visit the list's Web site at
> >http://listmail.temple.edu/archives/networker.html where you can
> >also view and post messages to the list.
> >=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
>
> Hi Randy,
>
> Do you use the Networker Module for Oracle to backup your DB's? I'll
> assume
> that you do based on the fact that the DBA has to change the pool name
in
> the RMAN scripts each time the backup level changes. What you should
look
> at
> doing is to create a second instance (instances...as you're running
RAC)of
> the client in Networker for the Archive Redo Log backups (Incr) and
place
> it
> in a different group with a seperate pool specifically for ARL
backups.
> This
> client instance will have a different RMAN script that specifies this
new
> pool and level (ARL).
>
> What also needs to be understood is that even though the logs are
being
> sent
> to disk, Legato still tracks all savesets in the media db and knows
what
> goes where and where it came from. So when it comes time to recover,
> Networker will know that in order for a point in time restore to be
> achieved
> it will require the latest full (dependant on recovery point
objective)
> and
> dependant level backups or in this case logs. It doesn't matter where
the
> savesets are located (on disk or tape) as long as they are available
and
> browseable.
>
> On the staging question, ideally you would stage to a clone pool but
> Networker will allow you to stage to the pool that holds the full
backups.
> I
> would recommend looking at allocating a seperate tape pool
specifically
> for
> the staging of the logs.
>
> HTH
> Stephen
>
> --
> Note: To sign off this list, send a "signoff networker" command via
email
> to listserv AT listmail.temple DOT edu or visit the list's Web site at
> http://listmail.temple.edu/archives/networker.html where you can
> also view and post messages to the list.
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
|
