On Sunday 16 February 2003 11:07, nawir wrote:
> Hi steven,
>
> What causing your /opt/openv so big?

My catalog! I back up an infrastructure in the neighborhood of 70 Terabytes, 
some of which I retain for 1 year. Plus, I log everything!

> What I know you backup your catalog to tape, so in my theory that means no
> increasing in your /opt/openv.

Nope. Backing up my catalog doesn't affect the size of the catalog itself. 
(except the entries for the catalog backup itself)

> Do you think even tough if I only specify to backup to tape that means my
> /opt/openv also will increase tremendously.
> If yes, that means not enough space if I only have 1gb root space left.
>

No. /opt/openv will not be affected by backing up your catalog to tape.  See 
above.

> here are my current setup.
> SunFire 6800 with L9 connected to it
> I am using Business Server, both Master/Media server is installed in the
> domain A. Domain A also client of Master/Media server.
>
> If space will be an issue and I can't add more disk to SF6800 because of
> scsi hd price issue. Can I change the design to be like this.
> I will move Master/Media server to a SunBlade 100 that will have let say
> 200GB IDE HD. And setup that SunFire 6800 only as client of SunBlade 100.
> Will SunBlade 100 can handle backup load of max 160gb/day?

I don't think a SunBlade is going to be enough to do what you want to do. 

Disk storage is only one consideration when sizing a server. Especially in the 
case of your media server, you want to make sure that you have sufficient 
system resources to push your tape drives. On both, I/O is something to 
consider, as well.

This conversation is regarding disaster recovery, right? The people making the 
$$$ decisions must understand what the money they are willing to spend can 
buy. We can only do so much on a limited budget. 

Conversations with management are always funny, regarding DR. You ask them 
what they want, and more often then not, it's 100% uptime, HA services 
infrastructure-wide. When they get the price tag on that, then the real 
discussions begin. But that's a whole other topic.

IMHO, adding additional SCSI drives is pretty cheap, in the grand scheme of 
things, assuming that you have the drive bays to spare.




> thanks
> ----- Original Message -----
> From: "Steven L. Sesar" <ssesar AT rcn DOT com>
> To: "nawir" <nawir AT esuria.com DOT bn>; "Prabhala, Rama"
> <RPrabhala AT forestpharm DOT com>; <veritas-bu AT mailman.eng.auburn DOT edu>
> Sent: Sunday, February 16, 2003 11:22 PM
> Subject: Re: [Veritas-bu] Solaris Master server recovery using veritas
>
> > On Sunday 16 February 2003 09:42, nawir wrote:
> > > Steven your solution is very good.
> > > I want to know how big hard disk space I need to prepare if I want to
>
> keep
>
> > > a catalog for 90 tapes of DLT 80GB max compressed with 3 month
> > > retention period.
> > > My current setting is to store catalog to tape instead of root hd
>
> because
>
> > > my root hd only 1gb space left.
> > > Can I still use rsync method since the database will going to tape not
>
> to
>
> > > /usr/openv.
> > > If can't, can I set alternate catalog save to a NFS mount point.
> > > If yes, how does it work. Will it save catalog in both tape and nfs or
>
> it
>
> > > will save 1st catalog into tape and 2nd catalog into nfs and so on
> > >
> > > thanks
> >
> > A couple of things:
> >
> > Where you back up your catalog, is irrelevant, in my solution. I back
> > mine
>
> up
>
> > to tape, as well. I don't think that you would want to back your catalog
>
> up
>
> > to the same disk that it resides on, anyway.
> >
> > In terms of disk space, it's not how many tapes/tape capacity that drives
>
> this
>
> > decision. In my system, /opt/openv is ~45G, and sits on a ~100G
> > partition.
>
> On
>
> > my master server, what I have done is concatenated 3 - 36G drives, and
> > mounted the resultant space on /opt/openv.
> >
> > So, I did the same thing on a media server, which lives in a different
> > building. Concatted 3 - 36G drives, but over there, I mount it on
> > /opt/backup. rsync mirrors master_server:/opt/openv to
> > media_server:/opt/backup.
> >
> > As you see, the entire NBU directory structure is getting mirrored via
>
> rsync.
>
> > Catalogs, and all.
> >
> > Let me know if you need more information.
> > --Steve
> >
> > > ----- Original Message -----
> > > From: "Steven L. Sesar" <ssesar AT rcn DOT com>
> > > To: "Prabhala, Rama" <RPrabhala AT forestpharm DOT com>;
> > > <veritas-bu AT mailman.eng.auburn DOT edu>
> > > Sent: Saturday, February 15, 2003 6:36 AM
> > > Subject: Re: [Veritas-bu] Solaris Master server recovery using veritas
> > >
> > > > Here's what I do:
> > > >
> > > > I have an rsync job running on my master server, which mirrors
>
> /opt/openv
>
> > > to a
> > >
> > > > drive in one of our media servers. It mirrors to /opt/backup
> > > > remotely.
> > > >
> > > > In the event we need to mobilize our DR plan, here's the procedure
> > > > (tested more than once):
> > > >
> > > > 1. kill NBU processes on media server.
> > > > 2. move /opt/openv /export/openv.ORIG
> > > > 3. move /opt/backup to /opt/openv
> > > > 4. edit crontab appropriately.
> > > > 3. add FORCE_RESTORE_MEDIA_SERVER <master> <media> to my bp.conf.
> > > > This
>
> is
>
> > > > necessary because we have a media server running on our master
> > > > server. 4. change hostname of media server to that of master server.
> > > > 5. reboot
> > > > 6. reconfigure robotics/drives
> > > > 7. The system is ready for restores/backups.
> > > >
> > > > This has proven to be a whole lot quicker than building a new box,
> > >
> > > installing
> > >
> > > > NBU, and recovering the catalog from tape.
> > > >
> > > > NOTE: This assumes SCSI-attached robotics/drives. If your drives are
> > >
> > > connected
> > >
> > > > to a SAN, this may not be enough....
> > > >
> > > > My next plan, is to install a second GIG interface on this media
>
> server,
>
> > > which
> > >
> > > > will simplify things, even further!
> > > >
> > > > HTH,
> > > >
> > > > Steve
> > > >
> > > > On Friday 14 February 2003 13:55, Prabhala, Rama wrote:
> > > > > Hi there gurus,
> > > > >
> > > > > I have a solaris master server which is also the backup server. I
> > > > > am preparing a disater recovery procedure for the server. I went
>
> through
>
> > > the
> > >
> > > > > veritas handbook and found something on it but i thought it was a
> > > > > little confusing.
> > > > >
> > > > > Does any one of you have  aprocedure for recovering a solaris
>
> server.
>
> > > Any
> > >
> > > > > links , books will do.
> > > > >
> > > > > Regards,
> > > > >
> > > > > Rama
> > > > > 493-7058
> > > >
> > > > _______________________________________________
> > > > Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> > > > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
> > >
> > > _______________________________________________
> > > Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> > > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
> >
> > _______________________________________________
> > Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu