I see all this exchange about backing up the NetBackup database and the
obnoxiousness of it all and it makes me question what I'm doing because I don't
do it the normal way. Let me run this by you to see if I'm setting myself up
for failure.
My NBU environment consists of two Sun E450's acting as Master servers (no
media servers) running Solaris 2.6 and NBU 3.2GA. Each server has it's own
attached tape library. When I built the NBU environment on each server, I used
Veritas Volume Manager to build a separate file system for the NBU database
directory and the and the NetBackup Volume Manager database directory. I do a
FULL backup each day of the NBU Master servers (after all the backups have
completed) and my thinking is if I had to recover the NBU databases, I could
just restore each file system. I came to that conclusion because the only
directories backed up in the built-in NBDB backup process are the NetBackup and
volume manager database directories. Obviously my operation is fairly small
(about 191 servers total) and it's probably less complicated since I have no
media servers, but I guess I need to know if anyone sees a flaw in my plan. I
have a pool of tapes in each library dedicated to the backup of each NBU master
server and they are located in Slot 1 & 2. Each tape has a different bar code
pattern than all others in the library and I can easily open the door and grab
those two tapes if I had to and import then into a new, bare bones recovered
server for use to recover my NBU master to its previous state. Each master is
backed up to the other and the masters are not located in the same building
(they are 10 miles apart).
What do you think? Am I a catastrophe just waiting to happen or is my plan
viable?
Regards,
Dennis
Quote: "Time is not a test of the truth"
Translation: Just because you've always done it that way, doesn't make it right
Dennis F. Dwyer
Enterprise Storage Manager
Tampa Electric Company
(813) 225-5181 - Voice
(813) 275-3599 - FAX
Visit our corporate website at www.tecoenergy.com
>>> "David A. Chapa" <david AT datastaff DOT com> 01/25/01 09:41AM >>>
Curtis:
A couple of thoughts
1. Why not have a pool of tapes dedicated to NBDB from
the beginning?
2. If you don't like that idea, then for your tape
selection you may want to use the DBBACKUP_CALLED file
to see which tape(s) have been used the last X times.
Using the dbbackup_notify script will cat the
DBBACKUP_CALLED file and included that in an email to
someone_who_cares for tracking the last tape used.
3. I'm not sure why you need to use bpsyncinfo to add
the mediaid for the db backup if you are manually
kicking off the bpbackupdb, unless you are doing it to
ensure the db paths are consistent. But you can get
that using bpsyncinfo and then direct the bpbackupdb to
use a pool other than NetBackup (bpsyncinfo -l and it
is the 14th field and following if I'm not mistaken)
4. You probably have a very good reason for doing this,
but what I do to streamline my selection process of
tapes in the robot is to use the -bx switch with
vmquery then grep for TLD (if it is Tape Library DLT)
for instance. Functionally, it doesn't affect your
script operation and what you are trying to accomplish
(as you well know), but just an FYI. I also use this
extended listing to see if a tape is truly 'scratch' by
testing the time assigned column, if it doesn't eq ---
then I know that it is in use.
Just some thoughts.
david
Quoting "W. Curtis Preston" <curtis AT backupcentral DOT com>:
> No, it works. I just have trouble with all the
things AROUND bysyncinfo:
>
> 1. Finding/getting an UNUSED tape in the NetBackup
pool.
> 2. Protecting the tapes that used to be the NDBD
backup tapes from being
> overwritten.
>
> Here's the script that I am currently using. I've
taken some of the error
> checking and functions out for brevity. I think it's
working in its
> current form. I'm open to suggestions.
>
>
>
########################################################
#####
> # FIRST
SECTION: #
> # INITIALIZE
VARIABLES #
>
########################################################
### #
>
> robot_num=0
> scratch_pool=`grep
SCRATCH_POOL /usr/openv/volmgr/vm.conf |awk '{print
> $3}'`
> pool_num=1
> db_pool_num=2
> VOLTYPE=dlt
>
>
########################################################
#####
> # SECOND
SECTION: #
> # BUILD A LIST ($VOLIDS) of volumes that are in the
scratch #
> # pool and in the robot that we are
using #
>
########################################################
### #
>
> #Identifies all volumes in the robot numbers current
volume database and
> redirects to outfile
> vmquery -b -rn $robot_num | egrep -v '^media|^ *ID|^-
------' |awk '{print
> $1}' \
> >/tmp/$$.tapesinrobot 2>/tmp/$$.errors
>
> #Identifies all volumes that are a part of the
Scratch pool on robot_num
> and assigns to VOLIDS
> VOLIDS=`vmquery -b -pn $scratch_pool | egrep -
v '^media|^ *ID|^-------' \
> |awk '{print $1}'`
>
> #For each VOLID in VOLIDS check to see if the tape is
part of the scratch
> pool
> #and in the robot, if true break
>
> for VOLID in $VOLIDS ;do
> grep $VOLID /tmp/$$.tapesinrobot >/dev/null
> [ $? -eq 0 ] && INVOLIDS="$INVOLIDS $VOLID"
> done
>
>
>
########################################################
######
> # THIRD
SECTION: #
> # For each volume in the list, try to assign it to
the db #
> # backup pool and back up the database. If it's
successfull,#
> # we move on, if not, we try with the next
volume #
>
########################################################
######
>
>
> for VOLID in $INVOLIDS ; do
>
> #Put the volume into the NetBackup pool
> vmchange -p $pool_num -m $VOLID 2>/tmp/
$$.errors
>
> if [ $? -gt 0 ] ; then
> continue
> fi
>
> #Put the volume into the "bpdb pool"
> bpsyncinfo -id1 $VOLID -d1 $VOLTYPE 2>/tmp/
$$.errors
>
> if [ $? -gt 0 ] ; then
> continue
> fi
>
> #I like the logic in the other script about
checking to see if
> #jobs are running. I think I'll add that
here in a while loop
>
> #Back up the database
> bpbackupdb -v 2>/tmp/$$.errors
>
> if [ $? -gt 0 ] ; then
> MYERROR=1
> else
> MYERROR=0
>
> #Take it OUT of the "bpdb pool"
> bpsyncinfo -delete $VOLID
>
> #Put into a special pool
> vmchange -p $db_pool_num -m $VOLID
2>/tmp/$$.errors
>
> break
>
> fi
>
> done
>
> if [ $MYERROR -gt 0 ] ; then
> ERROR="Could not backup NetBackup database
AT ALL!"
> Exit #Exit function
> else
> DATE=`date`
> echo "$DATE: \c"
> echo "Backed up database to volume $VOLID\n"
\
> |tee -a /tmp/$$.log
> echo $VOLID >>/usr/openv/local/data/db-
backup-tapes.txt
> fi
>
>
########################################################
######
> # FOURTH
SECTION: #
> # This takes the oldest tape (out of $NUMBDBTAPES)
and puts #
> # it back into the scratch
pool. #
> # This keeps a set of tapes that contain the most
recent db #
> # backup
tapes. #
>
########################################################
######
>
>
> DBTAPES=`grep -c . /usr/openv/local/data/db-backup-
tapes.txt`
> while [ $DBTAPES -gt "$NUMDBTAPES" ] ; do
> TAPE=`head -1 /usr/openv/local/data/db-
backup-tapes.txt`
> echo "There are over $NUMDBTAPES database
backup tapes: Relabeling
>
> $TAPE." \
> |tee -a /tmp/$$.log
>
> #It's assigned to the database backup, so we
have to deassign it
> vmquery -deassignbyid $TAPE $db_pool_num 0x0
>/dev/null
> 2>/tmp/$$.errors
>
> #Now we put it back in the scratch pool
> vmchange -p $scratch_pool_num -m $TAPE
2>/tmp/$$.errors
>
> #Label it just to make sure all is OK
> bplabel -ev $TAPE -d $VOLTYPE -p
$scratch_pool -o 2>/tmp/$$.errors
>
> grep -v "^$TAPE$" /usr/openv/local/data/db-
backup-tapes.txt
> >/tmp/$$.X
> mv /tmp/$$.X /usr/openv/local/data/db-backup-
tapes.txt
> DBTAPES=`grep -c . /usr/openv/local/data/db-
backup-tapes.txt`
>
> done
>
> echo "\nThis is the current list of database backup
tapes in the " |tee -a
> /tmp/$$.log
> echo "order they were created:" |tee -a /tmp/$$.log
>
> cat /usr/openv/local/data/db-backup-tapes.txt |tee -
a /tmp/$$.log
>
> - ---
> W. Curtis Preston, Principal Consultant at Collective
Technologies
> Email: curtis AT colltech DOT com (Best way
to contact me)
> Work : 408 452 5555 (Leave a
message.)
> Pager: 800 946 4646, pin#1436065 (If urgent.)
>
> Tap into the Collective Intellect (TM):
http://www.colltech.com
> Backup & Restore resources:
http://www.backupcentral.com
>
> -----BEGIN PGP SIGNATURE-----
> Version: PGPfreeware 6.5.3 for non-commercial use
<http://www.pgp.com>
>
>
iQA/AwUBOeIOnn9C1rqc6zS7EQISbACg+CjMAqj9kQFWWxvYYwZquzfZ
eBkAnj/I
> f3PShgyC+ngBlfXbiKTGtT6g
> =+zQd
> -----END PGP SIGNATURE-----
>
> At 04:04 PM 1/24/01 -0800, KevinB AT paccessglobal DOT com
wrote:
> >Are you saying that bpsyncinfo does not work per
it's usage statement?
> >
> >-----Original Message-----
> >From: W. Curtis Preston
[mailto:curtis AT backupcentral DOT com]
> >Sent: Wednesday, January 24, 2001 3:21 PM
> >To: Steve Moccio; david AT datastaff DOT com; Bob Bakh
> >Cc: W. Curtis Preston; veritas-
bu AT mailman.eng.auburn DOT edu
> >Subject: RE: [Veritas-bu] NBDB backup scripts?
> >
> >
> >Thanks for the script.
> >
> >This solves ONE of the problems with the NBDB
backup, but not the biggest
> >one.
> >
> >I've been working on a script to automated swapping
out the NBDB tapes.
> (I
> >really don't like the default setup that alternates
between two tapes.)
> >
> >At 01:18 PM 1/24/01 -0500, Steve Moccio wrote:
> > >Hello all,
> > >
> > >I too was in the advanced class and received a
handout about the script.
> >Not
> > >sure if this is what you are looking for Bob.
> > >
> > >Steve Moccio
> > > Bell Labs
> > > svm AT lucent DOT com
> > >
> > >
> > >
> > >#!/bin/ksh
> > >
> > ># find out how many currently queued, re-queued,
or active jobs
> > ># there are currently
> > >QUEUED=`bpdbjobs - summary | sed -e '/MASTER/d' |
awk '{print $2}'`
> > >REQUEUED=` bpdbjobs - summary | sed -e '/MASTER/d'
| awk '{print $3}'`
> > >ACTIVE=` bpdbjobs - summary | sed -e '/MASTER/d' |
awk '{print $4}'`
> > >
> > ># If there are no queue, re-queued, or active
jobs, then it is ok.
> > ># to initiate netbackup DB backup
> > >
> > >if [ $QUEUED -eq 0] && [ $REQUEUED -eq 0 ] && [
$ACTIVE -eq 0]
> > >then
> > > #Stop the request daemon
> > > echo "Stopping the request Daemon"
> > > bprdreg -terminate
> > > echo "bprd Daemon terminated"
> > > # Run netbackup database backup
> > > echo "Running netbackup database backup"
> > > bpbackupdb
> > > echo "Database backup complete"
> > > # restart request daemon
> > > echo "Restarting request daemon"
> > > bprd
> > >else
> > > echo "Could not initiate netbackup
database backup
> > >fi
> > >
> > >
> > >-----Original Message-----
> > >From: veritas-bu-admin AT Eng.Auburn DOT EDU
> > >[mailto:veritas-bu-admin AT Eng.Auburn DOT EDU]On Behalf
Of David A. Chapa
> > >Sent: Wednesday, January 24, 2001 11:47 AM
> > >To: Bob Bakh
> > >Cc: W. Curtis Preston; veritas-
bu AT mailman.eng.auburn DOT edu
> > >Subject: RE: [Veritas-bu] NBDB backup scripts?
> > >
> > >Bob:
> > >
> > >I teach that class, and I don't remember an NBDB
backup script?
> > >
> > >that's not to say that it can't be done, it most
definitely can.
> > >
> > >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > >David A. Chapa
> > >Consulting Manager
> > >DataStaff, Inc.
> > >847 413 1144
> > >
> > >On Wed, 17 Jan 2001, Bob Bakh wrote:
> > >
> > > > I remember one in the NetBackup advanced class,
if someone has taken
> >that
> > > > class they may have it. I can't find my course
material, but If I
> find
> >it
> > > > I'll forward it to you.
> > > >
> > > > Bob
> > > >
> > > > -----Original Message-----
> > > > From: W. Curtis Preston
[mailto:curtis AT colltech DOT com]
> > > > Sent: Tuesday, January 16, 2001 12:44 PM
> > > > To: veritas-bu AT mailman.eng.auburn DOT edu
> > > > Subject: [Veritas-bu] NBDB backup scripts?
> > > >
> > > >
> > > > Does anyone have what they consider to be a
good NBU database backup
> > >script?
> > > >
> > > > I've tried to write a script to do the
following, but have been
> > > > unsuccessful:
> > > >
> > > > 1. Swap in a new DB backup tape daily
> > > > 2. Protect the OLD DB backup tapes from use
> > > > 3. Backup the database on demand (whenever the
script is run)
> > > >
> > > > Does anyone have a good script for this?
> > > >
> > > >
> > > >
> > > > ---
> > > > W. Curtis Preston, Principal Consultant at
Collective Technologies
> > > > Email: curtis AT colltech DOT com (Best
way to contact me)
> > > > Work : 408 452 5555
(Leave a message.)
> > > > Pager: 800 946 4646, pin#1436065 (If
urgent.)
> > > >
> > > > Tap into the Collective Intellect (TM):
http://www.colltech.com
> > > > Backup & Restore resources:
http://www.backupcentral.com
> > > >
> > > > _______________________________________________
> > > > 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
> >
> >_______________________________________________
> >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
>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
David A. Chapa
Consulting Manager
DataStaff, Inc.
847 413 1144
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
|
