That is correct. Assign tape drives to the underlying host name and
create a stu that points to that host name.
__
Jonathan Marianu (mah ree ah' nu)
ATT Storage Planning and Design Architect
(360) 597-6896
Work Hours 0800-1800 PST M-F
Manager: Jim
Nathan,
Are you using RMAN to back up the database.
There are many settings available to optimize RMAN-nbu disk to disk
backup.
Consider using a vault profile or running a script that:
-displays all images created on a client in the last n hours.
-that list is passed as an argument to
I use DSUs for network media servers and either DSUs or VTL for large
clients that need to be dedicated media servers.
Here are some things to consider with network based backups:
Using VTL: A slow network client will hold onto a VTL tape. The other
images on that tape are then not available for
Thanks Curtis,
I generally do not dup while I am backing up, other than images under
128k. (another important discussion)
The backups run at night to DSU, 1900-0600 and the duplication jobs run
during the day 0800-1600.
There are often a handful of network clients that are still running jobs
Paul,
Each NIC on the media server should have a separate IP.
Assign each IP a name in dns
Ensure that the static routes are set up correctly between the media
server and the clients. This part is critical.
This means if you run a traceroute from teh media server to the client
it is sending the
What is the underlying reason for doing this?
Is the goal to improve the network throughput on the media server or do the
clients have an actual need to be on four different vlans?
If you did it that way you would create a virtual media server entry for each
ip.
The real challenge comes
Something you wrote didn't sound quite right.
Bpbkar writes to the child bptm using TCP sockets which is a bottleneck.
The child bptm process or processes, depending on MPX, write to shared
memory, The parent bptm reads from shared memory and writes it to the
tape.
I still use 5.1 so this may
My recollection is that during duplication from VTL to tape , it uses
the mpx originally set in the policy unless it is throttled down by the
vault policy but you can't increase it. The MPX is what is so
interesting to examine in truss because I observed that any multiplexing
will impact the
Part 1 of 2
Matt,
This applies to 5.1 and below.
Also, triple check all advice because I could be wrong. I don't have
access to production NBU systems so this is from memory and my notes.
Also I am writing without providing a full explanation. So I will
include some notes at the bottom
:[EMAIL PROTECTED]
Sent: Wednesday, November 14, 2007 20:22 PM
To: Marianu, Jonathan
Subject: Re: [Veritas-bu] Any command line gurus around?
Thanks so much... I will be documenting this in our wiki, with a plan
to publish it on the internet.
One question, it seems that by running this I add
Brian,
This link may address your question.
http://seer.support.veritas.com/docs/275608.htm
Feel free to ask any command line question. Before moving into
architecture I managed the netbackup environments and did so exclusively
by command line.
I have not used the 6.0+ version of NBU yet but I
I'd like to obtain a list of the file systems on our unix clients for
reporting. Bpcoverage uses some type of mechanism to obtain this list.
Can anyone provide further insight as to how bpcoverage obtains a list
of the remote client's file systems?
I have run strings against the binary and
Yes, I know of the bpmount command. I mentioned it in my original post.
I have considered putting the bpmount command in a bpstart_notify script
and then using bpgp to receive a copy of the output.
The challenge to that is deploying this to the vast number of clients.
I could use bpgp to
An alternate solution dawned on me.
bpgp the /etc/mnttab
This provides the list of the host file systems.
Thanks everyone
__
Jonathan Marianu (mah ree ah' nu)
ATT Storage Planning and Design Architect
(360) 597-6896
Work Hours 0800-1800 PST M-F
Here is a possible answer to your problem.
I have a script that combines these commands into a more condense
version but here I broke it into several logical commands to make it
easier to understand.
This command provides a list of the TLD robots defined on the control
host
tpconfig -d | sed
We have a separate policy for each database instance.
__
Jonathan Marianu (mah ree ah' nu)
National Enterprise Backup Restore
ATT/Cingular
(360) 597-6896
Manager: John Menzik
(425) 288-8690
___
Veritas-bu
Two ways to obtain a list of all tape drives logically in use by media
servers across a netbackup environment:
1 vmdareq (not always accurate)
2. bpstulist (to obtain a list of media servers)
vmoprcmd -h $MEDIA_SERVER_NAME -xdraw
The output provides a list of all drives in use.
To
Shashi,
The netbackup process on the client, bpbkar reads the files from the
file system and creates a tar file that it streams over the network to
the bptm process running on a media server. The file is written in a
standard gnu tar format to tape. If needed, you can read the information
from
There are a few things that could cause the status 69.
When it happened to me it was the client name. I was use the name
assigned to the backup interface on the host instead of the name
assigned within exchange.
The rule here is that the name of the client must exactly match the name
embedded in
Yes, if you expire a piece of media then the image copy/fragments on
that media will also expire.
If you don't expire but just change the expiration date on the media
then I *think* the images will all pick up the new date as well.
Try it out and check it with bpimagelist.
You can use bpimagelist or bpimmedia to determine how much data was
written to a particular tape.
___
Remedy information:
Group: National EBR Ops
Scope: Backend Applications
Category: NetBackup
http://wa-gsmrdb02-bth/ESBR/
I'm running BPIMMEDIA in the -l mode on various media ids and
paths and
I'm getting data similar to the following.
IMAGE PBAPAP03 7 PBAPAP03_1162359941 NT_PBAP 13 X_INCR_DAILY 1 1 2109
1163569541 0 0
FRAG 1 1 108288 0 2 13 24 IN3683 pbbkup03 65536 937224 1162294541 3 0
*NULL*
BPPLSCHED COLUMNS
Line 1
1 SCHED
2 SCHEDULE_NAME
3 TYPE (0=FULL, 1=INCR, 2=USRD)
4 MPX
5 FREQ (Run every ___ seconds, 86400=1 day) NOTE: Even on FULLS set it
to run daily but restrict the time to the particular day.
6 retention level
7 uwind/o/d (Future use)
8 internal attribute
9 internal
Reset JobIDs
Stop the daemons and either echo a new number or cat /dev/null to
/usr/openv/netbackup/db/jobs/jobid
___
Remedy information:
Group: National EBR Ops
Scope: Backend Applications
Category: NetBackup
http://wa-gsmrdb02-bth/ESBR/
1. Tape went down due to room temperature (Asif Iqbal)
Asif,
You can remotely up a drive but you first want to ensure that there is
no tape in the drive or else it will go down again, and cause a backup
falure.
From the host connected to the library robot use robtest, pick the
library and
David McWilliams,
Please ensure the oracle user has permission to read the bp.conf and
write to logs/dbclient
JM
___
Remedy information:
Group: National EBR Ops
Scope: Backend Applications
Category: NetBackup
http://wa-gsmrdb02-bth/ESBR/
The use of any multiplexing has an adverse effect on duplication
throughput.
To maximize duplication performance either turn MPX Off or set it to a
high number (as much as 32).
Setting MPX to 1 (off) provides the best duplication performance and
restore performance but if there are not enough
27 matches
Mail list logo