If it is possible, I would run 2 parallel environments, one being your
existing and the other being the new with your linux master. Start
backing up to the new and keep the old around until all of your backup
images expire. I went from an AIX master to Linux which is entirely
doable without a
We have had a chance to side-by-side a 5220 with a DD860 for the past
couple of years and each has its high and low points.
Most of that time we didn't have a backup replication solution in place
and depended on duplicating back to tape off of each device for offsite
storage. The Data Domain i
Is the hostname listed in your DataDomain's access list? I think that is
where the failure is happening.
From: mdglazerman
To: VERITAS-BU@MAILMAN.ENG.AUBURN.EDU
Date: 06/13/2014 03:08 PM
Subject:[Veritas-bu] ddboost ove FC - Questions RE: Storage Server
Credentials
Sent by:
Has anybody else performing this upgrade noticed certain jobs slowing
down? I have particular ones that depend on getting in and out of the
queue quickly (Database agent log rolls) and they hang in the queue for 10
minutes before actually finishing. It would take 10 seconds before
(7.5.0.6).
I have a requirement where a remote division of ours is replicating a
Windows file server to our location via DFS. We are backing up that
server already, but wish to do special backups in order to satisfy some
ISO compliance guidelines. I want to send these copies offisite at their
given inte
I am in complete control of the subnet and attached hosts, so that will
not be a problem. I also see this as a good opportunity to force all of
the clients to use DNS. Always becomes an issue at DR every year.
From: Daniel Otto
To: "scott.geo...@parker.com"
Date: 02/01/2013 09:35
I am moving my master server to new hardware, but am keeping the hostname.
I am also want to change the IP address and naturally have the DNS record
changed as well. I wish to move it to a subnet that utilizes jumbo
frames. Aside of all of the hardcoded clients that will become
immediately k
I encountered one of the "ghost" jobs this morning. It appears that
bpjobd can now be used on an up and running system. The empty jobs can
still show up, you just don't need to bring NBU down to clear them.
From: "Reynolds, Susan K."
To: ,
Date: 11/09/2012 10:31 AM
Subject:
The only way I have been able to get rid of these in 7.1 was to bring NBU
down and run "bpjobd -r " on every one.
I just upgraded to 7.5.0.4 yesterday, and it seems to be much cleaner,
meaning, none of these have shown up yet. I did see some empty jobs that
I was able to delete as well.
Are you using GUI from the client? My guess is that since this is a
cluster (I am assuming that from the name), it is using the local node's
IP address instead of the resources IP address and since there are no
backup images for the local node, it doesn't show any. My cluster
experience with
If you still have information about previous backups available, I would
query the catalog for available images on the media used for the backup.
This will tell you right away what is on the tape. You don't have
anything in place that changes the copy number of images, do you? Any
vaulting occ
Make sure you are selecting the proper policy type, i.e. "MS-Windows" for
Windows based systems or "Standard" for *NIX systems.
From: Grigore Petrisor
To: "VERITAS-BU@mailman.eng.auburn.edu"
Date: 07/25/2012 08:04 AM
Subject:[Veritas-bu] There are no files matching the spe
Greetings all,
I began using 2 STK T1C drives in my SL8500. Naturally, the bulk of
my clients will not be able to push data to the capacity of the drive, so
I have turned on multiplexing for certain policies and for the storage
group itself. I am running 6 jobs wide per drive, and seems t
Just so I didn't leave this hanging out there, I had 201 tapes with volume
expirations that had expired. It was a good lesson in the difference
between data expiration and volume expiration.
Thank you for all of the input!
From: David McMullin
To: "veritas-bu@mailman.eng.auburn.edu"
>From what I remember, the recommended catalog backup was the hot backup
when it first appeared (in 6.0?). I have had to use it a couple of times
with success every time.
Even if it could be dumped to disk and transferred/mapped to the new
system, it would be worth a try.
From: "Justin
Here is my environment:
NBU 7.1.0.3 on an AIX 6.1 Master (also acting as media server in this
instance)
My scratch pool has 201 tapes inside the robot. Robot is STK SL8500 with
20 9840C Drives. I am using ACSLS.
Overnight vaults failed with the duplication jobs ending in error 96
(unable to
I believe the default behavior is to run fulls every day until the first
scheduled full, even if an incremental is scheduled that day. We are in
the habit of running a full immediately after a client is added to a
policy or a policy is created.
We learned the hard way by filling up or virtual
We have the entire operating system of the master server on SAN (which
also happens to house NBU and its catalog), and it is replicated to our DR
site. Once we get there, we just power it up and fly. The caveat to
using SAN is that it looks like local disk to the master. I am not sure
about
Only on the older 10/100 NIC cards. They tightened up the RFC when
gigabit came around. If it don't autoneg, something is wrong.
From:
"WEAVER, Simon \(external\)"
To:
"Patrick" ,
Date:
09/27/2011 03:12 PM
Subject:
Re: [Veritas-bu] HELP
Sent by:
veritas-bu-boun...@mailman.eng.auburn
I think the question was: where is that file located when it is created,
and does it then get backed up by the job doing the backup itself?
From:
David Stanaway
To:
veritas-bu@mailman.eng.auburn.edu
Date:
09/07/2011 09:11 PM
Subject:
Re: [Veritas-bu] Dedupe ratios and checkpoint backups
Sent b
To answer your question, we tested both the NBU 5000 (Symantec) and the
DD860 (Data Domain) and the differences couldn't be more stark. I was
using a 30 minute checkpoint interval. I never achieved anything better
than 15:1 from the NBU5000 (which isn't bad). The DD860 hit 39:1 at the
time I
If at all, this would probably affect fixed-block solutions more than the
variable-block ones. The variable-blocked solutions would continue to
look for identical blocks, but in different positions of the data stream.
From:
"Stafford, Geoff"
To:
"veritas-bu@mailman.eng.auburn.edu"
Date:
0
We use Quest Authentication Services, and it has the ability to put a
dummy entry in the passwd and group files for the AD enabled users and
groups. This has a tendency to create havoc with other things, but for
something as this, it may work. In QAS, its called merging. In LikeWise,
you may
If you are currently dumping SQL to disk, more than likely, you are
compressing your data. Your dedup rate will not be good with this data,
because of the compression.
When we did our POC, I gave my DBAs specific instructions to send the
backups uncompressed. That way, the DataDomain would ca
In order to answer that question, another couple of questions would have
to be asked:
1. How long are you going to retain these dumps? I am assuming that they
are uncompressed. The longer you retain your backups, the better your
deduplication ratio will be. We are wrapping up a proof of con
That is your discretion. What that is telling you is who owned the files
when they were packed. If those users don't show up on your system, I
would chown them to the user:group that owns /usr/openv. If you want to
see what the user:group was in actual text form, you may be able to use
"tar
..or, you can use exclude_list, and the path gets excluded for all
policies.
Scott E. George
Technical Specialist - Lead
UNIX Systems Administrator
Phone: (216)896-2197
E-mail: scott.geo...@parker.com
From:
"Lightner, Jeff"
To:
"Jim VandeVegt" ,
Date:
07/06/2011 04:57 PM
Subject:
Re:
Funny, I ran into this yesterday, and removing and re-installing the
client did the trick for me. NBU 7.0.1 on AIX master here.
Clearing the cache seems to be a more elegant solution, though. :-)
From:
"Lightner, Jeff"
To:
"James Pattinson"
Cc:
"Martin, Jonathan" ,
Date:
06/02/2011 10:1
Currently running 7.0.1 on AIX, and I really want to move off to Linux.
The code maturity on AIX lags behind, in my humble opinion. I have one
EEB installed and one workaround in place to compensate for the lag.
New features are late comers to AIX as well.
I actually fielded the question to S
If you can handle the nuisance, IT Toolbox has a fairly busy Solaris
group. My ISP cut off direct access to newsgroups, but I think that
comp.unix.solaris still gets some activity. Darren Dunham used to be a
regular on comp.unix.solaris. He may have better recommendations as well.
From:
"P
It doubled in size for just about every other client too. It is all
inclusive, meaning that it contains all of the DB agents and deduplicating
client.
Personally, I am only rolling this out to where it is absolutely needed.
Scott George
From:
"Harry Tirrell"
To:
Date:
05/13/
We found this out all on our own, without Symantec's blessing.
They came a little late with their 2008 R2 support, so they let this go on
for CYA. We were doing this before we went to 7.0.1.
When you are caught in the dilemma of getting backups or being supported,
well, I choose the former.
We still bring back the 4.3.3 systems back at our DR exercise, but we use
mksysb for restoring the operating system. From there, we use NBU for the file
level restores, which do work.
-"Martin, Jonathan" wrote: -
To: ,
From: "Martin, Jonathan"
Date: 03/31/2011 08:17PM
Subject: R
Its not supported, but it still works. If you are still using that version of
a client, probably everything else about the client is unsupported too! We
have a couple of AIX 4.3.3 servers using that client. Backing up just fine.
The NBU client on these systems is the very least of my worries
I am curious if anybody has seen significant client backup slowdowns after
this upgrade (or one similar). My master is AIX 6.1 TL5. The slowdown is
being experienced on some previously better performing clients, mostly AIX
5.3. I would get 25m-35m/sec previously, where I am now getting ~30k/s
I performed the 6.5.5 to 7.0.1 upgrade on Tuesday, and made the amazing
discovery that my tapes are not expiring when the day/time comes. This
also means that I am not getting any of my media back from off site. Has
anybody else encountered this?
Thank you in advance!
Scott
"PLEASE NOTE: T
e it
serving up 110 drives across 6 adapters. I only needed one ovpass0 to
serve all 110 drives in 6.5.5, so I don't think I need to do multiples.
Any ideas?
Scott George
"PLEASE NOTE: The preceding information may be confidential or privileged.
It only should be used or d
ltiples.
Any ideas?
Scott George
"PLEASE NOTE: The preceding information may be confidential or privileged.
It only should be used or disseminated for the purpose of conducting
business with Parker. If you are not an intended recipient, please notify
the sender by replying to this me
ng up 110 drives across 6 adapters. I only needed one ovpass0 to
serve all 110 drives in 6.5.5, so I don't think I need to do multiples.
Any ideas?
Scott George
"PLEASE NOTE: The preceding information may be confidential or
privileged. It only should be used or disseminated for
Unless you are licensed on a per/TB model, then it is included.
From:
Nate Sanders
To:
"VERITAS-BU@MAILMAN.ENG.AUBURN.EDU"
Date:
02/22/2011 12:41 PM
Subject:
Re: [Veritas-bu] OpsCenter Analytics Licensing
Sent by:
veritas-bu-boun...@mailman.eng.auburn.edu
Apparently the answer is on a per-
20. I allow it to take up all of the available physical drives if
possible. My incrementals reside on VTL.
From:
ccosta@gmail.com
To:
scott.geo...@parker.com, veritas-bu-boun...@mailman.eng.auburn.edu,
VERITAS-BU@MAILMAN.ENG.AUBURN.EDU
Date:
02/22/2011 09:18 AM
Subject:
Re: [Veritas-bu
I have 2 vault jobs that duplicate everything from the previous day, one
for my fulls, and the other for my incrementals. For whatever reason, the
vault job for the incrementals piles a large amount of images into one
duplication job, sometimes up to 4TB worth, and it never finishes in time
fo
We are about to do a proof of concept of one in a couple of weeks. I will
post the outcome.
I would really like Oracle to release Solaris 11 with zfs deduplication.
When that happens, its going to drive down the pricing in the market.
In the mean time, I am in the market too, and shopping aro
How are you backing up the data? If it is a file level backup, then the
source or destination file system shouldn't matter. From BAR, you can
select the client with the vxfs file system as the source and the linux
client as your destination.I don't think block size matters, because
the fi
I forgot about the maillist.
If you are doing this with NBU, I would think backing up the Windows
client as a standard client and then restoring from that backup would
work. I had a slew of Windows servers get changed to standard clients and
they backed up just fine, but wouldn't show up as
I did this old school. I turned on the FTP server on the Mac, and used
Filezilla to mass copy everything over that I wanted.
It wasn't elegant, but it worked.
Scott E. George
Technical Specialist - Lead
UNIX Systems Administrator
From:
To:
Date:
09/09/2010 02:10 PM
Subject:
[Veritas-bu] r
David,
I am running 7.0 in test right now, and it does have the ovpass driver
installed and running. All of the associated drivers and scripts are in
/usr/openv/volmgr/bin/driver.
When I run lsdev, it is there and available:
corclv011:/usr/openv # lsdev -l ovpass0
ovpass0 Available 2U-08-2,0
NBU 7.0 requires TL7/MP5. Fix that and you are set.
Scott E. George
Technical Specialist - Lead
UNIX Systems Administrator
Phone: (216)896-2197
E-mail: scott.geo...@parker.com
From:
David McWilliams
To:
"scott.george"
Date:
08/05/2010 09:27 AM
Subject:
Re: [Veritas-bu] NBU 6.0 -> 7.0 upgr
After I sent the last e-mail, I went to Symantec's site and found an
interactive checklist generator:
https://vos.symantec.com/checklist/install#report
And generated this from the 6.0 GA to 7.0
Scott E. George
From:
David McWilliams
To:
NetBackup List
Date:
08/05/2010 09:08 AM
Subject:
[
It appears that you may be running a backlevel version of your C/C++
runtime libraries. Which version of AIX and what TL/SP is it? Also, what
does "lslpp -l|grep -i xlc" show?
Scott E. George
From:
David McWilliams
To:
NetBackup List
Date:
08/05/2010 09:08 AM
Subject:
[Veritas-bu] NBU 6
No special accounts here. They way they (Datalink) described it was that
the deadlock condition described looked like what I was experiencing.
Curiously, it may have addressed another nuisance issue I was having,
which was not always completing deferred ejects cleanly. What I was
having my op
Just to update everybody on this issue I have been having, it appears that
there is an EEB to fix nbrb due to a locking issue.
http://seer.entsupport.symantec.com/docs/346920.htm
I have been running it for a week with no issues. I have 21 days until I
see if it fixes my problem.
From:
sc
Judy,
I am curious as to what our differences may be, in regard why I am
experiencing this issue and you are not. It happened again over the
weekend, and I was able to determine that it happens every 28 days,
provided that I leave NBU up that long (which I normally do). I was
wondering if yo
Datalink is our support organization for our NetBackup support, and I do
have a case open with them. I am in the process of collecting logs and
waiting for it to happen again, which is an issue in itself. I have to
clear out the logs daily to avoid locking up the database.
Curiously, Ms. Hin
All that I have right now is a symptom. We have a hourly script that
utilizes vmquery to count the amount of scratch tapes, and sent the
results to a mailbox. If I leave the master server up for more than 2
weeks (I haven't determined the exact amount of time), my hourly e-mail
shows 0 scratc
It may, but I want to go to 7.0. I am dealing with a ~2 week recurring
problem on 6.5.5, and nobody can tell me if it is addressed in 6.5.6. I
have come to realize that AIX seems to lag behind in support, and wanted
to know if there was anybody else running 7.0 on it.
From:
"Baumann, Kev
I was just wondering if there is anybody running 7.0 in production on an
AIX master server. We need it mostly for the 2008 support. We are
currently running 6.5.5.
Thank you,
Scott
PLEASE NOTE: The preceding information may be confidential or
privileged. It only should be used or disseminate
I have tested this on some proprietary servers of my own, and the 5.1
client works with 7.0, although not supported. If you are running any
6.5.x with a 5.1 client, that isn't supported either, so no harm, no foul.
From what I understand, the code base isn't that different from 6.5 to
7.0, an
I have been working on this as well, trying to get media server dedup
working. I am caught in a type of chicken/egg loop. I can't create a
PureDisk storage unit, because I don't have a disk pool, but when I go to
create a disk pool, I don't have any volumes available. Just exactly what
volum
Getting ready to roll it into Test. They give you similar warnings that
they gave for the 5.1 to 6.0 upgrade, (i.e. database consistency), but I
have heard that it isn't that brutal and glides right through that
portion. I gotta test dedupe really bad, so my test system will be on it
in the n
What I think he is saying is that once you configure the IP address on the
L700 and connect it to the network, you should be able to access it via a
web browser. http:// No extra software needed. I just went
out and looked at mine and it does have an ethernet interface (unused by
me). I na
I can see an audit taking that long if there is considerable NBU/other
activity during the audit. We only have 3000 slots active for our SL8500,
and when there is other activity other than the audit, the robot stops
what it is doing, grabs a tape and mounts/unmounts it, and then gets back
to i
Greetings all,
We are testing BMR on RHEL 4 going to dissimilar hardware, and I was
wondering if anybody else had some input on getting over some hurdles that
we are experiencing. We are currently using bonding with our network
adapters at home, but don't care to in our DR test. Upon the firs
I'm a little fuzzy on Solaris, but it doesn't allow root login via ftp by
default. There are some modifications that need to be made to allow this,
some files that come to mind are /etc/default/login, or /etc/ftpusers.
Wait, I just did a google search and the files you are looking for are in
/
We just turned on BMR for a slew of Windows clients, and our tape
consumption has increased drastically. What am I missing? I thought the
BMR information was transferred back to the BMR Boot and Master servers,
but no additional media would be required.
PLEASE NOTE: The preceding information
Our last DR exercise, we tried using this file to switch quickly to our
duplicate copies. What we found out is that it works perfectly for
Windows clients, but doesn't work for UNIX clients unless bprestore is
used from the command line. We were using 6.0 MP5 last year, this year we
are using
I am making the assumption that you:
a) have no requirement for offsite storage or to put it in industry terms,
a disaster recovery requirement, and
b) are comfortable with keeping all of your backups on virtual storage.
Personally, I would be uneasy with this, because in the event of a
catast
We ran into a situation last week where we hit a tipping point that there
would be more scratch tapes needed than were coming back from off site.
Being the curious person that I am, I ran some commands to figure out what
we had coming back, what were already in the library, and others that fell
It appears that your solution may not be a software solution. It would
probably help if you gave us more information, like what kind of hardware
is being driven by the ACSLS server? Your hardware will differ on how it
indicates hardware problems, but my first guess is that you have a robot
do
John,
I believe that's it. Our documentation had references to the vmglob
command, but gave to reason why. That technote explains it.
As far as Sandor goes, what can you say about somebody who airs their
dirty laundry in a public forum? ;-)
Thanks John!
John Nardello
10/02/2009 02:34
Our last full blown DR exercise was done with 6.0 MP5, and our next one
will be done with 6.5.3.1. My master is on AIX 5.3 TL9.
I replicate my NBU server's operating system to another disk subsystem and
boot the copy on somewhat similar hardware in a remote location.
Everything works fine in t
My assumption has been that Linux clients are geared more around the
version of the kernel than the actual operating system. I don't see why
the 2.6 won't work.
smwoodcrafts
Sent by: veritas-bu-boun...@mailman.eng.auburn.edu
08/10/2009 02:42 PM
Please respond to
VERITAS-BU@MAILMAN.ENG.AUB
I want to pass my gratitude on to Neil; this was the fix for my issue. For
those who may be concerned, the NetBackup server starts the 5 minute
counter when it gives the message to empty the cap. For a server that has
to communicate to ACSLS to fill 2 - 39 slot caps, the timer will expire
befo
No, they are right on it, because they initiate the eject with the
deferred eject. What happens is the vault runs and does all of its dupes
and the finishes without ejecting anything. When they notice the vault is
done, they expand out "Vault Management" down to the job itself and right
click
All,
I have a NBU 6.5.3.1 server running AIX. and a STK SL8500 Library with
dual 39-slot caps, driven by ACSLS. Our vault jobs are to the point where
they are consistently more than 78 tapes, and when the deferred ejects are
performed, the remaining x-78 tapes will not eject. We just upgraded
I apologize for the stupid question, now that I have my test environment
back up and running (I asked the question blindly). For those who may
have been lurking, BMR gives you the option to create a new configuration
other than "current" based on older backups. No expiring of missing
images n
I have configured and tested BMR in my test environment, and am getting
ready to go to 6.5.3.1 next week in production. We replicate our NBU
server's rootvg (AIX) to SAN storage at our DR location, where similar
hardware sits ready to boot when the line is cut. Naturally, this will
keep the d
John,
Thanks for the advice, this one is the winner. We had the 4.5 client
installed on the AIX 4.3.3 server, for what seems like forever. The
previous admin (and my boss) told me that he attempted to install a newer
version without success. I took him at his word. After your post, I felt
Greetings all,
We have an AIX 4.3.3 client that is naturally unsupported, but still
working under NBU 6.0 MP5. We have scheduled to go to 6.5.3.1, but found
out that this client (I think it is a 4.5 client) will not even work
anymore. Now we have been told that this may put the brakes on the
The only way I have ever seen to backup Clear Case is to bring it down and
backup the directory. I do something similar where I backup a file system
that has to be syncronized with my DB2 database. Basically, you create a
separate policy that is user-initiated, and execute that policy from a
Greetings all,
I am testing BMR with 6.5.3 on a AIX 5.3 client. My SRT and my client are
patched to the same level. The entire restore process moves like it
should, but the bosboot doesn't succeed at the end, and the log gives me
this error:
0301-154 bosboot: missing proto file: /tmp/bosboot
Ken and Rusty,
That was exactly it. When I modified the vaults to do deferred eject, the
profile defaulted to a Robotic Volume Group that was not the one I
wanted. After changing it to the right one and re-ran the deferred eject,
everything worked!
Thanks again!
Scott
[EMAIL PROTECTED]
I am running NBU 6.0 MP5 with multiple MAP binary fix on AIX 5.3
I have one of three vault jobs that is not ejecting tapes. The vault jobs
initiates the normal duplicates during the process, and ends normally.
When I initiate the deferred eject for the job, it flows through like it
normally do
83 matches
Mail list logo
