Ladies and Gents,
As I'm changing jobs moving out of the NetBackup world for a while, I
just wanted to say thanks for the sharing of information, advice and
help on this list. And good luck with the interesting times ahead!
cheers, Phil
-
Egg is a
Thanks for highlighting this problem to the list, as we for one hadn't
even picked up that our zfs filesystems were no longer backing up. We
now have the engineering binaries, and our account manager at Symantec
has kindly chased up the TechAlert which should have been issued.
cheers, Phil
Is mtx any use to you (http://freshmeat.net/projects/mtx/). We use this
with a StorageTek L700 and Bacula.
regards, Phil
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of jonafc
Sent: 01 July 2008 17:49
To: VERITAS-BU@mailman.eng.auburn.edu
Subject:
If you are using an existing media server, then you just need an
Advanced Client licence (NetBackup 5.1, as of 6+ now called Snapshot
Client I believe).
Phil
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
dmdelfini
Sent: 18 June 2008 15:04
To:
Is this definitely correct? I was under the impression that it would
join the storage unit unless there were already 4 streams from that
schedule already, i.e. that it would mpx higher than the schedule limit
so long as it was with streams from other policies.
I guess I should test this all out
Sounds similar to this
http://seer.entsupport.symantec.com/docs/290900.htm which is definitely
a case of scary. Seems like there is a problem with the way MS DFS
Replication shares interacts with NetBackup. I don't do Windows either
but my MS colleagues assure me that there is nothing funny
We've using Clariion Cx300 with ATA drives for disk staging and
performance is awful; seems to deteriorate over time. Currently RAID5
LUNs (tried RAID3 which had similar performance and even tested RAID0
which wasn't noticeably better). I'm assuming the performance issue is
that with lots of
If you're lucky enough to have a configuration management or asset
management system, you may be able to find a way to report on all the
servers that you aren't backing up, then you can hassle the admins who
have created those servers. It's still not foolproof, as you end up
with a list of
Make sure your NetBackup patch levels match between the master/media and
the client. Check firmware levels on network cards / hba's throughout
the system.
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of rascal
Sent: 13 February 2008 16:51
To:
And another thing!
Both the update_clients and update_dbclients fail because they can't
find gunzip in the PATH on the client, saying PATH is set to
/usr/bin:/usr/sbin. Gunzip is in /usr/local/bin. I've tried setting
the path in /etc/profile, /.profile, /.kshrc but currently to no avail
and
Business Technology
Phone: 01384 26 4136
Mobile: 07748 333503
Jabber: [EMAIL PROTECTED]
Egg Banking plc
From: Ed Wilts [mailto:[EMAIL PROTECTED]
Sent: 30 January 2008 19:41
To: Weber, Philip
Cc: VERITAS-BU@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] NBU
libgfspfiST.so
-r-xr-xr-x 1 root bin 964540 Nov 16 17:24 librawprtfiST.so
-r-xr-xr-x 1 root bin 5879916 Nov 16 17:24 libnetappfiST.so
_
From: Weber, Philip
Sent: 25 January 2008 15:20
To: 'VERITAS-BU@mailman.eng.auburn.edu
For NetBackup 4.5 - 5.1 we had a script which let us install Solaris
clients from a central admin server (not a NetBackup server). This was
because we have firewalls between the servers and clients, can't use
ftp, can't ssh between the master and clients, etc. So our script :
hacked the client
If you're using Advanced Client backups (proxy copy for Oracle), you'll
get straight filenames for the datafiles, and some pieces (Oracle RMAN
pseudo filenames) for the logs, control files etc. (This is for NBU
5.1 which is the only version for which I have experience of proxy copy
backups).
I
I'm about to upgrade to 6.5.1. Probably also about to be made
redundant; so perhaps I can blame that on 6.5.1 :-O
I'm happy to help out with the FAQ.
cheers, Phil
Phil Weber MBCS
Storage Technical Services - Senior UNIX Technologist
Business Technology
Phone: 01384 26 4136
Mobile:
In our experience :
DSSUs don't work in Storage Unit Groups at NBU 5.1 as NBU isn't clever
enough to handle one unit being full and move onto another. Having said
that, it shouldn't fail jobs when the DSSU is 100% full unless it hasn't
found any old images which can be expired from disk
of each, i.e. disk only / disk tape / various error
statuses. Not what you'd want to run every 5 mins, but not too bad.
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Ed Wilts
Sent: 15 January 2008 12:37
To: Weber, Philip
Cc: veritas-bu
Not sure what you mean by if floating point is required. There are
compatibility matrices for OS libraries on support.veritas.com. We
are running NetBackup 5.1 on Sun T2000s and v880s.
cheers, Phil
Phil Weber MBCS
Storage Technical Services - Senior UNIX Technologist
Business Technology
Interesting, I had that problem (with the restore) and gave up in the
end did the database migration in a different way, didn't log a call,
tut tut...
I always find the support split between a pita getting to the right
people, then when you get a good engineer, they are fantastic.
Phil
Anybody any experience of Spectralogic's library-based encryption? We
are about to start out no implementing this - basically the encryption,
and optionally compression, is carried out in the fibre cards in the
library. We will be using LTO2 and LTO3 drives.
I have some concerns around impact
We install the NetBackup client in the global and local zones, i.e. we
back up the local zones locally as they will have the knowledge of their
application/DB status. We don't use loopback filesystems in the local
zones for our data filesystems, we either use devices exported from the
global to
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
Of Ed Wilts
Sent: 20 July 2007 01:29
To: 'Brandon Zermeno'; Veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] retentions
What we have actually done here is to push the failures back on
I've got a script which I was going to post after checking it for
site-specific stuff, but I think the page at that link is far more
comprehensive!
Phil
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
Of Dave Markham
Sent: 20 July 2007 12:26
To:
I've had this issue where there's a problem with connectivity (e.g.
firewall) between the master and the client, but the (different) media
server and the client can talk to each other correctly. It's like the
master can't query the client to find out what filesystems need backing
up, but the
I back up ALL_LOCAL_DRIVES using the cluster name, and ensure local
drives (only C: for us) are excluded. Then back up C: only on the
individual nodes. It's still a pain managing exclude lists; from time
to time someone changes the policy names which messes the exclude lists
up and we start
Hi, we've been getting the same problem with 5.1 MP6. See
http://seer.support.veritas.com/docs/247062.htm.
My colleague has been working on this, it looks like with the
bpstart|end_notify scripts in place as described and the Content
Indexing service set to Manual (ours was disabled by a
Exactly the symptoms we were getting.
GUI -- master properties -- Client Attributes -- Windows Open File
Backup tab (naming may be different for v5.0, I can't remember now).
-Original Message-
From: Dave Markham [mailto:[EMAIL PROTECTED]
Sent: 06 July 2007 11:11
To: Weber, Philip
I'd never seen this before but funnily enough just tripped over it on a
couple of clients I'm moving from NBU 4.5MP9 master to 5.1MP6 master.
Both clients are W2K3 remaining at 4.5MP9 for now (because I can't
access them to upgrade). bpclient -hn on the master fails to return the
IP address
Off topic a bit, but does anyone know how to map the st device names to
rmt device names in the iostat output? I've written a script to do it
but it seems a bit like hard work ... on my Solaris 9 boxes iostat -xn
lists the rmt device names.
thanks, Phil
-Original Message-
From: [EMAIL
If NetBackup can't move a tape back from the drive to its slot, it will
down the drive (at least in our environment - as the OS is Solaris
will log a message to /var/adm/messages).
We used to have a small robot where the OPs manually changed tapes by
opening the robot up moving tapes. As part
Master/Media Solaris9 NBU 5.1 MP6.
Client Solaris10 NBU 5.1 MP6.
Client has two network interfaces client-pub client-bk.
I want all backup traffic to go via the client-bk interface, but unless
I add a network route for the backup server subnet over the client's
backup interface, backups fail
finally been recognised as a data loss
issue to be fixed in a forthcoming MP.
-Original Message-
From: Anas Kayal [mailto:[EMAIL PROTECTED]
Sent: 20 May 2007 05:38
To: Weber, Philip; veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] Windows VSS Q
Phil,
You have an MSSQL Server
Master/media: Solaris 9 NBU 5.1 MP6
Cient: Windows 2003 NBU 5.1 MP6
Dear all,
Noticed a Windows 2003 full backup wasn't as big as expected, and found
messages similar to the below in the client's bpbkar log. I get these
regardless of whether Windows Open File Backups are enabled :
2:04:56.126
Sorry...should have been a new topic :-( I'll re-send this.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
Of Weber, Philip
Sent: 18 May 2007 15:08
To: veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Tuning Windows backups
Master
Master/media: Solaris 9 NBU 5.1 MP6
Cient: Windows 2003 NBU 5.1 MP6
Dear all,
Noticed a Windows 2003 full backup wasn't as big as expected, and found
messages similar to the below in the client's bpbkar log. I get these
regardless of whether Windows Open File Backups are enabled :
2:04:56.126
We still have a fairly busy 4.5 environment due to older OS versions so
I can't see us moving off our 5.1 environment for a while ... this for
us is the show-stopper for not upgrading.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Ed Wilts
Sent: 15 May
I believe you can get this when resources (memory?) are overloaded on
the master/media. Can't be more specific as only seen it very rarely
so haven't investigated thoroughly.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
Of Brooks, Jason
Sent: 24
I recently added a Solaris 10 media server/robot host to our existing
Solaris 9 master medias environment. If I recall, didn't really have
any problems. NBU 5.1 MP6. Not SSO (though we use SSO on the other
media servers).
You may need to run /usr/sbin/inetconv to convert the /etc/services
I parse bpdbjobs -all_columns output, pull what fields we are interested
in into a MySQL database, from where I produce various email and HTML
reports are produced cutting the data by various criteria which
management are interested in. It's taken quite a while to evolve but I
can now usually
This means the script called by the rman policy returned a non-zero
status code, so could be rman returned an error which was returned to
the script, or something else in the script returned a non-zero code
(including a scripting error). If there's no error in the log file
written to by the rman
I spent some time on this a year or two ago for backing up Oracle
without using the NetBackup Agent. I found
1. Do the start of backup work on the first stream that runs, which
isn't necessarily STREAM_NUMBER=1. Set a flag to indicate it's
complete. You'll have to do some checking to prevent
Are you backing up through a firewall? We have a similar issue where
the final acknowledgement isn't getting back from the client to the
media server, so the backup hangs at 100% complete. Only for occasional
backups, and maybe those with more streams than others. It isn't
clear-cut though, and
]
[mailto:[EMAIL PROTECTED] On Behalf Of Hindle,
Greg
Sent: Monday, December 18, 2006 9:30 PM
To: Weber, Philip; NB List Mail
Subject: Re: [Veritas-bu] Firewall setup
OK. So on the firewall you only open ports 13782, 13724 and 13720? And
then configure the clients to use the same 3 ports? All other
NetBackup 5.1 MP5, Solaris 9 master/media servers.
I have a Windows 2000 client with approx 900 Gb D: drive which I want to
split into multiple streams, e.g.
Stream 1 :
D:\Shares\shareddata\All Departments\folder1 (100 Gb)
Stream 2 :
D:\Shares\shareddata\All
To: veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] LTO3 Read/Write LTO2
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Weber, Philip
Sent: December 19, 2006 12:30 PM
To: Wessam Aly; Dave Brown
Cc: veritas-bu@mailman.eng.auburn.edu
Subject: Re
I've also heard that once you've used LTO2 media in an LTO3 drive, you
can't then read it in an LTO2 drive. Not sure where I got this from
though.
I've also got some new LTO3 drives having asked for LTO2...so far they
seem to be running slower than the LTO2s as I can't supply data to them
fast
We've gone firewall-mad over the last couple of years and now pretty
much all of our clients are behind at least one firewall from the
perspective of the NetBackup servers. In general we open ports 13782,
13724 and 13720 in both directions, to make life simpler. This can be
reduced so that only
You can set the Maximum Concurrent drives used for backup in the
storage unit, but that doesn't give you any control over WHICH drives
are used, which sounds like what you want. I think you could use
tpconfig to manually set which drives (that the OS can see) are
available to NetBackup, but I'm
We take tapes out regularly, using a script that checks which tapes
aren't going to expire until a certain month. When the script ejects
these, it changes the volume group for the tape to be the month name.
The tape then goes in the box similarly labelled with the month name.
When a restore
Haven't got SAN Media Servers but have got SSO I have had some luck
getting drives back by unloading and reloading the sg driver.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Wilkinson, Alex
Sent: 06 November 2006 03:07
To:
Bit of a problem if you've got icmp (and ftp, and...) disabled between
most servers clients like we have. Is there anything similar that can
work over the NetBackup ports which obviously will be open?
Phil
-Original Message-
From: Martin, Jonathan (Contractor) [mailto:[EMAIL PROTECTED]
Title: Message
I
think you could builda new media server (new name), before removing the
media server from the app server. Then use "bpmedia -movedb" to change
ownership of all backup images owned by the old media server, to the new media
server. Then decommission the app server
Title: Solaris 10 Zones
Hi all,
I'm fairly new to Solaris 10 trying to work out the best way to back up a Solaris 10 server with local zones (NetBackup 5.1). Any suggestions or pointers to best practice?
It seems that by default :
1. the global backup will include nearly
Title: Message
Does
nbu_snap support ufs filesystem?
-Original Message-From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Evsyukov, SergeySent: 16 October 2006 09:47To:
veritas-bu@mailman.eng.auburn.eduSubject: [Veritas-bu] NBU_SNAP
issue
Title: Message
These
are log files for the robot services. If I recall, they will create a log
file for each day. You can create under debug the directories tpcommand,
reqlib, ltid and daemon and different logs will be written to each. You
should be able to delete old log files, but
Title: Message
Trouble acquiring decent hardware.
Interested though, is that concurrent operation, i.e. writing at these
speeds and reading at these speeds concurrently? I'm not familiar with
bonnie++.
cheers, Phil
-Original Message-From:
[EMAIL PROTECTED]
Sorry if this has been done to death recently but...
What sorts of disk systems are people using as disk staging to keep
LTO3s happy? Our current EMC Clariion Cx300 struggles to supply data
fast enough to 3 LTO2 drives (even outside of NetBackup) and performance
goes through the floor if (as is
My understanding of CentricStor is that NetBackup only knows about
CentricStor's virtual tape devices. CentricStor offlines virtual tapes to
physical tape as and when the virtual tapes become full, and NetBackup knows
nothing about where the data is on physical tapes.
Interested, because
: Peter DrakeUnderkoffler [mailto:[EMAIL PROTECTED]
Sent: 22 September 2006 10:48
To: Weber, Philip
Cc: veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] LTO3 and Disk Staging
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Some quite valid points. The splitting of a DSU's to alternate
September 2006 12:37
To: Weber, Philip
Cc: veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] LTO3 and Disk Staging
On 9/22/2006 4:19 AM, Weber, Philip wrote:
What sorts of disk systems are people using as disk staging to keep
LTO3s happy?
We're using HP EVA with FATA (Fibre ATA
Title: Message
Sort of related ... when I try to shut down
NetBackup 5.1 there are almost invariably disk duplication jobs running.
If I kill these from the GUI they immediately spawn new jobs. Even after
bp.kill_all there are processes running because the duplication jobs have
Title: Message
We're
using advanced client for offhost backups, using nbu_snap (using the media
server as the data mover, not doing Flashbackup). Currently only on MP4,
haven't got around to MP5 yet though it is supposed to include fixes for some of
the issues we have had, namely :
Title: Message
We are
backing up Solaris 10 clients to a NetBackup 4.5 master (using the NBU 4.5
client). Only problem is, they end with status 1 as the client doesn't
understand the fs type of /system/contract and /system/object - putting these or
/system in the exclude list doesn't
We use the Advanced Client add-on for 5.1 (another licence - name may be
changed in v6.0). One of our media servers is configured to see the
same SAN disks as the servers which we want to back up. Throughput
seems to be very roughly twice what we would get over the 1 GB LAN.
Nothing is mounted
Title: Message
I
don't think so. With versions of Oracle previous to 8i the libobk.so
libraries were linked into Oracle which might have been a problem for updating
the NetBackup Agent for Oracle from the master server, but I have not had
problems with the UNIX client update and the
Title: Message
We get
around 4% failures, and that is only due to continuing focus on backups.
The product is not at fault though, more our infrastructure as a whole (i.e.
networks and firewalls) and hetrogeneous server estate.
-Original Message-From:
[EMAIL
I did once look into all of our Status 1 results and had several along
the lines of it's a WINS server and the WINS DB failed ... sounded
important to me. Unfortunately, Status 1 is still at the bottom of my
list of things to look at in terms of priority, because there are too
many things above
Issues we have had using disk staing at 5.1MP4 include
o disk staging jobs returning with errors if
- no images are found (i.e. staging is complete)
- an image listed for duplication is not found (maybe it was
expired during the staging activity)
- configured to do more
Title: RE: [Veritas-bu] Netbackup Device Configuration problems :-(
I've occasionally found it necessary to delete everything NetBackup knows about the attached robots and drives, and re-scan from scratch, using the steps below. Seems a bit drastic, but it does the job and the
Title: Message
Hi all, has anyone tried using bprestore from a Windows
client?
Acolleague wants to use bprestore to restore
permissions only, for a set of files, restoring to a new directory. We are
having trouble getting the renaming to work.
NetBackup 4.5 FP 9, Solaris
Title: Message
A lot
of our clients are on 100 meg interfaces we get better performance out of
our tape drives if we do not multistream, but do multiplex. Otherwise the
streams aren't coming from the clients fast enough to keep the tape drives
streaming nicely, without
71 matches
Mail list logo