Assuming you meant to say on the disk array side or box side I'd say the
array side. Hardware RAID is more efficient than software RAID because
the hardware typically has its own caching and CPUs.
Having said that I will note that on most enterprise level systems
you'll see hardware RAID but
?
When I used to have archiving of oracle logs in NB it wouldn't delete
the file until the 2nd time it was backed up to tape. Since we were
also doing log shipping and backed up both from the primary and standby
databases we typically had 4 copies of the logs at any given point.
-Original
Title: Large Oracle RMAN DB Backup Question
I havent done it but my coworker
here has indicated several times that RMAN isnt good for hot backup of
large DBs because it does stair stepping of the tables and takes
forever to backup due to this. Perhaps others on the list have more details
Title: Windows backup failing to Unix Server - Succeeds to Windows server
An odd one here. Our Windows Admin has attempted to setup a Windows machine to be backed up by our Unix Master. When he does he gets a Status 59 which means the server isnt recognized by the client. When he uses his
I wouldn't backup via autofs. First why backup via NFS rather than
backing up the filesystem from its native host? You still have the
network impact but don't have to worry about NFS' overhead and quirks.
Second if you really have to do NFS why do autofs? You could set up a
script to mount
Title: Message
For Unix/Linux clients your exclude list
file name should contain the name of the policy.
i.e. If the policy is named HOSTNAME-OS
rather than just creating:
exclude_list
You should create:
exclude_list.HOSTNAME-OS
From:
[EMAIL PROTECTED]
[mailto:[EMAIL
We did NOT give the DBAs access to the Netbackup GUI. We tell them the
name of the policies and they include them in their RMAN scripts. You
HAVE to run RMAN because Netbackup and Oracle only support RMAN for the
OCFS filesystems used on Linux. (We didn't do ASM - suspect its true
for that as
Sudo is a great idea for using utilities but any Unix Admin worth
his/her salt isn't going to give you a root shell via sudo - it defeats
the whole point of not giving out the root account in the first place.
The audit objections to having root is not the specific account but the
total power over
Title: Restricting tapes to a single stream
Is it possible to setup a policy in such a way that any stream will not use a tape previously used within another stream of the same backup?
Background:
We run our Production DB backup using 3 streams. This backup is used for doing refreshes to
as a single stream).
From: Greenberg,
Katherine A [mailto:[EMAIL PROTECTED]
Sent: Tuesday, January 31, 2006
10:22 AM
To: Jeff Lightner;
veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu]
Restricting tapes to a single stream
Why don't you try to identify and repair what
, January 31, 2006 11:08 AM
To: Jeff Lightner
Cc: veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Restricting tapes to a single stream
Well, there's no setting of this kind, so you'll need other
workarounds...
As you mentioned, single stream backup. Not very desirable, so that's
out
Title: Error 58 got me stumped
Interestingly Ive had the same issue.
You need to look at the admin guide discussion on firewalls.
Although bpcd does the original connection
there is a call back port opened that is likely blocked by your internal
firewall. (Im assuming you have one
]
[mailto:[EMAIL PROTECTED] On Behalf Of Chapman, Scott
Sent: February 6, 2006 12:17 PM
To: Jeff Lightner; Hindle, Greg;
veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] Error 58
got me stumped
check out the bpclient command. this
will allow you to use vnet without boucing the masters
You can backup across your WAN but it will
be very slow and will impact other traffic.
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of j. okabayashi
Sent: Wednesday, February 15, 2006
1:44 PM
To:
veritas-bu@mailman.eng.auburn.edu
Subject: [Veritas-bu] Backing up
to in my working DMZ config.)
Note: You have to do the selection for each DMZ client you add to the
list under the Master's Client Attributes box.
-Original Message-
From: Paul Keating [mailto:[EMAIL PROTECTED]
Sent: Tuesday, February 21, 2006 8:47 AM
To: Brooks, Jason; Jeff Lightner
I wonder if the type of failure determines this. If the tape is frozen
because NB suspects it to be bad one probably doesn't want it to just
continue from that point. If the tape was truly bad then you'd end up
with only a fraction of a backup.
-Original Message-
From: [EMAIL PROTECTED]
I've seen this on HP-UX. We have fibre bridges going to the SCSI drives and
usually booting the fibre bridge and power cycling the drive takes care of it.
In fact just power cycling the drive has taken care of it on a couple of
occasions.
-Original Message-
From: [EMAIL PROTECTED]
Whether it's a kill -9 of a process or a Glock 9 used on a fellow admin
you're likely to have problems afterwards that make you question the
wisdom of the kill. :-)
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of WEAVER,
Simon
Sent: Wednesday, March 01,
Of course if you're using a Windows workstation like most people there
is a Java client you can install there rather than running the GUI
directly from Solaris. You just use the one on your workstation to
login to your Solaris server.
-Original Message-
From: [EMAIL PROTECTED]
The restore log itself isn't telling you anything?
Typically I see things like:
Restore started 03/04/2006 07:11:38
07:11:43 (109902.xxx) Restore job id 109902 will require 3 images.
07:11:43 (109902.xxx) Media id SU0188 is needed for the restore.
07:11:43 (109902.xxx) Media id SU1532 is
Title: Weird restore problem
Have you verified the tape isnt in
use by another backup or duplication job?
What does vmquery report about the tape?
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Hindle, Greg
Sent: Wednesday, March 22, 2006
9:41 AM
To:
Title: Message
It shouldnt fail so long as your
policy is using the virtual name rather than the physical name. Ideally you
should have different backups configured:
1) policy-server1 = Backs up the OS and other local filesystems unique
to this physical host
2) policy-server2 = Backs
Title: Issues with NetBackup after applying Solaris patches
Havent used Solaris in a while but
this reminds me of a time I did an update on some Solaris production servers
and found the patch included a default st.conf that overwrote the specific
entries wed made for our AIT libraries.
For the separate backup network we simply add entries to the hosts file
on the backup master. The entries added there typically just have b
appended to the real host name. The appended to hostname is then used
as the client instead of the real hostname.
Example: Real hostname, exchange1,
Support organizations not coming up with obvious solutions is something
I'm fairly used to. In fact I've had occasions where I theorized as to
the correct answer but was told by support that wasn't it only to later
try it in desperation and find that it was indeed the fix.
Perhaps Katherine's
Title: Message
It means that you would have to recover
from whichever tape was last used.
The first backup uses the first tape.
The second backup uses the second tape. The third backup uses the first tape
again. And so on
If you had a failure after the fourth time
you did the backup
Didn't respond earlier because I thought your post regarding Sun only. On HP
the device's minor specifies settings of the drive's device entries - it will
have several. You need to make sure your minor numbers are for the Best
density which would include compression.
Also as noted by
FORCE_MEDIA_SERVER entry in bp.conf would be sufficient for the same
drives on another media server in the same SSO config. By newer and
cleaner drives it sounds as if you're using a different storage unit so
I'm not sure what that would require.
The entry in bp.conf would be:
From the bpminlicense man
page:
The bpminlicense
utility manages a NetBackup license file. The
preferred method to
manage NetBackup licenses is to use the Help
License Keys panel
in the NetBackup Administration console. For UNIX
servers, you may
use the get_license_key(1M) utility
Of course for facilities where you don't have and don't expect to have
more than 6 domains this isn't an issue. Therefore using
/etc/resolv.conf is adequate for most environments though I'll have to
admit I usually use the FQDN.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL
host name with a
b appended). We don't put those entries in DNS as they are used by
nothing other than the backup master and we don't want any other traffic
on those NICs.
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: Monday, May 01, 2006 11:06 AM
To: Jeff
, 2006 3:09
AM
To: Jeff Lightner; Bob Stump;
Marianne Berg; veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] can SAN
media server restore clients?
SSO = Shared Storage Option Jeff :-)
Allows Servers (Media as example) to share 1
Library!
Regards
Simon Weaver
3rd
I'm with you. Linux for enterprise is a reality. RedHat and Suse are
the two main commercial variants. We use RHEL here and they used it at
Cisco when I worked there as well.
Not supporting the two main commercial distros is showing a lack of
recognizing reality. My company has no plans to
This makes the assumption someone at Symantec actually looks at these
things. I don't know of anyone who has posted on this link that has
ever seen a response. I know I haven't.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Tristan
Ball
Sent:
Thats not the culture of this list
although it is on others I am subscribed to. The downside Ive seen to
what you propose is that many times people do not post the summary so one finds
more questions than solutions in the archives.
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
Ive run Netbackup on HP-UX for
years and dont see many major issues. Ive also run it on Solaris
though it has been a while since I did that. It always seemed to have more
issues on Solaris than HP-UX though once properly configured it seemed to be
acceptable on Solaris. There just seems to
There IS dollar savings even if you go the commercial Linux variants and
pay for support as compared to commercial Unix (especially the RISC
based solutions).
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Paul
Keating
Sent: Tuesday, May 16, 2006 12:06
-Original Message-
From: Jeff Lightner [mailto:[EMAIL PROTECTED]
Sent: Tuesday, May 16, 2006 12:49 PM
To: Paul Keating; Greenberg, Katherine A; Dhotre, Shekhar; List Veritas
List
Subject: RE: [Veritas-bu] Symantec and Linux support
There IS dollar savings even if you go the commercial Linux
Title: Message
No DBAs are demons subordinate to the REAL
spawns of satan Developers.
It has always amazed me that in an
environment where one runs an OS used by thousands or organizations, a DB used
by thousands of organizations and tools used by thousands of organizations in
which
the RISC systems are usually made by the same people making the
Unix that runs on it this discussion is usually internal to the vendor.
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: Tuesday, May 16, 2006 1:32 PM
To: Jeff Lightner; [EMAIL PROTECTED];
[EMAIL
Title: Message
SOX may only require 3 years but dont
forget for taxing authorities (IRS, states etc) they may go back 7 years.
Also for some lawsuits data may be required that is older than 3 years old. We
typically do yearly backups with infinite retention here of our Production
database.
Title: Experiences with DataDomain
We just saw a demo of the DataDomain technology. Anyone using this and if so how? What are your experiences? What would you do differently? Are you using it in SAN config like a library or with IP to each host?
Jeffrey C. Lightner
Unix Systems
The NetBackup inventory compares what NetBackup thinks the robot has to
what the robot says it has. This should be relatively quick. I've used
the Sun OEMed L700 (StorEdge) in the past with DLT drives and didn't see
the kind of delay you're talking about. Here we use the HP OEM and
don't see it
Wow this takes me back...
I used to work in a shop where we had Sybase and SQLBackTrack running on
HP-UX. There too we ran into issues with shared memory because the
Sybase available was compiled as a 32 bit binary. Even though the OS
itself was 64 bit we had the limitations of a 32 bit host
Status 1 is not a failure UNLESS it's a database backup. For
filesystems everywhere I've been we've shown that a status 1 is not a
failure but a warning. Unfortunately you won't get Veritas to
officially confirm that.
Some things I've seen cause a status 1:
Open files.
Files changed between
For activity monitoring you can't beat the GUI.
For newbies (and tasks done infrequently) the GUI is nice as it often
lets you figure out what to do from its menu selections.
For experienced users the CLI is more flexible because you can give the
flags you want to commands and can also do
the switch.
Pretty neat, actually.
Paul
--
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
Of Jeff Lightner
Sent: June 30, 2006 1:19 PM
To: Keith W; Veritas List
Subject: Re: [Veritas-bu] CLI or GUI
For activity monitoring you can't beat the GUI
I'll have to agree. Left it out of my prior post but I do usually use
the GUI for policies administration.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
[EMAIL PROTECTED]
Sent: Friday, June 30, 2006 2:20 PM
To: veritas-bu@mailman.eng.auburn.edu
Title: Message
Just putting the drives in wouldnt
make them available to the OS (HP-UX). You can cheat and simply reboot the
HP-UX host or you can do insf e to make it install the devices.
ioscan fn will show
you the hardware your HP-UX sees. Make sure the drives dont show as UNCLAIMED
Thats because all their binaries
have a new system call that executes a shell routine:
if ! $0
then exit 0 # Dont let the user
know it really failed
fi
J
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Stefanos Monovasios
Sent: Saturday, January 01, 2000
From the American perspective:
Was there some kind of game being played this weekend?
Just kidding :-)
Congrats to Italy.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of WEAVER,
Simon
Sent: Monday, July 10, 2006 2:04 AM
To: 'Alessandro da Silveira';
Are you sure you didn't intend THIS one for Justin :-)
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of WEAVER,
Simon
Sent: Thursday, July 13, 2006 9:29 AM
To: 'Brooks, Jason'; veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Backup Speeds
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Ed Wilts
Sent: Monday, July 17, 2006 7:41 AM
To: Drunen van, Marcel
Cc: veritas-bu@mailman.eng.auburn.edu; [EMAIL PROTECTED]
Subject: Re: [Veritas-bu] Unix Master, Windows Media server possible?
On Mon,
We moved from an HP-UX 11.0 server to HP-UX 11.11. Due to this we were
able to move the catalogs etc... without having to reimport media. To
dodge the server name issue we just added another Name with old IP to
the same NIC as the primary interface.
-Original Message-
From: [EMAIL
The problem is that fuser doesn't always show everything. For example
sometimes if you are running as a user and cd to a directory then su to
another user and cd out of the directory fuser won't show what is
busying it out. lsof on the other hand will.
I found out about lsof years ago and now
You might want to in order to do database refreshes by restoring one
machines database backup to another's database. We currently do that
with tape using the FORCE_MEDIA_RESTORE option in our SSO setup. I
don't think having separate disks would eliminate this need unless you
wanted to do a 2
We have been using the FC/SCSI bridges on
our STK 700 for a few years. We dont have the feature
turned on. They dont seem to cause us problems so I wouldnt say
to get rid of them. The only issue we saw with these was when we tried to add
them to our Linux servers to be used as media
Is the master a Windows machine as well? If its Unix/Linux you can
download and install tcpdump which will let you specify interface and IP
you wish to monitor. You can use ethereal to look at the packets
captured by tcpdump.
Not sure if there's a Windoze equivalent or version of those tools.
I'm starting to get the impression that in 6 NBU = No BackUps and MP =
Mangled Patch. :-)
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Major,
Rusty
Sent: Thursday, August 17, 2006 11:32 AM
To: Ed Wilts; veritas-bu@mailman.eng.auburn.edu
Subject: Re:
Title: Message
Sure NOW this gets post. Last week on
writing a script to update bp.conf with sed I had to actually read documentation
to figure this out. :--(
From:
[EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Clooney, David
Sent: Friday, August 25, 2006 5:53
AM
To:
Title: queued job never going away
In the activity monitor on a few occasions weve seen jobs staying queued forever. Is there a way to get rid of such queued jobs?
Canceling them or deleting them doesnt seem to help.
It appears they occur when a client initiated action occurs and goes
Sorry - hadn't seen this correction before I responded to your earlier
mail.
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: Friday, August 25, 2006 1:01 PM
To: [EMAIL PROTECTED]; Jeff Lightner;
veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] queued
:00
PM
To: Jeff Lightner;
veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] queued
job never going away
Document ID: 278498
http://support.veritas.com/docs/2784984
Matthew Johnson
Entertainment Partners
Backup Administrator
Work - 818-955-6357
Cell - 805-469-8867
Title: Message
It must be nice to live in Europe where
you get 4-5 week holidays instead of 2 one week vacations. You actually have
time to go somewhere nice. In a prior life I worked in hotels the last of
which was in Grenada and we had mostly British and German tourists there.
Yep. My company wouldn't let me interview.
Interesting that the guy says he's unaware of the issues. Do you
suppose support doesn't report the issues or that he just makes it his
business to stay unaware so that he can make comments like that to
reporters?
I didn't really see them mention the
or the grass.
Phil
456-3136
-Original Message-
From: Jeff Lightner [mailto:[EMAIL PROTECTED]
Sent: Thursday, September 07, 2006 9:30 AM
To: Mansell, Richard; veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] Fame at last.
Yep. My company wouldn't let me interview.
Interesting
Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
Of Jeff Lightner
Sent: September 7, 2006 11:18 AM
To: Koster, Phil; veritas-bu
Subject: Re: [Veritas-bu] Fame at last.
I didn't say they had a good case for libel but rather that they might
try such a case
at such suggestions?
-Original Message-
From: bob944 [mailto:[EMAIL PROTECTED]
Sent: Thursday, September 07, 2006 1:33 PM
To: veritas-bu@mailman.eng.auburn.edu; Jeff Lightner
Subject: RE: [Veritas-bu] Fame at last.
From some strange Oliver-Stoneish planet, Jeff Lightner
[EMAIL PROTECTED] wrote
By the way, the September issue of Storage magazine also has upgrade
path bumpy for major backup app as the cover story. It is about NBU 6
but seems to focus on one company's strategy for migrating to it and why
they are holding off on MP3. For whatever reason, the magazine's site
only goes to
Weve seen the activity monitor
failure when one of the fibre cards in our HP-UX master server has gone away.
On HP we can reset the fibre card with the server still on line using fcmsutil.
Not sure if its possible in Linux or how to do it if so.
From:
[EMAIL PROTECTED]
are SCSI via fibre bridges including the robotic control.
From: Dustin D'Amour
[mailto:[EMAIL PROTECTED]
Sent: Friday, September 08, 2006
11:22 AM
To: Jeff Lightner;
veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] Strange
Errors
Is this when the Master and Media
I read it as When it tells you it completed it may have only backed up
1K so is not really complete. At the time I read it I thought it meant
1K data so missed the files. It does seem to beg for some
clarification.
It does seem it could be read as only backed up 1000 files (or 1024
files) OR it
Are you backing up the UNIX machines with NetBackup? If so perhaps
that's your master and you just need to setup a client backup policy for
the Windows server.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of WEAVER,
Simon
Sent: Thursday, September 14,
, 2006 9:11 AM
To: Jeff Lightner; veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] Backup through firewalls
--
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
Of Jeff Lightner
Step by step notes I wrote when I did this:
FYI the following
:[EMAIL PROTECTED]
Sent: Friday, September 15, 2006 9:53 AM
To: Jeff Lightner; veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] Backup through firewalls
Please post the iptables information. We are adding Linux to our environment
and that information would help.
Thanks
-Original
/etc/sysctl.conf is the correct file.
You can learn more about it by reading the man page for it and also the
one for sysctl.
The available parameters are in /proc/sys/kernel. You can cat each of
these to see current value (assuming they are using defaults not shown
in sysctl.conf).
There are
Title: Stuck tapes survey
Assuming robtest works thats how we
do it.
I have seen occasions where physical
intervention was required so we had to have an operator or someone power cycle
the drive (or even fish the tape out). No matter whether it is robtest or any
other way its important
Title: Stuck tapes survey
That was so damn simple its never
occurred to me. I always end up doing physical intervention if the master cant
talk to it. Hopefully Ill remember this next time. (Knock on wood
havent had to do this in 3-4 months after having many issues for a few
before that.)
My bad. If I'd been paying attention I might have mentioned chatr.
FYI: Its not necessary on 64 bit HP-UX. What class of server are you
running the 32 bit on? Most servers since the HP 9000 K400 have PA-RISC
2.x processors. PA-RISC 2.x are all capable of running the 64 bit
version of 11.11.
to use them for NBU but did have to use them for Oracle
before it had a 64 bit version.
-Original Message-
From: bob944 [mailto:[EMAIL PROTECTED]
Sent: Tuesday, October 03, 2006 12:21 AM
To: veritas-bu@mailman.eng.auburn.edu; Jeff Lightner
Subject: RE: [Veritas-bu] Restoring a 1TB
Title: Exclude Examples for Windows
1) On Windows how does one setup exclude lists for specific policies? (On UNIX one just creates text files in /usr/openv/netbackup.)
2) Are there any general Windows Excludes (Win 2000, Win 2003, Win NT) that are recommended?
3) Are Windows Media
We use RMAN for a backup of a smaller (300 GB) DB. It doesn't require
stopping the database.
For a large (2 TB) Oracle DB we do the BCV thing outlined below. You
don't have to run Oracle on the media server because you do a standard
rather than an oracle backup. (E.G. do not use RMAN - just
Assuming youre using the oracle
backup type rather than the standard. If so Ive never
seen it work from the master but always from the client but admit that Ive
never delved into it since I could schedule from the client (and usually wanted
to anyway for RMAN and SAP backups).
Title: Message
Of course it depends a lot on the age and
class of the equipment. My RP8420 with over 10 TB attached boots much faster
than one of my old N Class machines with only 2 TB attached. Also he says HP-UX
without mentioning version. 9x? 10x? 11? 11i?. Itanium box or PA-RISC?
it wasnt
meant as an attack but rather to let him know it depends.
From: Hampus Lind
[mailto:[EMAIL PROTECTED]
Sent: Thursday, October 12, 2006
1:27 PM
To: Jeff Lightner; 'WEAVER,
Simon'; 'ZIMMER, RANDY K [AG/1000]'; veritas-bu@mailman.eng.auburn.edu
Subject: SV: [Veritas-bu] HP-UX
Media
PROTECTED]
Sent: Friday, October 13, 2006
5:28 AM
To: Jeff Lightner; Hampus Lind;
ZIMMER, RANDY K [AG/1000]; veritas-bu@mailman.eng.auburn.edu
Subject: RE: [Veritas-bu] HP-UX
Media Server to boot up
Guys
This is
slightly off topic of NBU, as it does not sound like this is the problem
In a prior incarnation I worked at a site
that did backups of Sybase on raw devices. Rather than backup the raw devices
we used SQL Backtrack software to do the backups. Not sure if it is still
available but I believe it (like RMAN) can be run with the DB on line. To have
a good point in
Ports are at the end but you need to modify NBU itself to let it know you only
want to use these 2 ports (otherwise it will open random ports):
Open Netbackup Java GUI
Go to Host Properties
Go to Master Servers
Double click on the master server
In Master Server
Title: Have to get rid of STK L180
Donating to a local college or university
might work. If the unit still has value on the books your company can probably
use it as a donation for tax purposes.
From:
[EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of Johnny Østergaard
I'm with David on this. You changed it on the server but did the
associated switch port get changed as well? Both sides should be hard
set to 100 Full.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Clooney,
David
Sent: Thursday, November 02, 2006
Finger pointing?
Shared memory and Semaphores are interprocess communication (IPCS). The
parameters are set in the OS even though it is typically the application
that requires the setting.
Veritas (Symantec) would rightly say that it is up to Sun to say HOW to
set the parameters. Sun would be
Your first unload is probably failing because it is still in the process
of mounting the tape. After you move a tape in wait a few minutes
before trying the unload.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Mike
Jackson
Sent: Monday, November 06,
Title: How Netbackup handles NFS
You would get anything under /data/iwmnt
EXCEPT /data/iwmnt/default as the latter is an NFS mount that you told it not
to follow.
You tried to restore main/internet/x/WORKAREA/2006
from under /data/iwmnt/default so of course it is 0 bytes as you dont
We do suspend/resume fine here on 5.1.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of WEAVER,
Simon
Sent: Tuesday, November 07, 2006 10:36 AM
To: 'Bobby Williams'; Michael Barrow
Cc: veritas-bu@mailman.eng.auburn.edu
Subject: Re: [Veritas-bu] NBU 5.1
We actually do dual (or more) HBAs to SAN so we'll have redundance for
disks (EMC Clariion/DMX) so can also SEE the tape drives on multiple
paths but only use one path for tapes. Having your tape drives stay on
line isn't quite as critical as having your disk drives do it so we've
never seen the
It was true of 32 bit. The limit was actually less than 2 GB. Even
without that if the application itself isn't compiled with largefiles
support it would hit this limit. Since most Linux is still 32 bit the
NBU for it may not have this support.
Of course the real issue here is that
Getting list of files in a given backup set use the bplist command:
Example: To see all files in the policy ATUBKS-1-ERPDB-DMXBCV that were
backed up on 12/04/2006:
bplist -k ATUBKS01-ERPDB-DMXBCV -s 12/04/2006 01:00:00 -e
12/05/2006 02:00:00 -R /
The -k is policy, -s starttime, -e
Here we use the public for some backups and the backup lan for others.
We don't put the IPs for the backup LAN in DNS - we just add it to the
/etc/hosts file on the master or media server we use as the host for
backups.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
HP when it sent notification for the OS patches also sent notification
for the need to update Java including a TZupdater tool for HP-UX.
Surprising they'd have a way to do the update and Sun who wrote Java
wouldn't.
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On
components are
only
the Admin GUI and the Backup Restore GUI. The rest of NetBackup is
distubuted in binary form.
That being said, I've always kept our JREs up to date, especially on
those
mission critical servers.
-- M
Jeff Lightner [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
01/10/2007
1 - 100 of 512 matches
Mail list logo