Client
W2K, TSM Server 5.1.7.0 on AIX 2 years ago.
We solved the problem by having users renaming directories and applying
stricter naming conventions.
Regards,
Markus Engelhard
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat
Hi,
using a NetApp-box: do you see the drives with a sysconfig -t and sysconfig
-m? You might also be stumbeling over ndmpd-authorisation on the Netapp. We
saw a default of 16 characters for the password...
On the NetApp get the WWN of the drive with a 'storage' command (not sure,
could be 'storag
Hi TSMers,
we are currently planning to move some database-heavy backups from an
3583-L72 with LTO2/SDG to 3583 LTO3 using TSM for SAN. Looking at
http://knowledge.storage.ibm.com/servers/storage/support/hbasearch/interop/hbaSearch.do
for HBAs to use with tis configuration, it does not seen to be
ANE4968I (Session: 100425, Node: xxx) Objects
compressed by: 0%%(SESSION: 100425)
07/29/2005 11:19:29ANE4964I (Session: 100425, Node: xxx) Elapsed
processing time:15:19:19(SESSION: 100425)
Currently running an oldish TSM Server 5.2.4.5 on 3466-C01 4GB ram.
regards,
Markus
Hi Andrew,
IMHO, IBM SCSI libraries such as 3583 need the IBM tapedriver, not the TSM
driver...
Checkin with search=yes is quite straightforward once the cartridges are in
the library (Open door, take out elephant, stuff cartridges into slots,
close door, wait for automatic inventory, checkin tape
policies, the user might need additional rights to manage security logs.
The CAD will also have to run under this account, so you will have to use
the web-interface for restores. Sell this as a security enhancement, as
local tampering can´t access your data.
Have fun with your domain admins :-),
Markus
Hi,
why not try
select stgpool_name as "Primary Pools", -
sum(num_files) as "Objekte", -
sum(logical_mb) as "MB" -
from occupancy -
where stgpool_name in -
(select stgpool_name from stgpools -
where pooltype='PRIMARY') -
group by stgpool_name
select sum(num_files)
Hi Mark,
try
/opt/IBMtape/tapelist -l
if you have multiple libraries, you can sort them out with
/usr/lpp/Atape/samples >mtlib -l /dev/smcX -qD -f /dev/rmtX
(substitute the X)
This will give you a martrix to keep track of the devices
have fun,
Markus
--
Diese E-Mail enthält vertrauliche und
Hi Mark,
I love this one. It potentially goes right deep down into the core of
priviledges and could be discussed forever.
My Idea would be to have the user use the WebClient service which could run
under a domain user account with the required privileges Doug
has pointed out (e.g. Backup Operator
Hi Geoffrey,
we had a similar problem after downloading code for Solaris. After ftping
the code from Windows (where we have web-access) to Sun the install would
fail. In Unix, EN_US is not en_US. Pardon me if our workaround looks
somewhat primitive, but we just renamed the fileset an installation
Hi Nancy,
Jurjen is right: just look at the kernel version to identify the linux
version involved. The standard TSM-Client plays nicely with the ESX host.
To back up the ESX-guests, you can choose according to your needs:
running ESX-guests on NAS, you can snap the guests (very fast and very neat
Hi Michal,
one more way would be backing up the TSM-database to a second TSM-Server
using virtual volumes. Setting up a TSM-Server on a workstation is a quick
task, and setting up the vv is rather straightforward. If you take a
machine that is backed up anyway, it will even be free of cost, as the
Hi all,
just my two cents on file-deviceclass: I agree with Kelly that Volumes
should be predfined to avoid fragmentation. In addition I would very
strongly recommend running the define volume command with a wait=yes to
format them sequentially. While this looks a pain, volumes will be
formatted w
the
archive retention of your default management class until you have checked
the archives table. Data will have to be moved as pointed out.
If your aim is just to collocate, why not hav a close look at collocation
grouping? Great effect at very little cost.
Regards,
Markus Engelhard
--
Diese
Hi Murugan,
sorry for the late response.
On W2K3 and SAN Drives I rather love the tsmdlst-utility with the flag -all
which gives you a good Idea of the library and drive state, including
fw-levels . Also check if the Library and drives are managed by the
IBM-Tape drivers. Last but not least deact
Hi all,
just a quick one we just stumbled upon: running incremental/full dbb using
following script:
***snip***
select * from db where LAST_BACKUP_DATE is NULL
if(rc_ok) goto ba_full
select * from db where num_backup_incr=20
if(rc_ok) goto ba_full
if(rc_notfound) goto ba_incr
ba_full: backup db d
Hi all,
thanks Ramakrishna, there was no extention or reduction on the database
since the last full, in fact, we had 13 successful incrementals before the
failure.
Andi, I´ll take a second look at the script, although it has been working
for over a year on 10-odd TSM-servers on Windows2003 an has
Hello Richard,
thanks for the hint, the incremental dbbackups have run daily, adding up to
13 incrementals in the series in question. Could backup_chg_pct equal or
greater than 100 lead to a full dbbackup being requested? I did´nt find
anything in the docs...
Regards,
Markus
--
Diese E-Mail en
Hi *sm ers,
I´m presently trying to consolidate 20 odd distributed TSM-Servers that
were indiviually managed and have very different logical setups. My aim is
to have identical policies across the country, so I thought having
standardized policy domains would be a good idea.
It´s quite straightfor
running the share, and presto, it worked ( Would make me happy). I'm afraid
I won't be able to persuade him to give me this privilege on the production
machine though.
Has anyone come upon this problem, and is there any circumventions. If
anyone could share$ his expertise and thougth
Hi TSMers,
we are looking at possible Platforms for TSM. Is there anybody out there
willing to share experience (online or offline) about TSM on zLinux? I
don´t seem to be able to find much about it on the list.
Mit freundlichen Grüßen,
Markus Engelhard
IT531 Storage & Recovery
Deut
t´s just a roumor...
If not, I would be highly interested where to find more. It´s still in the
list of supported devices:
http://www-306.ibm.com/software/sysmgmt/products/support/IBM_TSM_Supported_Devices_for_AIXHPSUNWIN.html
Mit freundlichen Grüßen,
Markus Engelhard
IT531 Storage & Recovery
D
Hi Richard,
some very long time ago, we hit the need to use two policy sets in
environments with two physical tape robots (old 3583, remember them?),
small disk pools and large backups. Backups would run up to 24 hours to two
tape devices, not enough drives to allocate for synchronous writes for
t
Dear Charles,
use the Config Manager approach, it makes life pretty easy and is
lightweight. We also had situations in the past when server to server
communication sucked, update server xxx forces=yes on all servers fixed the
problem most of the time, just script and run if needed.
Cheers,
Markus
Just a comment,
while using nfs datasores does have some other limitations, cbt will only
have the drawback that the initial backup will explode thin provisioned
partitions on the initial backup. So take care to have a destination for
the data that will compress on the fly, we get approximately 1:
Dear all,
we are currently struggling a bit with TSM4VE config and making it more
manageable regarding including and excluding vms from backup.
For disks, we can exclude individual disks by just adding a line to the
optfile looking like
exclude.vmdisk "vm1" "Hard Disk 2"
exclude.vmdisk "vm2" "Har
Dear all,
thanks for the feedback and great solutions! @Del, that was exactly the
RFE, so the vote has increased by 1. Should be doable with limited
resources, as it is displayed as a list in the GUI and the syntax is there
as in exclude.vmdisk, so please dont´t reject it.
Using vmfolders is a gre
zz\UID*" -ina -pick C:\temp\
...
Beware you will still need some domino instance to open the data.
Kind regards,
_________
Markus Engelhard
Dipl. Min.
Deutsche Bundesbank
Zentrale
Produktmanagement Infrastrukturen
Wilhelm-Epstein-Straße 14
60431 Frankf
Dear all,
the combination ISP for Enterprise Ressource Planning seems not to support
SAP in combination with Oracle12. Anybody an idea when official support
will be added?
Kind regards,
Markus
One more on the topic: the admin guide has always been a great starting
point for a one-stop doc to set aside printed documentation for a true dr:
printed out as an addition to the home-grown documentation and available
without internet or local file access. Apart from this it has also served
as a
Hi,
has anyone been able to use ssl with the cert256.arm? The ssl connection
using cert.arm runs fine, but we would like to move on to the cert256.arm.
We tried it on a Solaris10 server with ISP 7.1.4.100 and on a Windows 2012
Server with ISP 7.1.4.0, but have not been able to connect the admin
in
Thanks David, Bjoern for the great hints!
We are testing the self-signed certificates created by the server instance,
so at this stage, no third-party certificate is involved, using the dsmadmc
native commandline (no hub-server, no oc, no server-to-server, all down to
absolute basics. Using gskit
Dear experts,
Checksum is a great idea! I opened a RFE a couple of days ago not being
aware that IBM does deliver a checksum (just does not state which ;-()
when downloading from the Passport site via http. One more enhancement
would be to deliver the checksum via a secondary transport method e.g.
Hello Andreas,
sorry to hear tat this sort of thing is still happening this century...
Have you had a look at container pools and dedup? It would ease the
storage-side pain quite a bit, with drawbacks in taking copies to tape.
I have been testing the dedup + compression feature, an apart from hand
Dear all,
we have been testing the tagging option and it may be a way to go in
certain environments, but this moves the control to the vm-admins with no
clue whatsoever to what is really happening backup-wise.
Would putting it in the client optin set be a good idea? One line per
object, centrally
Dear all,
we have been thinking about a local quick fix for Security Bulletin: Buffer
Overflow from improperly formatted SELECT command in IBM Tivoli Storage
Manager (IBM Spectrum Protect) Server (CVE-2016-8998).
Gerd Becker and I have come across the following Idea that may ease the
preasure to u
Thanks to all,
Andy, thanks for the suggestions regarding the script and thanks to all the
others for suggestions. Thanks Richard, I too am anxious to have the
ability to recover my database, in fact it would be nice not to loose it in
the first place. But it does happen and it will happen, I have
Hi Doug,
querying the backups- and archives table has been discussed to some extend
in this forum. It is certainly NOT a query I would issue as a daily report,
as it is VERY time-consuming and puts some strain on your TSM database; be
ready to wait quite a while for the output; DON´T issue it if y
Hi Geoffrey,
why not just create a profile on the Server with the scripts in it and then
create a subscription on the TSM-server you want the script on? This way
you can keep scripts neatly aligned with almost no administrative or manual
overhead, as changes are replicated automatically. Works wel
Hi Kevin,
I like this one...
Rebinding will not cure the inactive file with no active file problem, nore
will archives be rebound ar correctly pointed out. So there are basically
two options left:
a) Create another TSM-server instance with one donain an all that with one
default management class
Hi Zoltan,
as I understand, you will be changing OS for all your TSM-instances from
AIX to Linux. There is no supported way to just simply make TSM data
written by AIX readable by another OS apart from painstakingly
exporting/importing the node data and database references.
My best guess would thu
. Keeping
clients with little data change on disk also saves mounts and tapes if you
run collocation, which is in my eyes vital for adequate restore
performance.
Regards,
Markus Engelhard
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige
Hi Zoltan,
been there, done that cross-platform AIX/SUN/Wintel. It worked beautifully
for us, just define libraries and drives/paths to all desired ressources.
Check in tapes for each LIBM, actually we just checked scratch tapes out of
the retiring TSM with remove=no and checked them into the new
Hello Brenda,
basically you need a user (prefeably a domain user) with backup operator
rights plus rights to manage security logs and run as service. Set up a
scheduler and web service on a Windows machine as a sort of "CIFS Proxy"
under these rights. In the option file, enter something like an ex
Hi Greg,
it really depends on how you are running your TSM environment. In a shared
library environment. I try to back up the library manager as frequently as
I can, just a few mb on sequential disk, takes a few seconds to back up and
restore. An audit will be needed after restore anyway to align
Hello Michael,
we are just moving from AIX to W2K3 by exporting node by node, which is
sometimes painstaking but helps clean up the system from old and obsolete
stuff. The TSM server name is neccesarily unique on each system. I
experience that the password frequently has to be updated after export
Hi Keith,
virtual volumes used to be a big benefit while bridging big distances for
DR of backup data via fc/scsi was extremely expensive.
Today, I use the feature in our multi-site infrastucture to do my
additional db-snapshots to one (central) tsm-server. No tape is involved in
this part of the
Hi Thomas and all the rest on the list,
I´ll add to the discussion with a few questions of my own, though not with
scope to a specific vendor:
How did those who are using a VTL instead of random/sequential disk size
your environment? Not talking about a disk-only solution, tape for longer
term st
Hi Paul,
we run AIX, W2K3 and Sun-TSM-Servers. A modern Win-TSM is great for TSM up
to mid-size setups and performance should not be an issue in your projected
environment. If it´s a Windows shop, stick to the os you know you can
handle. We run a couple of them with various IBM libraries including
Hello Dirk,
sorry for being late with my comments on NDMP backups. In my opinion some
of the stuff regarding reporting and administration is admittedly still
rather odd:
- the full backup will be kept until the next one is performed, but it will
no longer be reported on with the query, so don´t w
Hi John et al.,
I guess the real question is: do you want to rely on one vendor to do
primary and backup storage? The obvious risk ist to loose not only the
primary data in case of a severe bug or attack.
One more thought is: don´t rely on pit local copies (e.g. snapshots) to be
available in case
Hi Shawn,
as always, the decission should be based on OS knowledge and internal
support structures. Both OS are stable and mature for TSM, and I wouldn´t
really care (I run an odd dozen TSM on Windows and a dozen on Unix, all
rather stable and comfortable to administer). TSM look and feel doesn´t
Hi Guenther,
looks rather like a challenge, but doable. You certainly will have to
seperate this data from the remaining nodes in its own primary disk random
access storage pool. As I understand, all primary data stays on disk, so
while the solution does make the setup slightly more complex and al
nd regards,
_____
Markus Engelhard
IBM certified deployment professional - TSM 5.4
Homepage: www.bundesbank.de
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail
irrtümlich erha
Dear all,
has anyone come across the following eventlog error message while backing
up a W2K3 client?
Volume Shadow Copy Service error: Unexpected error calling routine
MoveFileExW
TSM backups work, nothing in the TSM error log or schedlog. No change,
client has been working happily for quite a
Dear Wanda, dear Tim,
thanks for the feedback. OFS is required for a couple of systems for BMR. I
fully agree that it does not take care of all aspects of recovery and
databases are a case to be taken care of with great caution in regard to
consitent backups. For this part of recovery I prefer to
Hi Fred,
export a node (no data) with the cloptset associated to it, the cloptset
gets exported as well.
Bit of straightforward with a sidestep (dance the chacha).
Regards,
Markus
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Ad
Hi Stefan,
I guess there is a developer works article on this topic, eg.:
http://www.ibm.com/developerworks/wikis/display/tivolistoragemanager/Tivoli+Storage+Manager+Version+6+Hybrid+Upgrade+Migration+Method
Upgrade can be achieved with a downtime of less than an hour when using a
new system or b
Dear Zoltan,
there is no need to net use the UNC shares. Just define the share in the
domain statement and in the objects field in the schedule. Don´t forget to
exclude the paths for snapshots! The TSM scheduler must run on an account
with priviledges of backup operator and manage security logs, d
Hi folks,
sorry if this is a bit special, but I hope to find some opinions for this
parameter. We are planning to move our tape environment to a clustered
Eternus CS. Environment is TSM is 5.2.2.1 on Sun Solaris 10.
We will still maintain a limited "landing zone" on SAN disk for all those
tiny chu
Hi Thomas,
allmost been there (only have 10 years retention for this data). Initially,
we retired an oracle database on NT4 (don´t remenber the oracle version,
current in 1999/2000) to our old TSM 3.7 on a mainframe. In the meantime we
migrated TSM to AIX and the to Sun, with two changes in tape t
Hi Erwin,
rather straightforward:
SERVERNAME stanza alias for node in dsm.sys in unix, one
stanza per node
NODENAMEname of the TSM node as defined on the TSM server
TCPSERVERADDRESSIP or DNS name for the TSM server
COMMMETHOD TCPIP
TCPPORT
Sorry for being late with this. Disable session and starting the server
with nomigrrecl (TSM instance has to be restarted) usually works. Can be a
pain in the ... in a large Library manager/Library client environment to
get the appropriate downtime for the Library for all library clients. Don´t
for
Hi Richard,
more of a theory than anything I have done myself, strongly depending on
what you actually do with your Library manager. Supposing it serves solely
as a libm, no NDMP, no storage agents native to the libm, no config manager
on libm, library clients have not been upgraded:
Why not come
Hi Del,
is there any movement into adding support for Domino, e.g. running on
N-Series? I don´t think this is an exotic configuration. I would be glad to
get rid of my home-grown scripts to achieve the offload backup via proxy
node.
Kind regards,
Markus
--
Diese E-Mail enthält vertrauliche un
Hi Del et. al.,
sorry if my wording was not clear. My problem is:
Flash copy manager is really a great approach to ensure fast and consistant
backups of large environments. Integration is so far into the VSS-world of
Exchange and SQL as well as into the upper midrange to enterprise IBM disk
storag
Hi Zoltan,
my experience has been: use fixed size preformatted volumes, and be sure to
format them sequentially, even if it seems to take a hell of a time. But
then, it´s a one-time action and highly automated, so just don´t try to
boost "performance" here. Make sure no one else is bogging perfs,
Hi Stefan,
does diskshadow work? I just played around with a case where I had to
register several dlls according to a MS article I found, basically you do
cd /d %windir%\system32
Net stop vss
Net stop swprv
regsvr32 ole32.dll
regsvr32 oleaut32.dll
regsvr32 vss_ps.dll
vssvc /register
regsvr32 /i s
Hi Andreas,
let me guess: data format would be native, should be nonblocked. No Problem
with the primary pool and move data, but you could be stuck with the copy
pools. With SSAM activated, there is no easy way to move the data, as you
can´t change the settings for an existing pool and move data t
cite:
...I do this by creating a virtualfs called "Filesystem_01YEAR" and
point that to the same file system...
Brilliant Idea, Shawn!
I really love the elegance of your solution, which totally avoids tinkering
with nodes and such, potentially adding confusion and complexity. Thanks!
I have bee
Hi Andy,
been there, done that, for years: no isues. In fact we ran (we have now
consolidated somewhat) TSM servers all over the country spanning approx.
800 km, using a low-level cheap box for the config manager job.
Config does not need much bandwidth and really saves a lot of time in
rolling ou
Dear Mehdi,
apart from Grigoris arguements, I would add:
- no additional repository is needed for backups and images
- very easily standartized using TSM (e.g. clopts)
- near automatic recovery using predefined Answer-files
- no additional load on the system
- no additional admin tasks involved af
Hi Del,
does that also apply for TDP for Domino, I certainly will stay tuned!
See you all in Dresden for the TSM symposium:
http://tsm2011.uni-koeln.de/10134.html
Markus
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat sin
Dear all,
I hope it is not inappropriate to announce a vacancy for a TSM related job
on this list. If so, I appologize.
After a year off I am back into TSM and on the list and can offer a
permanent position in the public sector for our backup and recovery
planning and support team based in Frankf
Dear Fred,
if you don´t need the TOC, I would go for defining the datamover and paths
directly to the library manager, there is virtually no impact on the log
and database consumption and thus the startup speed. If you want to do
filer-to-server backups with TOC, this will go into the storage hira
er is here? Normally it would be more than a HACMP or likewise
failover to declare desaster, time enough to have someone push the button
to run the scripts.
Kind regards,
_________
Markus Engelhard
Dipl. Min.
Deutsche Bundesbank
Zentrale
Produktmanage
Hi Eric,
the process is definitely single-thread, we learned the hard way on a
T5240: (one 1.4 GHz thtead out of 64 busy, all the others idle). My way
would be to have a look at the hybrid upgrade method in develloperworks,
the basic steps would be rather similar (just not the upgrade), looks
rat
Dear Amit,
looks like you are archiving the same tons of files and tons of data again
and again. Why no use backups to backup copygroups (same Management class
as for archives) set to versioning NOLIMIT, as to make them keep data just
like archive copygroups. Then set up incremental backups, which
Hi John,
have you tried "update server blah forces=yes" from both sides? Might do
the trick.
Regards,
Markus
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail
irrtümlich erhalten haben, informieren Sie
Hi Roger,
according to my infos, the vulnerability is reported in versions 5.5.0.0
through 5.5.4.x, 6.1.0.0 through 6.1.5.x, 6.2.0.0 through 6.2.4.x, 6.3.0.x,
and 6.4.0.0.
Regards, Markus
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richt
Hi Zoltan,
have a look at the symantec alerts site and secunia. 6.2.5, 6.3.1 and
6.4.0.1 should be fixed, haven´t found much for Solaris though. Workaround
according to IBM: run scheduler in prompted mode (have real fun) or
manage it by cad.
Regards,
Markus
--
Diese E-Mail enthält vertrau
Hi Wanda,
not sure about the cluster stuff, but why not try dsmcutil to change the
stored password?
Regards,
Markus
--
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail
irrtümlich erhalten haben, informie
Hi Steve,
we just went through this for a migration from Windows TSM 5.5.1 to Sun
with TSM 6.3.3. It worked flawlessly, even aligning the storage agents
during the move. We had plan b in case of errors because of support issues
(even starting from scratch would have been feasable), but no problem
83 matches
Mail list logo