you might want to try using the dd
command on AIX with a large test file as David proposes.
something like
#dd if=/test_dir/test_file of=/dev/rmt0
bs=256 /* bs=block size expressed
in factors or multiples of 4096. 256 should be the default LTO blocksize*/
in this way you should get the OS
You probably need to run a Cleanup Backupgroups process.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
Hi,
I wanted to candidate myself for the
residency.I know that I'm a couple of days late, but I only just got my
first and second line manager's approval.
can I submit through you or some other
way?
thanks.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53,
TSM 5.3 will support AIX 5.3.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
PROTECTED] Behalf Of
Gianluca Mariani1
Sent: Tuesday, August 10, 2004 4:42 AM
To: [EMAIL PROTECTED]
Subject: Re: AIX5.3 and power5 for pSeries
TSM 5.3 will support AIX 5.3.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
How about:
Select * from volumeusage where node_name='nomenodo' and
stgpool_name='nomestgp' and filespace_id= fs#
?
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL
Just a small addendum to Christo's explanation (I can subscribe writing my
signature in blood to it):
this issue about compression generates a lot of confusion because there's
no standard commonly accepted viewpoint. one thing is looking at it client
side, a totally different thing is looking
see below.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
The
see nothing
-Original Message-
From: Gianluca Mariani1 [mailto:[EMAIL PROTECTED]
Sent: Wednesday, March 31, 2004 4:26 PM
To: [EMAIL PROTECTED]
Subject: Re: Gen1 to Gen2 Upgrade/coexist
see below.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53
Maybe this is not the best place,
but I wanted, together with the other Tivoli GRT members, to express our
deep sadness for the 11 march events in Madrid.
for what it's worth we feel close to you and your pain of today.
coraggio fratelli.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global
I will be out of the office starting December 8, 2003 and will not return
until December 31, 2003.
risponderò ai vostri messaggi al mio ritorno.grazie.
reason being AIX 4.3.3 will be going out of support at the end of the
year.it's in the server readme anyway, first paragraph.
so you'll need AIX 5.1 or later, 32 or 64 bit.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
you should use the lb and mt definitions, and the native device drivers for
the drives which I think you are saying you do use.
you can check this, in w2k, in the management console under tsm device
drivers-reports-device information.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response
try:
on TSM
compressionno
TCPwindow 128 or 256
TCPbuffsize 32
on aix try to enable tcp_nodelayack. this is APAR IY31983.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
just a thought.
before putting the problem to the HW devices (tapes), SW backup programs
(TSM or whatever), and looking for a technology solution, I would ask
myself exactly *why* I have a 300+ GB file.one file, 300GB? is there
anything I can do to bring it to an acceptable size? how am I
Zlatko,
the last two months meditating in a monastery in Tibet have been very
helpful...
IBM employees who participate to this list cannot but be grateful to all of
you who understand why we do it.
it's a reward that falls outside of business metrics.
;-)
Cordiali saluti
Gianluca Mariani
Tivoli
I tend to think of TSM as a troubleshooter by now.
in my experience, in what is probably the majority of real world cases I've
been in, anything that happens in the environment in which TSM sits,
automatically reflects back to TSM which is the first application to
suffer.before anything else.
we
Progressive incremental backs up only new or changed files. during the
initial backup the client backs up all eligible files of course(full
backup). Subsequently, files are backed up again only if they are new or
have changed since the last backup. In TSMs case, a pointer to each version
of
What is EndTxn Verb? It takes 50% of the clients time.
- it indicates the end-transaction verb time on the server.that is
the time it takes to close the transaction.the fact that in the trace Data
Verb is high as well points to a problem on the server.
Cache Hit Pct.: 98.25
we'd like to
4KB
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
The Hitch
pour savoir dans quelles cassettes est divisè un volume vous pouvez
exècuter la commande
restore volname preview=yes wait =yes
il n'importe pas s'il s'agit de volumes offsite ou onsite. il faut se
rappeler, quand meme, qu'il peut se passer que TSM va ècrir sur des autres
cassettes (parce qu'il
A couple of things: the cache hit % is too low. we'd like to see 99%
ideally, but 90% is bad performance. and cache wait % should be 0
ideally.one reason could be the many volumes you are using, and the fact
that Veritas filesystem does introduce an overhead.
bufpoolsize should be normally around
have you tried Journal Based Backup? it looks as if you can still take
advantage from this feature, since you need to backup ca. 5% of your files.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
my feeling is that Resource Timeout will not change the situation.it should
help in giving insight in the lock status.
I am seeing this elsewhere but we have not yet found out the culprit. at
least in EMEA.
one workaround with the NT scheduler is to use AT commands to stop/start it
remotely if you
if you're not sure where the problem lies, as a first thing run a PERFORM
trace on the client to see where things are getting bottled up. if you
don't know how to interpret the result (or how to run a trace), look in
TSM's class materials. if you haven't followed a TSM administration class,
what combination of OSs are you running on server and clients? and what's
running on 32 bits and what on 64? did u upgrade just the server or the
clients as well?
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
John,
I cannot relieve your pains a lot. you are right basically. I can only say
that, as a rule of thumb, we generally consider the audit db to go through
1GB per hour. but this is really a rule of thumb and I suppose lots of
people will jump to my neck for giving this figure. it really depends.
respond
to ADSM: Dist
Stor Manager
Can anyone please explain the disadvantages of using journal-based backup?
-Original Message-
From: Gianluca Mariani1 [mailto:gianluca_mariani;IT.IBM.COM]
Sent: 25 October 2002 15:47
To: [EMAIL PROTECTED]
Subject: Re: Very
I would just like to add something. the manuals are sometimes misleading.
ESS runs at a theoretical speed of 1280 MB/s yes. that's because you
count the fiber channel connections towards the outside. but the internal
busses are 4 133MB/s PCI busses. that makes a max throughput from the cache
to
I would recommend to wait next week for the 4.2.3 server release.
it's a very stable version and fixes many issues. a lot of work has gone
into it.
the guys from developement have done a really good job.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53,
Placing TSM DB volumes over multiple physical volumes can improve
performances because the DB is read-oriented, the LVM will spread the load
over the various volumes. if the number of volumes is too high, though,
your performance will be hit by the overhead on the Logical Volume Manager.
the
I will be out of the office starting August 30, 2002 and will not return
until September 16, 2002.
you can try to reach me telepathically.
Matthew,
can I propose you for Customer of the year award? It's difficult to try
to get this across, but support is more complex than it seems...
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
when you upgrade your licences must be registered again. have you done
that?
see pages 22 and 30 of the Quick Start manual.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL
they are not supported any longer.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
TSM 5.1 supports Netware, but not Netware clusters.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
Guillaume Gilbert
Chris,
TCPWINDOWSIZE should be set to 63 on Win2k.
apart from that ,have you tried
TxnGroupMax 256
TxnByteLimit 2097152(for BU direct to LTO or DLT)
TcpNoDelay yes
and maybe you can take an instr_client_detail trace and take a look at the
DataVerb figure, see if it's
Our official recommendation is for 63. IBM's position on a value 63 is
Since Windows 2000 supports TCP window scaling, it may be beneficial to use
a larger TCP window size for Windows 2000 systems that are communicating
exclusively with other Windows 2000 or UNIX systems.
Cordiali saluti
]]On Behalf Of
Gianluca Mariani1
TCPWINDOWSIZE should be set to 63 on Win2k.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
You need to check your documentation. Windows 2K machines should have
TCPWINDOWSIZE set to 64 (assuming 10 or 100Mb Ethernet).
I'm somewhat
just to point out that legal requirements are only met, generally, by
ablative WORM media. that is, physically unalterable media once data has
been written. not tapes.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
the error numbers are AIX errno. look them up and see if that gives you
some better idea.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
the H drives are not GA yet. TSM already supports them. a readme will be
available on the website.
Cordiali saluti
Gianluca Mariani
Tivoli TSM Global Response Team, Roma
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
John,
I'm not sure I understand your question. are you saying you want one
machine in a cluster to be registered to one policy domain and the other
machine in the cluster to be registered to another policy domain? you would
not be using a cluster this way, what would happen in case of failover?
this is from the tsm 4.2 Technical Guide Redbook (SG24-6277-00); it should
be your case, if you have W2k TSM servers. otherwise refer to the library
sharing paragraph in the same redbook.problem is, this way you can't
Disaster Recovery; if you need two different physical environments, things
get
16 or 32kB
Cordiali saluti
Gianluca Mariani
Tech Support SSD
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
[EMAIL PROTECTED]
Paul Zarnowski
[EMAIL PROTECTED]To: [EMAIL PROTECTED]
IBM LTO is, objectively, the best performer in the field. the drives are
more robust, have much fewer mechanical or firmware problems and are, in
general, very reliable. the 3584 and 3583 are very solid, performing
libraries. it is not by chance that in the LTO market IBM is doing so well.
even
but it's 13k objects having to be inspected for the incremental. only 29
have actually been backed up. maybe some more RAM or CPU could lower that
1:45 hr figure. one would have to look at the client options to see if any
setting can be changed. as said, the 47 secs are the time it took to backup
Wolf,
this sounds like a communications problem from exchange. it could be useful
if you run microsoft's BackTest utility that's in the exchange SDK. that
should tell you what the actual exchange return code is (425 is, as William
shows you, the TDP return code, not Exchange's) and clear up if
the way I know it, 5.1 does support Win9x client side. the web site is
correct. haven't tried to mix it up with a 4.2 server, but haven't heard of
problems there yet. what exactly is the generic error you get?
have you tried doing a q actlog to see what the server says?
Cordiali saluti
Gianluca
I'm not sure what is meant by library sharing here. the way I know it is
kind of like Mark here says: you can share a 3584 in the sense that, since
the library has multiple control paths you can partition it between
heterogeneous OSs (which is what you cannot do with the 3583, being that
single
Francisco,
those numbers are normal. 3590 drives have a native transfer rate of
12MB/s. that makes around the 50GB/h mark. to have more than that you have
to compress your data, and the 3590 can compress up to 3:1. whether you can
do that or not depends on your data (3:1 compression ratios are
Fran,
one parameter you can change is Bufpoolsize on the server. for 1GB of real
RAM you should put it to 131072. the other thing is to turn compression
off, in a single client, fast client, fast server, fast network it gives
better performance.
then on the AIX server try
no -o rfc1323=1
no -o
I agree with what Zlatko and Juraj are saying. I just want to add that in
disk subsystems such as ESS that do use RAID5 a lot of care has gone into
the introduction of cache destaging algorithms that mitigate the write
penalty. that is, writes are held in cache until they can be destaged
Hi Gerald,
I think that the answer to your question really depends on how your disk
subsystem works and spreads volumes through the RAID arrays. the object of
the game is always the same: spreading data as much as possible on all
disks in the subsystem and try to parallelize everything. in my
Fabio,
I don't know if I understand correctly your question, but if you want to
see log files and traces on a TSM Client machine (if those machines are
servers for other applications TSM doesn't care) you would need these
files:
dsmerror.log
dsmsched.log
dsm.opt
dsm.sys (if unix)
TSM server
Hi Lisa,
the main points in the AIT vs LTO contest to me are:
1. AIT is a proprietary format dveloped for the then niche market of
digital media. It is true it has faster access times than LTO. this is,
though, just about the only advantage it can count over LTO. AIT-2
cartridges have 50GB
I would use the IBM LTO drivers for the drives and TSM drivers for the
medium changer. then check the SCSI chain and be sure of which drive will
have element 256 and which element 257. there may be a switched definition
as Kelly points out.
then delete Devclasses and Stgpools and redifine
sorry, I was forgetting about AIT-3. my mistake.
this is a better option, though it has a smaller data buffer than LTO.
with the ADIC libraries you get good capacities, but still transfer rates
for AIT-3 are slower. but maybe in your case, with many small files it's an
option.
I would still go
58 matches
Mail list logo
