Matthew,
As Richard Sims said, the system may be waiting for a device reply that is
never going to occur. If this is indeed the case, then you will probably
find that you cannot bounce the tsm server instance - it will have a
driver thread in kernel, and will never end - and you can't kill a zomb
Guys,
One of our Domino admins has asked the following. Does anybody know if the
current TDP or future versions of the TDP will support this Domino
functionality.
1. Domino supports a shared mail system (also known as Single Copy Object
Store, or SCOS) - in which documents are transparently share
David,
Do you have the en_US locale installed (SUNWnamos and SUNWnamox, I think).
If so, try setting the en_US locale before starting the server:
LC_ALL=en_US
export LC_ALL
Regards,
-=Dave=-
Sungard Vaulting Services
+44 (0) 20 7608 7140
No matter how bad things get, there's always a new low j
Eric,
Is the database being restored very large? SQL server can take many hours
to recreate and initialize the database tables it has asked for the tape
but *before* it actually starts reading the tape. In these cases, we have
had to set our server timeouts to 4-6 hours during the period of the
Hello all,
Does anybody have any experience with TSM performance of STK 9940's ?
We are running 9840's with colocation and have found the performance for
large restores of a small file nature to be very good.
The only problem with 9840 is the 20GB limitation. I know that large
servers will span mu
Hi,
In the past we too have seen problems with the scheduler daemon not
honouring a nohup under Solaris, HPUX, and AIX. We have found that using
"at" on unix-type systems causes the least problems.
So, our (ksh) startup files tend to look something like this, and can be
called during init, or fr
John,
If you are using library sharing, check apar IC37251. This bit us, but not
nearly so badly as IC37289 did!
Regards,
-=Dave=-
Sungard Vaulting Services
+44 (0) 20 7608 7140
"Rome did not create a mighty empire by having management meetings; they
did it by mercilessly slaughtering all tho
Fernando,
As far as I know the "slot table" is only initialized when the library is
defined.
Whenever we have upgraded the capacity on one of our libraries, we have
done the following:
- note any scratch tapes
- checkout all volumes remove=no
- delete all drives
- delete library
- define library
Jamie,
If you are running Solaris, (or any unix, I think), then the problem is
simple. Create a new, additional, instance on the same machine as
described in the Administrator's Guide Chapter 16 - Starting and Halting
the Server -> Running multiple servers. With Solaris, you can also browse
the
Hello readers everywhere:
Each night I run a backup of an SQL server
On this server I have an 'off-line' databases called Bob (I know it is case
sensitive)
I have added the following to the dsm.opt for the TDP:
EXCLUDE "\...\Bob\...\*"
The issue is that when the backup runs (using this dsm.opt fi
Lee,
I assume you meant 5.1.6.2.
Not as far as I'm aware. However, once the data is migrated, you could try
"move nodedata fromstg= tostg= type=ARCHIVE" as a short-term
alternative
Regards,
-=Dave=-
+44 (0) 20 7608 7140
"If my son wants to be a pimp when he grows up, that's fine with me. I
Matthew,
Have you had any interesting records in /var/adm/messages that you can
match up to when one or more of the tapes was mounted? (ANR8468I volume
dismounted is a good search key).
We have only seen this on san-attached devices, and then only when a RSCN
has occurred on the fabric. Durin
Not a problem.
Your library manager instance must be at the same level or higher than the
latest library client. Consider 5.1.7, since this fixes apar ic36335.
You will need to set up paths in the library manager for all instances, no
matter what version they are running; however only v5 library
Hi,
Several of our clients have upgraded their Domino servers to v5.0.11, and
the tdp promptly stopped working. The tdp worked fine with previous
versions of Domino on these servers. We have tried using the very latest
tdp version from the ftp site, but no success there.
The symptoms we see are
Hi,
This is a manifestation of apar IC34488.
Fixed in a ptf of 5.1.5 - may be 5.1.5.4, but I'm not sure
Regards,
-=Dave=-
+44 (0) 20 7608 7140
"I tried an internal modem, but it hurt when I walked..."
Chris,
This one is *really* bad news. The error is coming from the file expansion
section of the client.
Mark the input tape as "unavailable" or "destroyed" to force the restore
onto your offsite copy, then at a convenient point, issue a restore vol on
the input tape.
Regards,
-=Dave=-
+44 (0
Also, ensure that there is a compatible devclass on the MANAGER to match
those on the calling client. It does not need to be the same devclass
name, but should be defined identically. Without this, the any request for
scratch tapes will fail.
Regards,
-=Dave=-
+44 (0) 20 7608 7140
(A)bort, (R
Bruce,
I don't think so.
To tell, look at the occupancy object count for your sys-objects. I would
think that <= 20,000 for an average system is not unreasonable, but the
figure varies with the copygroup settings, whether active-directory is in
use, etc.
We have had counts > 1.1 _million_ for s
Michelle,
Check the "maxscratch" setting for the "next" (or subordinate) storage pool
Regards,
-=Dave=-
+44 (0) 20 7608 7140
Give a man a fish and you feed him for a day; give him a freshly-charged
Electric Eel and chances are he won't bother you for anything ever again.
Hi all,
Has anybody had a resolution to failed restores, with the lines like the
following shown in the dsmerror.log:
10/29/2002 10:02:09
The 942th code was
found to be out of
sequence.
The code (2013) was
greater than
(1197), the next
available slot in
the
string table.
10/29/2002
Christoph,
what were the functions in the trace-back? Anything like this:
08/01/2002 16:58:43 ANRD Trace-back of called functions:
08/01/2002 16:58:43 ANRD 0x000100077208 pkFree
08/01/2002 16:58:43 ANRD 0x0001006B0138 SmDoEventLog
08/01/2002 16:58:43 ANRD 0x00
The default domain for unix is all-local.In order to backup nfs mounts,
they need to be specified in a domain statement:
A statement like this should do the trick:
domain all-local /nfs_mnt1/nfs_mnt2
Of course if you don't want _any_ of the host, take out the all-local.
Regards
Da
APAR IC33270
Regards
Dave
--
+44 (0) 20 7608 7140
Qui custodiet ipsos custodes
Kent Monthei
cc:
Sent by: "ADSM: Subject: Follow-up, Re:
URGENT - ANR0102E asalloc.c(): Error 1
Kent,
This is a known error, (I can't remember the apar), such that when a volume
becomes emptied, not all the pointers get deleted, and an audit volume
fix=yes does not correctly delete them either.
It is a volume-related problem, so if you mark the tapes involved "unav",
or if they are scratch,
Dirk,
You are correct the Clustered drive is not an issue as you use a
Clustergroup dsm.opt file
What is still an issue is that to backup the NDS you must enter NDS: SYS:
VOL1: in the DOMAIN statement
As I do not have control of the Netware servers I want to backup any drive
what ever it is calle
Hi *sm people:
I have an issue with the latest Netware client 5.1.5
I have been waiting for the cluster support in the 5.1.5 client
Now it seems I can not backup all clustered drives and NDS
I see others have used nds: sys: vol1: etc
I can not specify the drives to backup in the option file becau
Hi,
Does anybody else have problems with backups which have the
resourceutilization option set to greater than one ?
When the option first became available, we experimented with different
values. However, we found that the controlling thread would frequently go
into IDLEWAIT for long periods of
Microsoft Certified System Engineer
[EMAIL PROTECTED]
Nyherji Hf Simi TEL: +354-569-7700
Borgartun 37105 Iceland
URL:http://www.nyherji.is
-Original Message-
From: ADSM: Dist Stor Manager [mailto:[EMAIL PROTECTED]]On Behalf Of
Frost, Da
try adding the parameter "-outfile" to the dsmc command. From unix..
dsmadmc -id=xx -pa=yy -outfile -tab select volume_name from volumes \
where devclass_name like \'3590DEV\'
regards,
-=Dave=-
--
+44 (0)20 7608 7140
A Bugless Program is an Abstract Theoretical Concept.
-Ori
Bill,
I am not sure whether the snmp monitor is sufficient... not that I've ever
managed to get it working in solaris (!). Nor is simply checking for the
presence of the process good enough. We've found that the server can on
occasion fail or timeout backup sessions, yet still accept admin
conn
Jim,
Thanks for that. But...
A lookup in the tivoli/support/knowledgebase/ for apar ic34015 returns 0
records. Nor is it documented in the readmes for the 4.2 or 5.1 client in
"maintenance" or "patches" at boulder.
It would seem that this apar is not publically available.
regards,
-=Dave
Dear All,
Does anybody know if there is any plan in future releases of the TSM client
for W2K, to implement a facility to not replace the hardware dependent
driver information during a system object restore.
My reason for asking ;
We have a requirement to fully restore a W2K server back to diff
Zlatko,
We use the free telnet replacement "teraterm". You need to play with the
keyboard.cnf file a bit to get it correct, since the edit keys are
mis-defined in the issued version.
regards,
-=Dave=-
--
+44 (0)20 7608 7140
No electrons used in the production of this message were harmed o
My colleagues have all installed and use the old 3.1 gui, only using the web
browser if the gui can't do something.
Personally, I've _never_ used the gui. The only time I choose the web
browser over the cli is for management-class and copygroup management, when
the cli becomes cumbersome because
How about putting the output file as part of the dsmadmc command:
dsmadmc -id=admin -pa= -outfile=myoutput -tab select * from processes
regards,
-=Dave=-
--
+44 (0)20 7608 7140
Any time things appear to be going better, you have overlooked something.
-Original Message-
From:
set nomail
regards,
-=Dave=-
--
+44 (0)20 7608 7140
***He's dead, Jim. Mmm-mm, there'll be good eatin' tonight!***
www.guardianit.com
The information contained in this email is confidential and intended
only for the use of the individual or entity named above. If the reader
of this mess
Joe,
The default port number for the server listens on for connections from
clients is 1500. This will also include the admin clients. You may have
set the http _admin_ port to something like 1501; however unless you are
admining from OUTSIDE of the firewall using a web browser, you do not need
We too have found the same, using stk and adic libraries when labelling
direct from bulk.
A way round is to checkin the new volumes from bulk, with checklab=no, then
use checkout libv checkl=n rem=n on the volumes, followed by a label
libv search=y labels=bar.
Sounds cumbersome, but the checkou
Tony,
I haven't seen a reply to this yet, so here goes...
The number of slots in a library is read when the library is defined. So to
get tsm to recognise the new slots the library has to be redefined.
q libr f=dand save the output
q dr f=d and save the output
Checkout all the libvols
Cheri
This is our biggest headache with TSM which has NO suitable facility to
provide what you want. Backupsets and archiving are not viable ways of doing
it.
It has been raised as an issue before and there was an excellent description
of how Tivoli could resolve this submitted as a product sugge
Hi *SM-ers
Rather than use TSM to backup systemobjects we are using NTBACKUP to backup
systemstate. This allows us to do a bare metal restore which TSM doesn't
support. However we appear to have no option but to backup systemobjects
with TSM because we feel much safer specifying all-local rather
Hi Rene,
We have also found the "nohup" is not being honoured on solaris/hpux/aix. I
think it is something to do with the posix standards.
The best work around which somebody posted to the list some time ago (MANY
thanks whoever you are!!) is to use "at" to restart the daemon:
echo "/bin/dsmc
hi all,
We have had several of the Netware 5.1 nodes that we back up upgraded to
service pack 3.
Since then, most of the nodes have given considerable trouble, especially
those using a scheduled backup. Even worse, those nodes doing a client
initiated backup - either manual or via portal - have
Yes, _auditing_ the library is safe. Audit will check the current tsm
logical libvols against the physical ones it finds, checking out any that
are no longer visible. If you weren't running acsls it would also refresh
the cartridge<=>slot mapping.
What isn't safe is to run a checkin libv search
Hello *smers,
We are currently having a problem with backing up one particular volume out
of several in a Novell Netware v5 cluster; tsm client version 4.1.1.0, tsm
server version 4.1.3.0
We can backup all the other clustered volumes without any problem except for
the volume containing (and call
EMAIL PROTECTED]
Subject: Re: tsm mirroring vs pprc
Dave, what the level of the TSM server that caused this problem?
On Tue, 12 Jun 2001, Frost, Dave wrote:
> Richard,
>
> "BTDTGTTS"
>
> We are running tsm using san attached emc for all disk storage. On
initial
> set
Richard,
"BTDTGTTS"
We are running tsm using san attached emc for all disk storage. On initial
setup of the emc, we decided to stop using log/db mirroring, and use the emc
mirroring instead.
Soon afterwards, we had 2 instances where we could not restart the server.
Both times the server failed
Does add_drv lb work?
Check that you are running the Solaris 64 bit kernel using "isainfo -b".
TSM server is not supported at 32 bit on Solaris 7 or 8.
Confirm that the mt and lb binaries agree with your kernel ("file mt" will
tell you if it is 64 bit or 32 bit).
Be warned - if you are using sa
Seth,
We have found that the error is often associated with a specific tape
volume, and will occur when you try to WRITE to the tape during migration or
reclamation.
We have resolved it by identifying the tape concerned and marking it
readonly - you may need to do some experiments to find it, th
How to backup open files??
We are trying to backup a database file (not one that has an agent). The
database remains "open" at all times, but is not opened exclusive. We back
it up SHAREDSTATIC. We have noticed that we very rarely back this database
up and would probably not have a recent version
Hi *SMers,
we are running tsm version 3.7.4 on Solaris 7, and Sony ait2 tape drives in
an adic scalar 1000.
We are seeing a very large number reposition faults - especially with OP=RSR
- along with a scattering of actual read or write errors, spread across 10
drives and 500 tapes. In one partic
Hi *SMer's
Following a server crash, we have been forced to restore the database from
the latest backup. The database restored ok, then started running through
the dblog records at the end of the dbdump. During this process, it
suddenly aborted with
Sequential media log undo pass in progre
Hi *SM'ers
Is anybody using TSM to backup a Stratus VOS? Apparently this can be done
by nfs mounting the filesystem, but there is some concern that using this
techique could give problems on restores, since the data will be put back in
a physically different position, and physical position is ve
hi all
has anybody seen this problem:
NT Server 4 running MSSQL7.
Before TDP was installed, user used use an MSSQL7 function called LINK
SERVER.
This allowed them to connect to an AS400 database.
Since TDP was installed the function now gives
user32.dll error 3 initialization failure
For the *really* lazy, what I do is remove the volumes from the library, and
then run audit libr checkl=bar.
dave
--
"on a clear disk, you can seek for ever."
> --
> From: Shekhar Dhotre[SMTP:[EMAIL PROTECTED]]
> Sent: 21 August 2000 14:42
> To: [EMAIL PROTECTED]
> Su
ep, they are all the same..
> It's the files in the directory that you should worry about. Do you know
> what file in the WinNT directory that is getting updated on a daily basis.
> It should be in the log if you have it set for verbose.
>
> Mark
>
>
> -Original Mes
T directory was not restored?
> As far as using a special dirmc, the WinNT directory should still be there
> as long as it is still on the client.
>
> Mark
>
>
> -Original Message-
> From: Frost, Dave [mailto:[EMAIL PROTECTED]]
> Sent: Friday, August 18, 2000 8
Hi *SMers
We were restoring an exchange server to a point-in-time - both the server
itself and exchange. While we were restoring the server, we found that
there were no "winnt" directories for the date we wanted. Investigating
this, we have found that if the nt server is NOT running exchange, t
58 matches
Mail list logo
