Hi All!
System environment:
One TSM server and six TSM client. Oracle financials application and Oracle 8.1.6
database.
The system environment consists of 4 nodes and one extention node type S7A. The S7A is
the
TSM server, the nodes are the TSM clients. AIX file level daily incremental
Following my tests before Christmas with TSM 4.2.1.9 server on AIX I posted a
note regarding mount point mis-behaviour. I have now done some more testing
and I think the problem I saw related to the subtle behaviour when you have
no tape drives defined to a server : processes do not wait. If
I think the wide range of answers as to what this output means is a good
indication that IBM would do well to develop a more concise tool for
reviewing resource utilization. As for my opinion I would start with
'vmtune' and then look at 'iostat'. Good Luck
Thanks,
Jon Martin
-Original
Along thos same lines maybe if there are three file systems with the
following naming conventions
/work
/work/data
/work/data/martine
Would TSM back up each of these filesystems separately thus giving you three
different file names?
Thanks,
Jon Martin
-Original Message-
From: Vo An
HI,
Can you give me the output of these commands:
/usr/samples/kernel/vmtune
iostat 1
vmstat 1
svmon -G
netstat -m
netstat -in
netstat -v
ps aux |more
ps avx | more
ps -ef | sort +3 -r | more
ps -ef | egrep -v STIME|$LOGNAME | sort +3 -r | head -n 10
nfsstat -s
nfsstat -c
no -a
After this
We solved this exact problem by reducing BUFPOOLSIZE in dsmserv.opt. It was
set very high, causing TSM to (apparently) pin that much memory for its own
use - then the OS didn't have enough. Symptoms: high paging, rotten TSM
performance, admin sessions took minutes to connect.
Recommend using
I went ahead yesterday and applied the patch to my production sigh, I don't yet
have a test environment servers. I didn't see any evidence of mount point wait
failures from last night, nor any drives becoming unavailable due to mount point
errors.
So, hopefully, 4.2.1.9 DID fix the problem!
I am currently in the process of replacing all of my 3590 tapes in my ATL
with new 3590 EXTENDED length cartridges. I have a buyer for the 3200 3590
tapes I am removing from TSM/ATL. How do I go about making sure all data
has been erased from the old 3590 cartridges? It is my understanding
There is a better tool. It was introduced in AIX 4.3.2 and it's called
topas (it's very similiar to monitor or top) Simply enter topas at
the AIX command prompt. As to the problem, if the system is really paging,
it's most likely your bufpool is set too high. I would try lowering your
bufpool
Gill,
Any groups you create on 1 server must also be defined
on the other server(s) in the cluster otherwise, the
cluster will not start between the servers this is if
you're using HACMP as your cluster application. After
defining each volume group on each server in the
cluster you have to sync
Stu,
Creating mirrors is good way to move your log volumes and database. Format
your log volumes, the same size as your current logs, then add them as
mirrors. When the log volume mirror is in synch you can delete the current
log volumes.
Jim Sporer
At 02:29 PM 12/26/2001 -0500, you wrote:
Hi
Jim
Can I create one volume for each through this process? I understand that I
need to have equiv. or better size for the new volumes prior to mirroring,
but there are currently 4 log volumes and 3 database volumes. I want to
amalgamate them and have one of each properly sized, then mirror,
Hi all!
I have open registration on the server and at first login dsmc starts asking
the questions about password, login and contact information.
Does anybody knows how to pass answers on PC to that questions from script?
Thank you in advance,
Yuriy Us.
We're using RH7.2 and TSM 4.1.2 client here.
Some things to note :
(1) If you're using a journaled filesystem, like ext3, you need to put
a the following in your dsm.sys for each mount point defined in /etc/fstab
for your local harddisks and/or NFS mounts.
VirtualMountPoint /
Stu,
Sure you can do that. Format the new volume you want to use for the
database, then add the volume to the data base and extend it. Delete the
current volumes one at a time by reducing the data base by the size of the
volume you are deleting and then just delete it. TSM will move the data
There are several ways to do this:
1) Use another backup utility to write over the tape. You can use tar,
cpio, NT backup, or whatever to write over either the whole tape or
merely the TSM label.
2) Just do # 1 above to the TSM DB backups since the remaining tapes are
useless without the DB
Hi,
We currently have a problem, where our ESA 1 disk array will crash when
the disk that our database backs up to is full. Has anyone else suffered
from this problem.
Thanks
Stuart
The information contained in
FYI #3 won't do anything to the data...
I talked with someone that had relabeled (by mistake) some tapes
(with valid data) on new 3590 E1A drives
they then put the labels back on with the 3590 B1A drives and experienced
only limited data loss. I believe just the first file on the tape(s).
I presume you have a server that has the TAPEUTIL utility (NTUTIL on NT).
If so, there is a capability to erase the tape using the same
functionality as the mainframe dataset erase (DSE) otherwise know as Erase
on Scratch with RACF. This can also be issued in a batch mode. The Magstar
I've got this SQL statement for backup, archive and restore stats.
select activity,sum(bytes)/1048576*.001 from summary where activity in
('BACKUP','ARCHIVE','RESTORE') and start_timTIMESTAMP(CURRENT_DATE -1 DAY,
'08:00:00') group by activity order by 1
I'm guessing that was at least
Or just the opposite - a volume with data checked in as scratch, tried to
be used and (auto)changed by TSM to private to prevent scratch volumes
allocation become blocked.
Zlatko Krastev
IT Consultant
Doug Thorneycroft [EMAIL PROTECTED] on 03.01.2002 00:08:56
Please respond to [EMAIL
Be aware that the filesystem is not equal to all directories and files
under the mount point tree!
So continuing with your example and assuming we have filesystems /work,
/work/data and /work/data/martine:
a file /work/dummy_file will give us FILESPACE_NAME='/work', HL_NAME='/',
(1) is an option but time consuming (3200 tapes!)
(2) might not be an option - it depends how paranoic you are. If someone
(usually a virus) fills sector 0 on a PC disk with garbage this does not
automatically mean the data is lost. But it might take too much efforts. If
the tapes are not
Are these by any chance NFS mounted directories? We do a lot of
archives with loopback-mounted filesystems, and over the course
of multiple iterations of the dsmc client we have found that
ADSM/TSM have treated these filesystems differently wrt to what
is FILESPACE_NAME and what is HL_NAME.
The
Chris - tapas is part of which fileset and in which directory does it
reside?
On Thu, 3 Jan 2002, Chris Gibes wrote:
There is a better tool. It was introduced in AIX 4.3.2 and it's called
topas (it's very similiar to monitor or top) Simply enter topas at
the AIX command prompt. As to the
Hi,
Can you give me some hints about offsite backup automation? Can anybody
send me some schedule samples to automate offsite backups?
Best Regards,
Burak
Hi,
I have 2 3583 library with 3580 LTO drives on them. I am
confused about firmwares. Do firmwares coming with appcode.exe
include FMR firmwares for the drives or do they include only library firmware?
I can easily update library firmare with serial cable as download -C1
-FVn_n_n.lif
but I
27 matches
Mail list logo