IS anyone backing up Filenet with TSM?
My company is implementing a new Filenet document archiving system on AIX. The problem this is going to give is that it stores the data in an Oracle database and this is stored on one large raw logical volume. This is going to start at 200GB and will grow to 2 TB. Even at 200 GB its going to be slow to backup as it will have to be a full volume backup and I can't see a way to multistream this to more than one drive. I'm not sure if we can use the Oracle database agent to back this up at document level. Any advice is much appreciated. Mark
Re: ADSM questions
I am in need of the following information and do not know what scripts to create to get the information. Amount of data backed up daily. That can be obtained from the TSM accounting records. # of tapes used for backups on a daily basis. The simplest way is to do daily queries of your backup storage pool volumes complement, and track the increase. If I were to create a copypool how many tapes would be needed on a daily basis for the copypool data. Given that it's a copy, it will essentially be the same number of volumes as in your primary storage pool, assuming same device type and format. Richard Sims, BU
Re: sun solaris and adsm
What is the latest version out there for the clients? See http://www.tivoli.com/support/storage_mgr/adsercli.htm for TSM; see http://www.tivoli.com/support/storage_mgt/adsm/adsercli.htm for ADSM. See ftp://ftp.storsys.ibm.com/tivoli-storage-management/patches/ for TSM fixes. Richard Sims, BU
counting tapes
# of tapes used for backups on a daily basis. The simplest way is to do daily queries of your backup storage pool volumes complement, and track the increase. Or you could look in the actlog for msgs 1340,1341,1360,6684... Or at the tail of your volhistory disk file: unix tail dsmserv.volhistory 2000/10/17 15:48:42 STGDELETE 0 0 0 MAGSTAR3590 M00569 2000/10/17 16:30:33 STGDELETE 0 0 0 MAGSTAR3590 M00724 2000/10/17 17:04:37 STGDELETE 0 0 0 MAGSTAR3590 M00833 2000/10/17 17:39:14 STGNEW 0 0 0 MAGSTAR3590 M02030 2000/10/17 17:45:36 STGDELETE 0 0 0 MAGSTAR3590 M01467 2000/10/17 18:16:50 STGDELETE 0 0 0 MAGSTAR3590 M01175 2000/10/17 18:47:06 STGDELETE 0 0 0 MAGSTAR3590 M00468 2000/10/17 20:23:20 STGDELETE 0 0 0 MAGSTAR3590 A3 2000/10/17 20:53:12 STGDELETE 0 0 0 MAGSTAR3590 M00060 2000/10/18 00:11:36 STGNEW 0 0 0 MAGSTAR3590 M02031 ADSMq actlog begind=-1 msg=1340 Date/TimeMessage -- 10/17/00 09:02:15 ANR1340I Scratch volume M02025 is now defined in storage pool TSMDBB. 10/17/00 10:32:35 ANR1340I Scratch volume M02026 is now defined in storage pool BACKUP3590_OFFSITE. 10/17/00 11:31:02 ANR1340I Scratch volume M02027 is now defined in storage pool BACKUP3590_OFFSITE. 10/17/00 13:42:07 ANR1340I Scratch volume M02028 is now defined in storage pool BACKUP3590_OFFSITE. 10/17/00 15:20:02 ANR1340I Scratch volume M02029 is now defined in storage pool TAPEPOOL. 10/17/00 17:39:17 ANR1340I Scratch volume M02030 is now defined in storage pool TAPEPOOL. 10/18/00 00:11:40 ANR1340I Scratch volume M02031 is now defined in storage pool TAPEPOOL. tsm: ADSMq actlog begind=-1 msg=1360 Date/TimeMessage -- 10/18/00 05:01:16 ANR1360I Output volume M02032 opened (sequence number 1). tsm: ADSMq actlog begind=-1 msg=1341 Date/TimeMessage -- 10/17/00 15:03:06 ANR1341I Scratch volume M00031 has been deleted from storage pool BACKUP3590_OFFSITE. 10/17/00 15:03:09 ANR1341I Scratch volume M00042 has been deleted from storage pool BACKUP3590_OFFSITE. 10/17/00 15:48:45 ANR1341I Scratch volume M00569 has been deleted from storage pool TAPEPOOL. sm: ADSMq actlog begind=-1 msg=6684 Date/TimeMessage -- 10/17/00 15:03:07 ANR6684I MOVE DRMEDIA: Volume M00031 was deleted. 10/17/00 15:03:09 ANR6684I MOVE DRMEDIA: Volume M00042 was deleted. 10/17/00 15:03:11 ANR6684I MOVE DRMEDIA: Volume M00092 was deleted. -- Richard
3583 support
Hi, in the list of devices supported by TSM I am missing the "3583 Ultrium Scalable Tape Library". Does anybody know, if and when this support will come? -- Reinhard MerschWestfaelische Wilhelms-Universitaet Zentrum fuer Informationsverarbeitung - ehemals Universitaetsrechenzentrum Roentgenstrasse 9-13, D-48149 Muenster, Germany Tel: +49(251)83-31583 E-Mail: [EMAIL PROTECTED] Fax: +49(251)83-31653
Re: Error msg
Hello, I'm running ADSM Server 3.1.2.50 on an OS/390 and I have several NT server with ADSM Client 3.1.0.8. One of the servers posts an error when running an scheduled backup: ANS1410E Unable to access network path. I have been talking to my NT people and they say there is no error on the network. The msg appears allways backing up the last disk. This problem does not occur when I do backups from the Client GUI. Angel - That's one of those poorly documented product messages... As of version 3.1.0.5 of the client, ADSM now uses the dreaded UNC names for the files. This means that the machine name is part of the file name, which as a composite is "network path(name)". This is a name issue, and has nothing to do with communications networking, as the message description indicates. It may be that the filespace name on the ADSM server needs to be renamed/adjusted so that the UNC name matches that of the machine and disk. You can probably discern this through queries. Reference APAR number IC24166. Richard Sims, BU
TSM 3.7 Macintosh scheduler
Recent postings have noted that the TSM 3.7 Macintosh client scheduler does not work. Some customers using are using TSM 3.7 clients because of advanced Mac OS levels, but can't really use the TSM 4.1 client fixtest package because they still have ADSM 3.1 servers, and the mix is "not supported". I worked with fellow customer Bernd Knaak, of Germany, on his instance of this problem. It occurred to me that a good compromise might be to download and adopt just the 4.1 fixtest scheduler into a 3.7 Mac client. (ftp://ftp.storsys.ibm.com/tivoli-storage-management/patches/client/v4r1/Mac/v411/readme41113.txt) Bernd tried that and reports that it works: "I downloaded TSM scheduler daemon 4.1.1.13, replace the origin one from TSM 3.7.2.0, replace the alias in the startup folder of the MAC OS 9 and restart the MAC. ProcessWatcher shows the daemon running and after a while the TSM scheduler starts, get its schedule and stop. After a view minutes - I set the schedule-starttime to the time showing by my clock - the TSM backup began and runs." Others with this problem may want to try this approach and see if it works as well for them. If any problems, please let us all know. Richard Sims (Mac bigot in hiding), BU
Re: counting tapes
Or you could redirect (append) a scheduled select after the days backup to compare growth over time: select stgpool_name,count(*) from volumes group by stgpool_name [EMAIL PROTECTED] 10/18/00 08:40AM # of tapes used for backups on a daily basis. The simplest way is to do daily queries of your backup storage pool volumes complement, and track the increase. Or you could look in the actlog for msgs 1340,1341,1360,6684... Or at the tail of your volhistory disk file: unix tail dsmserv.volhistory 2000/10/17 15:48:42 STGDELETE 0 0 0 MAGSTAR3590 M00569 2000/10/17 16:30:33 STGDELETE 0 0 0 MAGSTAR3590 M00724 2000/10/17 17:04:37 STGDELETE 0 0 0 MAGSTAR3590 M00833 2000/10/17 17:39:14 STGNEW 0 0 0 MAGSTAR3590 M02030 2000/10/17 17:45:36 STGDELETE 0 0 0 MAGSTAR3590 M01467 2000/10/17 18:16:50 STGDELETE 0 0 0 MAGSTAR3590 M01175 2000/10/17 18:47:06 STGDELETE 0 0 0 MAGSTAR3590 M00468 2000/10/17 20:23:20 STGDELETE 0 0 0 MAGSTAR3590 A3 2000/10/17 20:53:12 STGDELETE 0 0 0 MAGSTAR3590 M00060 2000/10/18 00:11:36 STGNEW 0 0 0 MAGSTAR3590 M02031 ADSMq actlog begind=-1 msg=1340 Date/TimeMessage -- 10/17/00 09:02:15 ANR1340I Scratch volume M02025 is now defined in storage pool TSMDBB. 10/17/00 10:32:35 ANR1340I Scratch volume M02026 is now defined in storage pool BACKUP3590_OFFSITE. 10/17/00 11:31:02 ANR1340I Scratch volume M02027 is now defined in storage pool BACKUP3590_OFFSITE. 10/17/00 13:42:07 ANR1340I Scratch volume M02028 is now defined in storage pool BACKUP3590_OFFSITE. 10/17/00 15:20:02 ANR1340I Scratch volume M02029 is now defined in storage pool TAPEPOOL. 10/17/00 17:39:17 ANR1340I Scratch volume M02030 is now defined in storage pool TAPEPOOL. 10/18/00 00:11:40 ANR1340I Scratch volume M02031 is now defined in storage pool TAPEPOOL. tsm: ADSMq actlog begind=-1 msg=1360 Date/TimeMessage -- 10/18/00 05:01:16 ANR1360I Output volume M02032 opened (sequence number 1). tsm: ADSMq actlog begind=-1 msg=1341 Date/TimeMessage -- 10/17/00 15:03:06 ANR1341I Scratch volume M00031 has been deleted from storage pool BACKUP3590_OFFSITE. 10/17/00 15:03:09 ANR1341I Scratch volume M00042 has been deleted from storage pool BACKUP3590_OFFSITE. 10/17/00 15:48:45 ANR1341I Scratch volume M00569 has been deleted from storage pool TAPEPOOL. sm: ADSMq actlog begind=-1 msg=6684 Date/TimeMessage -- 10/17/00 15:03:07 ANR6684I MOVE DRMEDIA: Volume M00031 was deleted. 10/17/00 15:03:09 ANR6684I MOVE DRMEDIA: Volume M00042 was deleted. 10/17/00 15:03:11 ANR6684I MOVE DRMEDIA: Volume M00092 was deleted. -- Richard
Several tapes with zero capacity and zero utilization
I'm new with the ADSM system, so I hope I don't sound to dumb here! I've been looking at the list for a couple of months and I've gotten a lot of good information from it so far. Here we go, I run a select command every few days to check on the number of reclaimable tapes we have in the system. When I get several tapes with a PCT_RECLAIM above 70-80% I run the reclamation process. We do the tape reclamation manually. Here is the select command we use: select volume_name,est_capacity_mb,pct_utilized,pct_reclaim from volumes where stgpool_name='COPYPOOL' order by pct_reclaim desc The following is part of the result I get: VOLUME_NAME EST_CAPACITY_MB PCT_UTILIZED PCT_RECLAIM SG0113 19317.595.9 4.1 SG0117 14658.097.1 2.8 SG0007 0.0 0.0 0.0---?? SG0009 0.0 0.0 0.0---?? SG0014 0.0 0.0 0.0?? SG0027 0.0 0.0 0.0?? SG0030 13671.799.9 0.0 SG0046 0.0 0.0 0.0-?? SG0047 13533.1 100.0 0.0 SG0075 0.0 0.0 0.0?? SG0108 0.0 0.0 0.0?? SG0124 0.0 0.0 0.0-?? SG0133 0.0 0.0 0.0?? SG0135 11764.2 100.0 0.0 SG0139 10122.4 100.0 0.0 How can I have a tapes with 0 capacity and 0 utilization? We use 3590 tapes in a 3494 library attached to an IBM Netfinity server running NT 4 using ADSM version 3.7.3. Any ideas? David Tyree Microcomputer Specialist South Georgia Medical Center 912.333.1155
Re: Web client access problems after 3.7.3 SM upgrade
Yes. This worked like a charm. btw-I should have read the manual first. Its just that we never used this 2 server addresses before in adsm 3.1.2.40 and I did not even know they existed! Ken Sedlacek Information Technology Infrastructure - Kyrus Corp. W: 864-322-4260 Cell: 864-444-8375 Text Page: 864-444-7243, follow prompts [EMAIL PROTECTED] |+--- || "Charlotte | || Brooks/Almade| || n/IBM" | || [EMAIL PROTECTED]| || .COM| || | || 10/17/00 | || 06:04 PM | || Please | || respond to | || "ADSM: Dist | || Stor Manager"| || | |+--- ---| | | | To: [EMAIL PROTECTED]| | cc: (bcc: Ken Sedlacek/GV/Kyrus)| |Groups: Don't Expand | | Subject: Re: Web client access | | problems after 3.7.3 SM upgrade | ---| Ken - try QUERY SERVER. If it shows no servers defined you need to define one using the DEFINE SERVER command. You can use the minimum option, eg DEFINE SERVER tsm1 PASSWORD=tsm1 HLaddress=n.n.n.n LLAddress=1500 So the servername and password can be anything, the HLaddress is your server's TCP/IP address and LLADdress is the server port. After this, the console should come up. Hope this helps .. Charlotte Brooks Email: [EMAIL PROTECTED] Project Leader, Storage Management, ITSO Almaden http://ibm.com/redbooks Ken Sedlacek [EMAIL PROTECTED]@VM.MARIST.EDU on 10/17/2000 01:54:22 PM Please respond to "ADSM: Dist Stor Manager" [EMAIL PROTECTED] Sent by: "ADSM: Dist Stor Manager" [EMAIL PROTECTED] To: [EMAIL PROTECTED] cc: Subject: Web client access problems after 3.7.3 SM upgrade Last night we upgraded ADSM 3.1.2.40 to TSM 3.7.3 without much difficulty. I am having problems with the IE5.0 web client access though. When trying to access the TSM server via IE5, I get an error message on the Java console saying: "Server parameter is missing. Please Set serverhladdress and serverlladdress." I have been trying to find these 2 parameters: serverhladdress and serverlladdress Does anyone know where they are located and what should go in them? TIA Ken Sedlacek Information Technology Infrastructure - Kyrus Corp. W: 864-322-4260 Cell: 864-444-8375 Text Page: 864-444-7243, follow prompts [EMAIL PROTECTED]
Re: Several tapes with zero capacity and zero utilization
When I get several tapes with a PCT_RECLAIM above 70-80% I run the reclamation process. David - Another approach to scratch pool management is to wait until the number of scratches reaches a comfortable minimum, and only then start reclamation. I prefer to let attrition empty tapes as much as possible, rather than subject tapes to wear in copying remaining contents sooner than necessary - considering also that that remainder may expire soon anyway, making that much copying wasteful. How can I have a tapes with 0 capacity and 0 utilization? What your display doesn't tell you is the state of the tapes. Particularly for a copy storage pool, REUsedelay should be in effect, and the volumes would thus be Pending. Richard Sims, BU
Re: Several tapes with zero capacity and zero utilization
I did the q vol status=pending and it has the same tapes. They are pending. I had forgotten that the tapes stay in the pending status for three days. Thanks! -Original Message- From: Doug Thorneycroft [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 18, 2000 10:41 AM To: [EMAIL PROTECTED] Subject: Re: Several tapes with zero capacity and zero utilization They are probably pending, do a q vol status=pending and see if the list matches your zero capacity tapes. On Wednesday, October 18, 2000 7:10 AM, Tyree, David [SMTP:[EMAIL PROTECTED]] wrote: I'm new with the ADSM system, so I hope I don't sound to dumb here! I've been looking at the list for a couple of months and I've gotten a lot of good information from it so far. Here we go, I run a select command every few days to check on the number of reclaimable tapes we have in the system. When I get several tapes with a PCT_RECLAIM above 70-80% I run the reclamation process. We do the tape reclamation manually. Here is the select command we use: select volume_name,est_capacity_mb,pct_utilized,pct_reclaim from volumes where stgpool_name='COPYPOOL' order by pct_reclaim desc The following is part of the result I get: VOLUME_NAME EST_CAPACITY_MB PCT_UTILIZED PCT_RECLAIM SG0113 19317.595.9 4.1 SG0117 14658.097.1 2.8 SG0007 0.0 0.0 0.0---?? SG0009 0.0 0.0 0.0---?? SG0014 0.0 0.0 0.0?? SG0027 0.0 0.0 0.0?? SG0030 13671.799.9 0.0 SG0046 0.0 0.0 0.0-?? SG0047 13533.1 100.0 0.0 SG0075 0.0 0.0 0.0?? SG0108 0.0 0.0 0.0?? SG0124 0.0 0.0 0.0-?? SG0133 0.0 0.0 0.0?? SG0135 11764.2 100.0 0.0 SG0139 10122.4 100.0 0.0 How can I have a tapes with 0 capacity and 0 utilization? We use 3590 tapes in a 3494 library attached to an IBM Netfinity server running NT 4 using ADSM version 3.7.3. Any ideas? David Tyree Microcomputer Specialist South Georgia Medical Center 912.333.1155
Re: Several tapes with zero capacity and zero utilization
That sounds like an excellent idea. At the moment we are running short on tapes until we get a new batch in. Do you know of an _easy_ way to tell just how many scratch tapes we have? We are using select statements using the info in the volumes table. The problem is the tapes only show in stgpool_name as either copypool or tapepool. I run a select statement to get the number of copypool tapes then another to get the number of tapepool tapes. Then I have to add then together and subtract that from the number of tapes we put in the library to begin with. It's a convoluted process but that's the best I can come up right now. I can't find a direct way to tell me the number of scratch tapes I have. -Original Message- From: Richard Sims [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 18, 2000 10:44 AM To: [EMAIL PROTECTED] Subject: Re: Several tapes with zero capacity and zero utilization When I get several tapes with a PCT_RECLAIM above 70-80% I run the reclamation process. David - Another approach to scratch pool management is to wait until the number of scratches reaches a comfortable minimum, and only then start reclamation. I prefer to let attrition empty tapes as much as possible, rather than subject tapes to wear in copying remaining contents sooner than necessary - considering also that that remainder may expire soon anyway, making that much copying wasteful. How can I have a tapes with 0 capacity and 0 utilization? What your display doesn't tell you is the state of the tapes. Particularly for a copy storage pool, REUsedelay should be in effect, and the volumes would thus be Pending. Richard Sims, BU
Database Design Decision - TSM 3.7.3.8 - AIX 4.3.3
Hi all, Recently, I migrated our TSM server from an RS/6000 H50 to an H80. During this process, my TSM database changed from 5 - 9.1 GB SSA drives to 3 - 18 GB SSA drives. (Acutally, 10 - 9.1's and 6 - 18's utilizing TSM mirroring). In the old config. I had used raw volumes, as recommended by a consultant when the system first arrived. IBM TSM Support felt there was no discernible performance difference between raw and jfs filesystems for the TSM DB volumes, so I opted for jfs on the new config. I can't judge the difference as my performance has increased by about 40%, but I went from a 4-way H50 to a 4-way H80, and from 80 MB/s SSA to 160 MB/s SSA. Has anyone tried raw vs. AIX jfs for TSM DB performance? The other question I posed to IBM TSM Support was how to configure the 3 -18GB SSA disks: Three separate filesystems, each with a DB volume file, or one filesystem striped across all three disks, providing TSM one DB volume. They were unable to tell me whether TSM would manage the three DB vols in parallel or serial, but felt that the three vs. one striped was the way to go. Now that I am done, I see, using iostat, that during a DB backup it is only accessing one drive of the three at a time, showing me that TSM uses these volumes sequentially, not in parallel. Would I not get better TSM DB performance with one logical volume/filesystem striped across all three drives? If not, why? If so, large or small stripe size? Thanks, Joe Goshko ADC Software Systems The information contained in this e-mail is confidential and intended solely for the use of the addressee(s) listed above. If you have obtained this message in error or otherwise, please notify the sender and destroy all copies of this message immediately.
Platforms for TSM
Before Implementing TSM on solaris think about "mksysb " and ufsdump , incase you need to undergo Disaster recovery Logical volume manager and format command or solstice disk suite (In Solaris ) smit/smitty (Best A systems management tool , I am using couple of flaovours of Unix , HP is having SAM ,and sun Admintool , sco - scoadmin IRIX , Linux -linuxconf , but see the difference between smit and all others. Dynamic hardware configuration (cfgmgr in AIX ) : and boot -r (Solaris). while installing software AIX automatically extends /usr if needed , Sun hangs (error ) and try comparing other features of OS .. and of course Advantages of JFS I am for AIX ... Shekhar...
Re: Several tapes with zero capacity and zero utilization
For 3494's, you can use the mtilb command to count scratch tapes, or insert tapes - whatever class you want. I've got a shared library - 4 TSM servers, one 3494, and I run this script daily: echo "Scratch tapes on RTP1" /home/nickpc/data/scratchcount mtlib -l /dev/lmcp0 -qC -s012E | wc -l /home/nickpc/data/scratchcount echo "Scratch tapes on RTP2" /home/nickpc/data/scratchcount mtlib -l /dev/lmcp0 -qC -s013E | wc -l /home/nickpc/data/scratchcount echo "Scratch tapes on RTP3" /home/nickpc/data/scratchcount mtlib -l /dev/lmcp0 -qC -s014E | wc -l /home/nickpc/data/scratchcount echo "Scratch tapes on RTP4" /home/nickpc/data/scratchcount mtlib -l /dev/lmcp0 -qC -s015E | wc -l /home/nickpc/data/scratchcount echo "Number of insert tapes in UR13494" /home/nickpc/data/scratchcount mtlib -l /dev/lmcp0 -qC -sFF00 | wc -l /home/nickpc/data/scratchcount mtlib -l /dev/lmcp0 -qL | grep Intervention /home/nickpc/data/scratchcount more -e /home/nickpc/data/scratchcount I have a second script that runs this one and then mails the output file to me. Nick Cassimatis [EMAIL PROTECTED] "I'm one cookie away from happy." - Snoopy (Charles Schulz)
TSM Migration
In an environment with two ADSM 3.1 servers migrating to a single TSM 4.1 server, what is the general procedure for doing such a task? I want to verify Im not overlooking some considerations and want to see how others would go about it. Appreciate any input! Gerald Wichmann RS Engineer Sansia System Solutions 408-844-8893 work 408-844-9801 fax
Re: Several tapes with zero capacity and zero utilization
I have a script that counts scratches- my 3494 is shared betweeen two tsm
instances, so I use
a table of category codes to sort out private/scratch for each tsm instance.
This, of course, is
from the library/category point of view, which may not necessarily balance 100%
to TSM's!
Here's the (ksh) script:
lookup_code()
{
table=/usr/tivoli/tsm/etc/category_table
desc=`grep ^$1 $table|awk -F= '{print $2}'`
}
echo
echo
echo "--- TSM carts ---"
echo
mtlib -l /dev/lmcp0 -qI|awk '{print $2}'|sort |uniq -c|while read count cat;do
lookup_code $cat
printf "%5d " $count
echo $desc
done
Here's the category table (tsmSF1 and tsmSD1 are my tsm instance names) :
0190=tsmSF1 private
0191=tsmSF1 scratch 3490
0192=tsmSF1 scratch 3590
01F4=tsmSD1 private
01F5=tsmSD1 scratch 3490
01F6=tsmSD1 scratch 3590
F00E=volume in error
FF10=eject pending
FF11=bulk eject
FF00=inserted. (unassigned, in ATL)
FFF7=pending
FFFA=manually ejected
FFFB=purge volume
FFFC=unexpected volume
FFFA=purge volume
FFFD=3590 cleaner
FFFE=3490 cleaner
=private volume
Cleaner carts don't seem to show in the normal mtlib -qI command, so you can
add the following to count these:
echo
echo "--- Cleaner Carts ---"
echo
mtlib -l /dev/lmcp0 -qC -s fffd|awk '{print $2}'|sort |uniq -c|while read count
cat;do
lookup_code $cat
printf "%5d " $count
echo $desc
done
But this still doesn't show us how many cleaning cycles are available on these
cleaning carts, so add the following to display that (plus the library empty
cell count):
echo
echo
echo "--- Library Status ---"
echo
mtlib -l /dev/lmcp0 -qL|egrep "available cells|3590 cleaner cycles"
echo
Sample output:
--- Cleaner Carts ---
4 3590 cleaner
--- TSM carts ---
106 tsmSF1 private
82 tsmSF1 scratch 3590
46 tsmSD1 private
263 tsmSD1 scratch 3590
50 inserted. (unassigned, in ATL)
--- Library Status ---
available cells311
avail 3590 cleaner cycles..374
Comments, corrections, and other code swaps welcomed!
JRS.
Re: Several tapes with zero capacity and zero utilization
Hi, I have the following alias in my .profile to provide scratch tape numbers. It's pretty simple (but it avoids the select commands that sometimes use lots of resources on my system). alias scratch="dsmadmc -id=qid -password=qpass q libv |grep Scratch |wc -l" Regards, Ben Date: Wed, 18 Oct 2000 16:51:18 +0100 From: "Warren, Matthew James" [EMAIL PROTECTED] Subject: Re: Several tapes with zero capacity and zero utilization To: [EMAIL PROTECTED] MIME-version: 1.0 Content-transfer-encoding: 7BIT Hello David, I use, select count(*) from libvolumes where status='Scratch' Matthew. Do you know of an _easy_ way to tell just how many scratch tapes we have? We are using select statements using the info in the volumes table. The problem is the tapes only show in stgpool_name as either copypool or tapepool. I run a select statement to get the number of copypool tapes then another to get the number of tapepool tapes. Then I have to add then together and subtract that from the number of tapes we put in the library to begin with. It's a convoluted process but that's the best I can come up right now. I can't find a direct way to tell me the number of scratch tapes I have. -Original Message- From: Richard Sims [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 18, 2000 10:44 AM To: [EMAIL PROTECTED] Subject: Re: Several tapes with zero capacity and zero utilization When I get several tapes with a PCT_RECLAIM above 70-80% I run the reclamation process. David - Another approach to scratch pool management is to wait until the number of scratches reaches a comfortable minimum, and only then start reclamation. I prefer to let attrition empty tapes as much as possible, rather than subject tapes to wear in copying remaining contents sooner than necessary - considering also that that remainder may expire soon anyway, making that much copying wasteful. How can I have a tapes with 0 capacity and 0 utilization? What your display doesn't tell you is the state of the tapes. Particularly for a copy storage pool, REUsedelay should be in effect, and the volumes would thus be Pending. Richard Sims, BU
Re: Database Design Decision - TSM 3.7.3.8 - AIX 4.3.3
Has anyone tried raw vs. AIX jfs for TSM DB performance? Joseph - This has historically been discussed on the List. The Tivoli manuals advise against the use of raw logical volumes, but you can certainly do so if you want. Overall, performance is pretty much a "wash". The other question I posed to IBM TSM Support was how to configure the 3 -18GB SSA disks: Three separate filesystems, each with a DB volume file, or one filesystem striped across all three disks, providing TSM one DB volume. They were unable to tell me whether TSM would manage the three DB vols in parallel or serial, but felt that the three vs. one striped was the way to go. Now that I am done, I see, using iostat, that during a DB backup it is only accessing one drive of the three at a time, showing me that TSM uses these volumes sequentially, not in parallel. Would I not get better TSM DB performance with one logical volume/filesystem striped across all three drives? If not, why? If so, large or small stripe size? Tivoli management should do something about TSM Support people who don't know squat about TSM. It's well known that TSM performance benefits from more volumes in devoting a thread to each volume, and that threads are parallel. This will result in better performance when there are multiple client sessions occurring. But ADSM/TSM utility functions have historically been single-threaded operations. That will hopefully improve over time. But, generally, you want more volumes, not fewer. Richard Sims, BU
Scripts/Tools to monitor drive usage?
For the sake of capacity planning, I'd like to monitor my drive usage, and find out at what times we are maxed, and for how long. I'm thinking a combo of a cron job that does a 'q mount' every 10 minutes, to gather the information, and then using gnuplot or excel to plot it. But before I go about this, I'd like to know if anyone out there has created similar solutions they are willing to share. Note: We don't have the ADSM ODBC driver configured yet at our site.
Re: Scripts/Tools to monitor drive usage?
Here is a sample html I create by querying the actlog.
reclaimed volume won't go pending
I have this 3590 volume, offsite, which I have reclaimed by setting the reclaim percentage on the storage pool. The volume is now empty. I know this because "q content xxx" returns "ANR2034E QUERY CONTENT: No match found using this criteria.". The problem is that the volume won't go pending, I.E. "q v xx" returns pct util = 0.0 status=Filling. Any ideas on how to get this volume back into the fold? Environment: ADSM 3.1.2.55 Ken Chamberlain University of Toronto Phone:(416) 978-1582 Fax: (416) 971-2085 E-mail: [EMAIL PROTECTED]
Re: IS anyone backing up Filenet with TSM?
If you haven't implemented this Filenet system yet I would look into it real closely. We have filenet here at two of our sites and we recently had a meeting with our DR provider and the filenet folks and they say they don't recommend using anything other then their own utilities to backup "filenet". It supposedly uses Oracle under the covers but has some additional proprietary files that need to be backed up in sysnc with the Oracle files and their utility is the only one that does it properly. They have a process that puts the output of their utility to a file that you can pickup with TSM but nothing that can go straight to TSM such as a TDP. Please keep me posted on your progress. JIm - Do You Yahoo!? Yahoo! Messenger - Talk while you surf! It's FREE.
web client doesn't like IE?
Recently switched from using Netscape Navigator to Internet Explorer. The ADSM web client seems to not like it. I point the URL to the same place: server:1581, it looks like it tries to load (get ADSM Web Backup/Archive Client as the header of the browser window), but the menu doesn't appear on the page. The dsmwebcl.log lists a connection entry that looks ok: 10/18/2000 13:37:27 (dsmcad) ANS3006I Processing request for the ADSM Web Client Any ideas what it doesn't like about IE? Tracy __ Do You Yahoo!? Yahoo! Messenger - Talk while you surf! It's FREE. http://im.yahoo.com/
Re: reclaimed volume won't go pending
Have you tried audit vol xx f=y ? On Wednesday, October 18, 2000 11:43 AM, Ken Chamberlain [SMTP:[EMAIL PROTECTED]] wrote: I have this 3590 volume, offsite, which I have reclaimed by setting the reclaim percentage on the storage pool. The volume is now empty. I know this because "q content xxx" returns "ANR2034E QUERY CONTENT: No match found using this criteria.". The problem is that the volume won't go pending, I.E. "q v xx" returns pct util = 0.0 status=Filling. Any ideas on how to get this volume back into the fold? Environment: ADSM 3.1.2.55 Ken Chamberlain University of Toronto Phone:(416) 978-1582 Fax: (416) 971-2085 E-mail: [EMAIL PROTECTED]
Re: Scripts/Tools to monitor drive usage?
Rich, Sample html was not included. Can you send it to me Andy --- -Original Message- From: richard cowen [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 18, 2000 2:39 PM To: [EMAIL PROTECTED] Subject: Re: Scripts/Tools to monitor drive usage? Here is a sample html I create by querying the actlog.
Generate Backupsets for D/R?
Does anyone use Generate Backupsets for D/R? Do you store the backupset on some type of independent media so that you don't need your ADSM server for D/R? I was wondering if this ability was worth recommending a move of ADSM off the mainframe to Unix type box, so that we can write backupsets to some kind of tape that NT boxes or Novell boxes can use. Can this be done? Thanks. Julie Phinney
Re: web client doesn't like IE?
Our standard browser is IE and has no problem loading the interface. My problem has been with the way the interface works in general (at least on ADSM clients 3.1.6 and 3.1.7). Are you using IE5 or higher? Damon Burkhart Operations Analyst Server Operations Kmart Corporation (248)614-0629 -Original Message- From: Tracy T. [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 18, 2000 2:52 PM To: [EMAIL PROTECTED] Subject: web client doesn't like IE? Recently switched from using Netscape Navigator to Internet Explorer. The ADSM web client seems to not like it. I point the URL to the same place: server:1581, it looks like it tries to load (get ADSM Web Backup/Archive Client as the header of the browser window), but the menu doesn't appear on the page. The dsmwebcl.log lists a connection entry that looks ok: 10/18/2000 13:37:27 (dsmcad) ANS3006I Processing request for the ADSM Web Client Any ideas what it doesn't like about IE? Tracy __ Do You Yahoo!? Yahoo! Messenger - Talk while you surf! It's FREE. http://im.yahoo.com/
Re: Several tapes with zero capacity and zero utilization
I have the following script defined in TSM. All I have to do, is: run q_n_scratch (or just click on the script I want to run, from the Web GUI) /* -*/ /* Script Name: Q_N_SCRATCH*/ /* Description: Display number of scratch */ /* for each libraries defined */ /* Parameter: none*/ /* Example: run q_n_scratch*/ /* -*/ select LIBRARY_NAME, count(*) as - "Number of Scratch" from libvolumes where - status='Scratch' group by LIBRARY_NAME Serge Gaudet, CMI Consultant "Tyree, David" To: [EMAIL PROTECTED] david.tyree@cc: SGMC.ORGSubject: Several tapes with zero capacity and zero Sent by: utilization "ADSM: Dist Stor Manager" [EMAIL PROTECTED] RIST.EDU 18/10/2000 10:09 AM Please respond to "ADSM: Dist Stor Manager" I'm new with the ADSM system, so I hope I don't sound to dumb here! I've been looking at the list for a couple of months and I've gotten a lot of good information from it so far. Here we go, I run a select command every few days to check on the number of reclaimable tapes we have in the system. When I get several tapes with a PCT_RECLAIM above 70-80% I run the reclamation process. We do the tape reclamation manually. Here is the select command we use: select volume_name,est_capacity_mb,pct_utilized,pct_reclaim from volumes where stgpool_name='COPYPOOL' order by pct_reclaim desc The following is part of the result I get: VOLUME_NAME EST_CAPACITY_MB PCT_UTILIZED PCT_RECLAIM SG0113 19317.595.9 4.1 SG0117 14658.097.1 2.8 SG0007 0.0 0.0 0.0---?? SG0009 0.0 0.0 0.0---?? SG0014 0.0 0.0 0.0?? SG0027 0.0 0.0 0.0?? SG0030 13671.799.9 0.0 SG0046 0.0 0.0 0.0-?? SG0047 13533.1 100.0 0.0 SG0075 0.0 0.0 0.0?? SG0108 0.0 0.0 0.0?? SG0124 0.0 0.0 0.0-?? SG0133 0.0 0.0 0.0?? SG0135 11764.2 100.0 0.0 SG0139 10122.4 100.0 0.0 How can I have a tapes with 0 capacity and 0 utilization? We use 3590 tapes in a 3494 library attached to an IBM Netfinity server running NT 4 using ADSM version 3.7.3. Any ideas? David Tyree Microcomputer Specialist South Georgia Medical Center 912.333.1155
Re: web client doesn't like IE?
You didn't say what version of IE you were using, for what version of the Web client. However, I have no problems with IE 5.00.2919.6307 for TSM 3.7.3 Server, or the TSM 3.7.2 Clients of my (current) environment. Serge Gaudet, CMI Consultant "Tracy T." adsmigmo@YAHTo: [EMAIL PROTECTED] OO.COM cc: Sent by: Subject: web client doesn't like IE? "ADSM: Dist Stor Manager" [EMAIL PROTECTED] RIST.EDU 18/10/2000 02:51 PM Please respond to "ADSM: Dist Stor Manager" Recently switched from using Netscape Navigator to Internet Explorer. The ADSM web client seems to not like it. I point the URL to the same place: server:1581, it looks like it tries to load (get ADSM Web Backup/Archive Client as the header of the browser window), but the menu doesn't appear on the page. The dsmwebcl.log lists a connection entry that looks ok: 10/18/2000 13:37:27 (dsmcad) ANS3006I Processing request for the ADSM Web Client Any ideas what it doesn't like about IE? Tracy __ Do You Yahoo!? Yahoo! Messenger - Talk while you surf! It's FREE. http://im.yahoo.com/
Re: web client doesn't like IE?
IE 5.5 ADSM client 3.1.7 ADSM MVS server 3.1.2 --- "Burkhart, Damon" [EMAIL PROTECTED] wrote: Our standard browser is IE and has no problem loading the interface. My problem has been with the way the interface works in general (at least on ADSM clients 3.1.6 and 3.1.7). Are you using IE5 or higher? Damon Burkhart Operations Analyst Server Operations Kmart Corporation (248)614-0629 -Original Message- From: Tracy T. [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 18, 2000 2:52 PM To: [EMAIL PROTECTED] Subject: web client doesn't like IE? Recently switched from using Netscape Navigator to Internet Explorer. The ADSM web client seems to not like it. I point the URL to the same place: server:1581, it looks like it tries to load (get ADSM Web Backup/Archive Client as the header of the browser window), but the menu doesn't appear on the page. The dsmwebcl.log lists a connection entry that looks ok: 10/18/2000 13:37:27 (dsmcad) ANS3006I Processing request for the ADSM Web Client Any ideas what it doesn't like about IE? Tracy __ Do You Yahoo!? Yahoo! Messenger - Talk while you surf! It's FREE. http://im.yahoo.com/ __ Do You Yahoo!? Yahoo! Messenger - Talk while you surf! It's FREE. http://im.yahoo.com/
Re: Generate Backupsets for D/R?
I am trying to generate this from morning , . once i get it done i will pass it to you. shekhar "[EMAIL PROTECTED]/P=Internet/A= /C=us" on 10/18/2000 03:22:55 PM Please respond to "[EMAIL PROTECTED]/P=Internet/A= /C=us" @ X400 To: "[EMAIL PROTECTED]/P=Internet/A= /C=us"@X400 cc: Subject: Generate Backupsets for D/R? Does anyone use Generate Backupsets for D/R? Do you store the backupset on some type of independent media so that you don't need your ADSM server for D/R? I was wondering if this ability was worth recommending a move of ADSM off the mainframe to Unix type box, so that we can write backupsets to some kind of tape that NT boxes or Novell boxes can use. Can this be done? Thanks. Julie Phinney
Re: TSM 3.7 Macintosh scheduler
Richard, Is the unsupported mix noted below limited to the Macintosh/ADSM enviroment? Richard Sims wrote: Recent postings have noted that the TSM 3.7 Macintosh client scheduler does not work. Some customers using are using TSM 3.7 clients because of advanced Mac OS levels, but can't really use the TSM 4.1 client fixtest package because they still have ADSM 3.1 servers, and the mix is "not supported". I worked with fellow customer Bernd Knaak, of Germany, on his instance of this problem. It occurred to me that a good compromise might be to download and adopt just the 4.1 fixtest scheduler into a 3.7 Mac client. (ftp://ftp.storsys.ibm.com/tivoli-storage-management/patches/client/v4r1/Mac/v411/readme41113.txt) Bernd tried that and reports that it works: "I downloaded TSM scheduler daemon 4.1.1.13, replace the origin one from TSM 3.7.2.0, replace the alias in the startup folder of the MAC OS 9 and restart the MAC. ProcessWatcher shows the daemon running and after a while the TSM scheduler starts, get its schedule and stop. After a view minutes - I set the schedule-starttime to the time showing by my clock - the TSM backup began and runs." Others with this problem may want to try this approach and see if it works as well for them. If any problems, please let us all know. Richard Sims (Mac bigot in hiding), BU -- Jim Kirkman AIS - Systems UNC-Chapel Hill 966-5884
I/O error on 3570 drives attached to a TSM Solaris Server
I'm getting some strange errors when trying to use the 3570 drives of a 3575 tape library attached to a Sun Solaris TSM Server. Everything seems to be ok when configuring the drives to the operating system. We have even run some tests using the 'IBM Tapeutil' utility and the drives work just fine. The define drive command run with no errors. But when I try to label tapes, the first one is labeled without problems but when TSM tries to use the drive to label another tape or perform a backup, I/O errrors appear on the console and the operation is not performed. Detailed information on logs and definition follows. Any suggestion to this? Thanks in advance. Regards Gerardo Zapata Gonzalez IT Specialist ISM IBM de Mixico 52704000 Ext. 04224 tsm: TSM_SATURNOq drive Library Name Drive NameDevice Type DeviceON LINE --- --- --- ROBOT 3570DRV1 3570 /dev/rmt/2stc Yes tsm: TSM_SATURNOq drive f=d Library Name: ROBOT Drive Name: 3570DRV1 Device Type: 3570 Device: /dev/rmt/2stc ON LINE: Yes Element: 17 Allocated to: Last Update by (administrator): ADMIN Last Update Date/Time: 10/13/00 12:50:36 Cleaning Frequency (Gigabytes/ASNEEDED/NONE): NONE tsm: TSM_SATURNOq libr Library Name: ROBOT Library Type: SCSI Device: /dev/rmt/1smc Private Category: Scratch Category: External Manager: Shared: No saturno# pwd /dev/rmt saturno# ls 0st 0stcbn 2stb2stcn 3stc4st 4stcbn 5stbn 5stn7stc 0stb0stcn 2stbn 2stn3stcb 4stb4stcn 5stc6smc7stcb 0stbn 0stn2stc3st 3stcbn 4stbn 4stn5stcb 7st 7stcbn 0stc1smc2stcb 3stb3stcn 4stc5st 5stcbn 7stb7stcn 0stcb 2st 2stcbn 3stbn 3stn4stcb 5stb5stcn 7stbn 7stn saturno# When labeling volumes: tsm: TSM_SATURNO q act 10/13/00 11:58:08 ANR2017I Administrator ADMIN issued command: LABEL LIBVOLUME ROBOT SEARCH=YES LABELSOURCE=BARCODE CHECKIN=SCRATCH OVERWRITE=YES 10/13/00 11:58:08 ANR0984I Process 2 for LABEL LIBVOLUME started in the BACKGROUND at 11:58:08. 10/13/00 11:58:08 ANR8799I LABEL LIBVOLUME: Operation for library ROBOT started as process 2. 10/13/00 11:58:08 ANR0984I Process 2 for LABEL LIBVOLUME started in the BACKGROUND at 11:58:08. 10/13/00 11:58:08 ANR8799I LABEL LIBVOLUME: Operation for library ROBOT started as process 2. 10/13/00 11:58:08 ANR0609I LABEL LIBVOLUME started as process 2. 10/13/00 11:59:04 ANR8810I Volume 00C5B3 has been labeled in library ROBOT. 10/13/00 11:59:18 ANR8427I CHECKIN LIBVOLUME for volume 00C5B3 in library ROBOT completed successfully. 10/13/00 11:59:49 ANR8302E I/O error on drive 3570DRV (/dev/rmt/0stc) (OP=READ, CC=0, KEY=00, ASC=00, ASCQ=00, SENSE=00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00- more... (ENTER to continue, 'C' to cancel) .00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.0- 0.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.- 00.00.00.00.00.00.00.00.00.00, Description=An undetermined error has occurred). Refer to Appendix B in the 'Messages' manual for recommended action. 10/13/00 11:59:53 ANR8810I Volume 00DBFA has been labeled in library ROBOT. 10/13/00 12:00:06 ANR8427I CHECKIN LIBVOLUME for volume 00DBFA in library ROBOT completed successfully. 10/13/00 12:00:27 ANR8302E I/O error on drive 3570DRV (/dev/rmt/0stc) (OP=READ, CC=0, KEY=00, ASC=00, ASCQ=00, SENSE=00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00- .00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.0- 0.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.- 00.00.00.00.00.00.00.00.00.00, Description=An undetermined error has occurred). Refer to Appendix B in the 'Messages' manual for recommended action. 10/13/00 12:00:31 ANR8810I Volume 00DC15 has been labeled in library ROBOT. 10/13/00 12:00:45 ANR8427I CHECKIN LIBVOLUME for volume 00DC15 in library ROBOT completed successfully. 10/13/00 12:01:06 ANR8302E I/O error on drive 3570DRV (/dev/rmt/0stc) (OP=READ, CC=0, KEY=00, ASC=00, ASCQ=00,
reclamation issue and an ANR9999D
Here's one. I've got a server that dies during reclamation on a copy pool tape with the following error: ANRD aferase.c(528): Invalid logSizeRatio 4048033.806786 (logSize=0.372, size=0.1061, aggrSize=0.4860) for aggregate 0.9306677 Based on what I've read around here I did a bit of research on the old reclamation issues and the Audit Reclaim, etc. tools, since it looked like that was the problem here. However, Audit Reclaim says that it 'isn't required' and doesn't run. Simply ditching the tape would be perfectly acceptable (since it's just a copy pool tape), but any attempt to delete the volume with a discarddata=yes just returns the same error. It's an AIX server currently at 3.1.2.15. Yes, it most certainly is problem code and needs to be updated, but I'm a little bit concerned about attempting to update the code before fixing the issue. Should I be? Anyone have any ideas? regards, Paul
