RMAN on ORA-01403
Hi guys and girls, haven't been here for a while and it's nice to be back. I need help from one of the Oracle-Gurus on this list. One of my customers moved some of his Oracle datafiles to another filesystem. Unfortunately, he/she did so without telling RMAN. So even though the Redbook says RMAN needs to know about any structural change of the target database ..., what's been done has been done. So now, I tried to convince RMAN to accept the new database layout. The first thing that happend was a: RMAN-06004: ORACLE error from recovery catalog database: RMAN-20003: target database incarnation not found in recovery catalog So I did a RESET DATABASE to set things back, which worked until resync'ing the catalog. There it blows with: RMAN-03006: non-retryable error occurred during execution of command: resync RMAN-07004: unhandled exception during command execution on channel default RMAN-10035: exception raised in RPC: ORA-01403: no data found RMAN-10031: ORA-1403 occurred during call to DBMS_RCVCAT.CHECKBACKUPPIECE Does anyone here know what is going on and how to resolve it? I am by no means a database expert so I did not really understand what some of the Oracle documentation stated. CYA Lars
dsmcad running scheduler
Hi folks, I am sure this question has been asked before. One of our customers is running dsmcad (Client ver 5.1.1.0) in webclient schedule mode. Last night the schedules got missed. After I checked the log files this morning, I realized dsmcad set with a HTTPClientport 1582 claims to have shifted to port 1583: 10/09/03 16:58:27 Trying port number 1583 10/09/03 16:58:27 Obtained new port number on which to listen. However, a netstat -a reveales that it is still listening on 1582. Is there a reason why dsmcad would lie to the TSM server? CYA Lars
Re: AIX Question
Hi there, I guess you already got the right answer from Rene. Just add another entry to dsm.sys with tcpserveraddress pointing to the ip address of the GB card. The scheduler will be started with the default entry and handle connections through the 10/100MBit network while you will need to start your backup client either with the option -se=gbit_address or with using an alternative dsm.opt file employing the -optfile=dsm_gbit_address.opt option. Also, do not forget to define or modify your schedules appropriately by setting the option field for the relevant schedules to one of the above. CYA Lars On 08.09.2003 12:17:26 ADSM: Dist Stor Manager wrote: Hi TSMers, Sorry for doing this...but does any one know how to configure a gigabit card on AIX ...my scenario is as follows:- I want to use the gigabit card purely for tsm backups and for my nfs mounts..it is on a gigabit switch. I am using the en0 at present 10/100mb networkthe application will connect via this card but I need my backups and nfs mounts to connect via the gigabit card. Can you specify a second host name on the server for the gigabit card address ? Any ideas Lindy Crawford Information Technology Nedbank Corporate - Property Asset Finance *+27-31-3642185 +27-31-3642946 [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] This email and any accompanying attachments may contain confidential and proprietary information. This information is private and protected by law and accordingly if you are not the intended recipient you are requested to delete this entire communication immediately and are notified that any disclosure copying or distribution of or taking any action based on this information is prohibited. Emails cannot be guaranteed to be secure or free of errors or viruses. The sender does not accept any liability or responsibility for any interception corruption destruction loss late arrival or incompleteness of or tampering or interference with any of the information contained in this email or for its incorrect delivery or non-delivery for whatsoever reason or for its effect on any electronic device of the recipient. If verification of this email or any attachment is required please request a hard-copy version
Two questions regarding NDMP backups
Hi guys and girls, I have just been busy implementing NDMP for a NetApp filer under TSM 5.2.0.2 last week. The Module runs just fine including file level restore capabilities. However, two questions popped up in my mind that do not get answered in the relevant redbooks so maybe you on the list can help me with these: 1. For the backup copy group running NDMP backups you can set all the normal version and retension parameters, but NDMP backups still use the conventional approach of doing fulls and differentials. What will TSM actually consider as a version (verexist) since a differential will prove to be quite useless if there is no related full? 2. Is there any way to actually delete data on the backup image level from TSM storage? I understand you can always delete the full filespace of a client node. Next time you do a NDMP backup the filespace gets recreated with a new ID (!) but I was not able to figure out if there is a trick to delete lets say just one particular differential backup image. Thanks for helping with this issue Lars
Re: AIX Question
Hi, this is actually a good and valid question. Since I am a strong believer of KISS I actually do not see any obvious reason for separating the traffic even though TSM makes it possible to do so. The only reason I would figure is that the traffic should flow via two separate subnets with different requirements but thats network infrastructure so Lindy will have to answer this question. CYA Lars On 08.09.2003 17:10:55 Miles Purdy wrote: Hi there, I have a related question: what is the benefit to using two interfaces? Why not just use the gigabit card, for all TSM traffic? Is it better to keep the high-speed backups on the giga bit interface and keep the scheduler traffic on its own? It seems like a lot of extra work and confusion. Consequently, I have a very similar setup, and I'm always interested in ways to make the backups run faster. Thanks Miles [EMAIL PROTECTED] 08-Sep-03 6:15:05 AM Hi there, I guess you already got the right answer from Rene. Just add another entry to dsm.sys with tcpserveraddress pointing to the ip address of the GB card. The scheduler will be started with the default entry and handle connections through the 10/100MBit network while you will need to start your backup client either with the option -se=gbit_address or with using an alternative dsm.opt file employing the -optfile=dsm_gbit_address.opt option. Also, do not forget to define or modify your schedules appropriately by setting the option field for the relevant schedules to one of the above. CYA Lars On 08.09.2003 12:17:26 ADSM: Dist Stor Manager wrote: Hi TSMers, Sorry for doing this...but does any one know how to configure a gigabit card on AIX ...my scenario is as follows:- I want to use the gigabit card purely for tsm backups and for my nfs mounts..it is on a gigabit switch. I am using the en0 at present 10/100mb networkthe application will connect via this card but I need my backups and nfs mounts to connect via the gigabit card. Can you specify a second host name on the server for the gigabit card address ? Any ideas Lindy Crawford Information Technology Nedbank Corporate - Property Asset Finance *+27-31-3642185 +27-31-3642946 [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] This email and any accompanying attachments may contain confidential and proprietary information. This information is private and protected by law and accordingly if you are not the intended recipient you are requested to delete this entire communication immediately and are notified that any disclosure copying or distribution of or taking any action based on this information is prohibited. Emails cannot be guaranteed to be secure or free of errors or viruses. The sender does not accept any liability or responsibility for any interception corruption destruction loss late arrival or incompleteness of or tampering or interference with any of the information contained in this email or for its incorrect delivery or non-delivery for whatsoever reason or for its effect on any electronic device of the recipient. If verification of this email or any attachment is required please request a hard-copy version
TDP for Oracle on Linux
Hi guys and girls, does anyone of you know if there are plans to support the TDP for Oracle on Linux in the near future? If so, when can we expect public availability? Thanks a lot Lars
Re: Upgrading HSM and AIX at the same time?
Hi doing the upgrade the usual way employing an AIX migration installation would preserve all filesystems in the rootvg except /tmp. Only the the filesets delivered with AIX will be upgraded but not the ADSM/TSM/HSM client filesets. After you have done the AIX upgrade just upgrade the TSM client. To be on the save side you could also go the alternate rootvg installation way, so if something goes wrong you can easily switch back provided you have a spare disk to use as the new rootvg disk. CYA Lars On 20.02.2003 13:59:14 ADSM: Dist Stor Manager wrote: Hello, we have an HSM client version 3.1.20 running on AIX 4.2.1. For some reason we need to upgrade AIX to 4.3.3 which means to do an install from scratch. It would make sense to upgrade HSM to the current level. As far as I understand from the manual upgrading HSM alone is not a big deal. The upgrade procedure would do everything automatically. However, after installing AIX no HSM would be there to upgrade. What we would do is to export the HSM filesystem, install AIX, TSM client and HSM. Then we can import the HSM filesystem. How can we tell HSM to run on this filesystem without thinking it is new and destroying it? Advice would be appreciated. Regards Gerhard --- Gerhard Rentschleremail:[EMAIL PROTECTED] Regional Computing Center tel. ++49/711/685 5806 University of Stuttgart fax: ++49/711/682357 Allmandring 30a D 70550 Stuttgart Germany
Schedule twice weekly
Hi guys and girls, anyone got a good idea of how to start a schedule twice a week say Tuesday and Thursday? It could be nice if TSM would allow more than one entry in the dayofweek field. Thanks Lars
DB2 backup problem
Hi guys and girls, we are running weekly offline backups of a DB2 database ver 7.1.0.65 EEE on AIX. The server is a NSM at 4.1.6.0 and the clients at 3.7.2.0. The last backup failed on one of the nodes showing lots of these entries in the dsmerror.log: sessInit: Starting communications initialization 01/19/03 17:59:29 sessInit: Transitioning: sInit state === sTRANSERR state Can someone explain to me what this means? I searched this list and found this question being asked a couple of times but there was never an answer to it. The only hint I was able to find was in DB2's diag logs: 2003-01-19-18.02.51.509653 Instance:db2minst Node:001 PID:51956(db2agent (DB2MPROD) 1) Appid:*N1.db2minst.030119170156 database_utilities sqlubcka Probe:20 Database:DB2MPROD fbf5 556e 6162 6c65 2074 6f20 6f62 ÿÿûõUnable to ob 7461 696e 2073 7570 6572 2065 7863 6c75 tain super exclu 7369 7665 2063 6f6e 6e65 6374 696f 6e00 sive connection. Anyone there to translate?? Thanks Lars
Re: 3590 under Linux
Hi there, thanks everyone for the replies. So I figured out the 3590-B11 now. If it runs in Random mode tape as well as changer commands get channeled through the scsi id of the tape drive. So basically there will be only be lun 0 for the drive. In 2Lun-Random mode the changer gets lun 1 and should be seen by the scsi hba. Whether to choose Random or 2Lun Random mode actually depends on the application (or so the manual says). We had the beast running in 2Lun Random mode on AIX/TSM so I thought I this should be the preferable choice. So here's what I did for everyone of you who ever wants to run a B11 on Linux: First, I had to switch on multiple lun scanning in the hba's (AHA-3944) BIOS. It is actually not switched on by default. Linux will still see no lun 1 even though the hba does. Anyone who ever compiled a Linux kernel will remember there's an option for the scsi module that switches on multiple lun scanning (CONFIG_SCSI_MULTI_LUN). And h IBM, what arest thou writing in the file IBMtape_Magstar.ReadMe: The 32 bit 1.2.2 version of IBMtape Linux Device Driver operates under the following i386 compatible 32 bit environments: 1. Red Hat 7.2 with the 2.4.9-31 smp and uni-processor kernels 2. Red Hat 7.2 with the 2.4.9-31 enterprise kernel . . The kernel rpm packages are available on the following ftp site: ftp://ftp.redhat.com/pub/redhat/support/enterprise/isv/kernel-archive/7.2/2.4.9-31 Beware folks (and everyone from IBM who reads this list), this kernel is compiled WITHOUT multiple lun support enabled. So I had to download RedHat's 2.4.9 kernel source and recompile the kernel with CONFIG_SCSI_MULTI_LUN=y in the config. Watch out that you set your kernels name -31, -31smp or -31enterprise via EXTRAVERSIONS in the main Makefile otherwise the IBMtape modules will probably refuse to load. Afterwards just run IBMtapeconfig and a device /dev/IBMchanger0 will be created that you will use as the device entry for defining a path from the tsm server to the library. Have a nice weekend, Cya Lars On 13.12.2002 13:26:41 ADSM: Dist Stor Manager wrote: I am currently busy testing a TSM server (5.1.5.4) under Linux (RH7.2). I have a 3590-B11 attached to the machine. The only device that shows up in /proc/scsi/scsi is: Host: scsi0 Channel: 00 Id: 05 Lun: 00 Vendor: IBM Model: 03590B11 Rev: A429 Type: Sequential-AccessANSI SCSI revision: 03 When I run IBMtapeconfig it creates the special device files /dev/IBMtape0 and /dev/IBMtape0n. Since I want to configure the 3590 as a library would I not also need a device for the media changer? The docu says there will be a device called /dev/IBMchangerX for libraries. However, no such device shows up when the adapter scans the bus. I was setting the 3590 to random mode as well as random 2-lun mode still, there is no changer device visible. What can be done, did someone of you ever test this beast with Linux?? Thanks Lars Hi, forgive me if I'm wrong, but isn't the 3590 just a tape-drive? If you've placed it in some sort of library, that must either be a 3494 or a special made STK library, in either case, the library is not controlled via the drive, but via some other channel (ethernet, serial, scsi). 3590 is the identifier for a type of tape technology, which encompasses a variety of model having a variety of features. The B11 model is a medium changer, like on the old 3480s - sort of a tiny library hanging off the front of the drive. Lars, you didn't say what doc you are following. The IBM TotalStorage Tape Device Drivers: Installation and User's Guide manual would seem the definitive one. I see that it says on p.130: IBMtapeconfig removes all the existing /dev/IBMtape and /dev/IBMchanger special files, and creates new ones based on the information in /proc/scsi/IBMtape and /proc/scsi/IBMchanger. so I would first look to see that both of those /proc files are in place, and that they seem to contain the right information. Recheck things based upon that manual to see if all is in place. If IBMtapeconfig is a shell script, you can invoke it with verbosity to record what it does. Richard Sims, BU
Re: TSM in a MS Cluster environment
Hi there, you are perfectly correct. There's only one thing left. dsmcad is actually the one responsible for accepting connections on the http ports set in dsm.sys. For accessing the mentioned webclients just create two more option files pointing to the relevant server entry. Under AIX (that's what I tested so far) you can then start dsmcad with -optfile=dsm_whatever.opt and then login on 1581, 1582 and 1583. Have a nice day Cya Lars On 12.12.2002 21:26:35 ADSM: Dist Stor Manager wrote: Hey gang, I'm thinking about using the TSM web client for a customer of mine. I understand how to set up the 1581 port and what not. The only thing I'm confused about is the following. On this one server they have two oracle databases which we back up with two different server instances and one o/s instance. To backup the o/s and the two databases the following TSM schedulers must be running: dsmc sched = for the o/s backup dsmc sched -server=db_instancename_1 = for the first database instance. dsmc sched -server=db_instancename_2 = for the second database instance. My question is when the client logs onto this web interface by going to the web address: http://client_hostname:1581;. I'm assuming they will get the node definitions of the default server stanza in dsm.sys which will give them access to the o/s data only. Now what if they want the web interface to have access to the other two instances/nodes as well? Which will give them access to there database data. I'm guessing the option: httpport in the dsm.sys file. This way have 1581 for the regular o/s, 1582 for the second database instance and 1583 for the second database instance. This way: http://client_hostname:1581 = will give them access to the o/s data. http://client_hostname:1582 = will give them access to there first database instance information. http://client_hostname:1583 = will give them access to there second database instance information. Am I right to assume this? Am I missing anything with the configuration of this sort of environment. Any assistance or procedures would be appreciated. Thanks!. --Justin Richard Bleistein Unix Systems Administrator (Sungard eSourcing) Desk: (856) 866 - 4017 Cell:(856) 912-0861 Email: [EMAIL PROTECTED]
Re: Upgrading firmware on 3583 library and 3580 drives
Hi, Not at all. The devices just have to be reset, ie you do a rmdev -l followed by a mkdev -l for the appropriate devices. After, you should confirm with lscfg -v -l that the new firmware level is correct. Bye Lars On 12.12.2002 11:21:38 ADSM: Dist Stor Manager wrote: Hello everybody, When you upgrade the firmware in your drive and/or library, is it necessary to (under AIX) 'rmdev -dl' the library and drives and (in TSM) 'delete drive' and 'delete library' after updating the firmware? The drives went from 22T4 to 25D4 and the library went from 2.72.0012 to 2.80.0041. Thanks, -- Jurjen Oskam PGP Key available at http://www.stupendous.org/
3590 under Linux
Hi guys and girls, I am currently busy testing a TSM server (5.1.5.4) under Linux (RH7.2). I have a 3590-B11 attached to the machine. The only device that shows up in /proc/scsi/scsi is: Host: scsi0 Channel: 00 Id: 05 Lun: 00 Vendor: IBM Model: 03590B11 Rev: A429 Type: Sequential-AccessANSI SCSI revision: 03 When I run IBMtapeconfig it creates the special device files /dev/IBMtape0 and /dev/IBMtape0n. Since I want to configure the 3590 as a library would I not also need a device for the media changer? The docu says there will be a device called /dev/IBMchangerX for libraries. However, no such device shows up when the adapter scans the bus. I was setting the 3590 to random mode as well as random 2-lun mode still, there is no changer device visible. What can be done, did someone of you ever test this beast with Linux?? Thanks Lars
DB2 backup problem
Hi guys and girls, hope there is some kind of DB2 guru reading in this list. When doing a DB2 backup we are getting the following messages in the appropriate DB2 diag log: 2002-12-08-18.00.35.226529 Instance:db2minst Node:001 PID:90534(db2agent (DB2MPROD) 1) Appid:C0A80196.046F.021207182454 base_sys_utilities sqleflsr Probe:3 Database:DB2MPROD DiagData 0150 ...P 2002-12-08-18.00.35.260127 Instance:db2minst Node:001 PID:90534(db2agent (DB2MPROD) 1) Appid:C0A80196.046F.021207182454 base_sys_utilities sqleflsr Probe:3 Database:DB2MPROD RIDS/sqleflsr MS/ABNOR_TERRM PRCS/3 0150 ...P 2002-12-08-18.00.35.270169 Instance:db2minst Node:001 PID:90534(db2agent (DB2MPROD) 1) Appid:C0A80196.046F.021207182454 base_sys_utilities sqleflsr Probe:4 Database:DB2MPROD DiagData 5351 4c43 4120 2020 0088 fb38 SQLCA ÿÿû8 2020 2020 2020 2020 2020 2020 2020 .. 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 5351 4c45 4941 4754 SQLEIAGT 0001 2020 2020 2020 2020 2020 2020 2020 2020 2002-12-08-18.00.35.293557 Instance:db2minst Node:001 PID:90534(db2agent (DB2MPROD) 1) Appid:C0A80196.046F.021207182454 base_sys_utilities sqleflsr Probe:4 Database:DB2MPROD RIDS/sqleflsr MS/ABNOR_TERRM PRCS/45351 4c43 4120 2020 0088 fb38 SQLCA ÿÿû8 2020 2020 2020 2020 2020 2020 2020 .. 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 5351 4c45 4941 4754 SQLEIAGT 0001 2020 2020 2020 2020 2020 2020 2020 2020 2002-12-08-18.00.51.894542 Instance:db2minst Node:001 PID:26964(db2agent (DB2MPROD) 1) Appid: base_sys_utilities sqleMergeSqlca Probe:20 Database:DB2MPROD Received sqlcode 1496 for request 801e from node number 2 Data Title:SQLCA PID:26964 Node:001 sqlcaid : SQLCA sqlcabc: 136 sqlcode: 1496 sqlerrml: 0 sqlerrmc: sqlerrp : SQLESRSU sqlerrd : (1) 0x (2) 0x (3) 0x (4) 0x (5) 0x (6) 0x0002 sqlwarn : (1) (2) (3) (4)(5) (6) (7) (8) (9) (10) (11) sqlstate: 2002-12-08-18.00.55.880949 Instance:db2minst Node:001 PID:102342(db2spmrm 1) Appid:none oper_system_services sqloopenp Probe:36 errno: 0011 2002-12-08-18.01.15.702362 Instance:db2minst Node:001 PID:95324(db2bp) Appid:*N1.db2minst.021208170115 database_utilities sqlubckp Probe:26 DiagData fbf5 ÿÿûõ 2002-12-08-18.01.18.341954 Instance:db2minst Node:002 PID:55850(db2bp) Appid:*N2.db2minst.021208170118 database_utilities sqlubckp Probe:26 DiagData fbf5 ÿÿûõ The database is running on AIX 4.3.3 and is backed up up offline after stopping it with a db2stop force and starting it with db2start. Your help is very much appreciated Cya Lars
Re: Linux Client on Debian?
Hi there, either way, using alien to convert the packages or using rpm --nodeps worked for me so far for the client and the server. I was installing them on woody (stable) as well as sarge (testing), using distro as well as customized kernels ver 2.2.x as well as 2.4.x. Just make sure you are using alien -c when converting the server rpm so things get setup correctly. The server will not start otherwise. Haven't tested the server much up to now but the client runs like a charm. The thing thats still missing with the client is an image functionality to save a whole partition. Anyone on this list got an idea if it will be available soon?? Cya Lars On 06.12.2002 00:44:31 ADSM: Dist Stor Manager wrote: Hi, Has anyone succesfully run a Linux client on a Debian distro? I was going to try using alien to turn the .rpm into a .deb files and give it a try. ANybody got any other suggestions?
Re: aix client installation
On 06.12.2002 11:20:57 ADSM: Dist Stor Manager wrote: dear all! I just intalled 2 aix clients as described in the readme file (aix 4.3, tsm client 4.2.2.1) all worked fine, but I am not sure, how to start the scheduler correct. I inserted the following lines to /etc/inittab tsm::once:/usr/bin/dsmc sched /dev/null 21 tsmws::once:/usr/tivoli/tsm/client/ba/bin/dsmcad /dev/null 21 on an other older aix machine (aix 4.3, tsm 4.1.2) the inittab is different: tsm:2:once:startsrc -s dsmc /dev/console 21 #Start TSM Client Scheduler tsmws:2:once:/usr/tivoli/tsm/client/ba/bin/rc.adsmws /dev/console 21 #Start Webshell why can't I use startsrc on the 1st machine? Hi there, as far as I know there's no src functionality for the tsm client available per default. So unless someone seriously got srcmstr convinced, it would not start the dsmc in scheduling mode. You will have to start the client side scheduler with dsmc sched like in the first example. However it should be possible to add a subsystem to the system resource controller with something like: $ mkssys -s dsmc -p /usr/tivoli/tsm/client/ba/bin/dsmc -u0 -R -S -n15 -f3 -a sched Hereafter it should be possible to start and stop the scheduler via the startsrc and stopsrc commands. I would actually not name the subsystem dsmc but something like dsmsched to avoid confusion with invoking the dsmc command line client. Cya Lars
Continual media feed for dlt
Hi guys and girls, here is a real teaser for you. I want to export a client nodes spacemanaged filespace containing about 250GB of data to a dlt library. The dlt got 7 scratch volumes. Last time I did the export it failed with having copied 224 somthing gigabytes. Since the last of the volumes was filled completely all the volumes where returned to scratch status. The dlt library got only one drive and has no barcode reader. What I would like to do is checking out exported volumes lets say after 6 volumes are full and feed the library with new scratches so the process does not stop with the seventh tape. Is there a way to temporary suspend a process to change tapes and then resume the process (similar to a shells CTRL-Z)? Thanx Lars
Continual media feed for dlt
H, good point. Will TSM not complain when it comes across the first volume? I am sure it will, so what will TSM do after: just skip the unavailable volume and continue or wait for a timeout period to make the volume available and then die? Thanks a lot Lars On 06.12.2002 14:41:48 bob.tetstall wrote: You could make some of the volumes to be exported UNAVAILABLE to TSM will not try and export them. Run your partial export then update the rest of the volumes. I do it for my year end backups Best Regards Bob Tetstall Unix System Specialist ab Conferencing, TCICS Team, Aylesbury. *mailto: [EMAIL PROTECTED] *Mobile: +44 7092 306483 *Tel: + 44 1296 618301 *Fax: + 44 1296 339083 *: +44 870 2412996 pin 7940534# (Meetme) Visit our website at www.conferencing.bt.com British Telecommunications plc Registered office: 81 Newgate Street London EC1A 7AJ Registered in England no. 180
Re: 4.2.3 Upgradedb - ??
Hi there, this kind of question popped up some days ago. You should still be able to find the thread. One of the things I realized was the following: If you do the upgrade while the dsmserv process was still running next time you start it it will fail and tell you to invoke it with the upgradedb option. If you do the upgrade without dsmserv running, the upgrade of the database will start automagically at dsmserv startup. But since I have not done every possible upgrade this may also vary from release to release. Cya Lars On 06.12.2002 15:00:18 ADSM: Dist Stor Manager wrote: At 13:54 -0800 12/5/02, Sam Sheppard wrote: I have just installed the 4.2.3 server upgrade and have been running the UPGRADEDB process for almost 3 hours now. When we upgraded from 4.2.2 to 4.2.3, we didn't do an Upgradedb. We didn't see anything in the install doc that said it was necessary. How do we know which version upgrades require an upgradedb? And what happens if we don't do it? TSM still seems to be running OK .. or as close to OK as it ever gets. --
Re: Help
Hi, you will find the drivers on ftp://service.boulder.ibm.com/storage/devdrvr/Win2000. Beware, what IBM says in the README: IBMMag.Win2k.exe - Win/2000 TotalStorage (Magstar) device driver (not to be used with Tivoli Storage Manager (ADSM) ) However, I have seen these drivers running fine with TSM. Just make sure you: 1. Use the right SCSI controller (AHA 2944 UW) is said to be the only one supported. 2. Since the 3570 drives are also supposed to be natively supported under W2K, make sure the IBMmag driver and the native driver are not getting in each others way. I will send a copy of the driver file directly to your email address in case your proxy does not allow for ftp since we do not want to blast this list. CYA Lars On 02.12.2002 15:01:04 ADSM: Dist Stor Manager wrote: Hello TSM people, Please help, if there is someone who knows where I can get the 3570 Magstar library drivers for Windows 2000. I have checked the IBM site and I can not find any there. Cheers. Thabani Bengani Business Solutions IT Boe Corporate Ext. 364 2113 Cell. 0825716778 [EMAIL PROTECTED] WARNING: Any unauthorised use or interception of this email is illegal. If this email is not intended for you, you may not copy, distribute nor disclose the contents to anyone. Save for bona fide company matters, BoE Ltd does not accept any responsibility for the opinions expressed in this email. For further details please see: http://www.boe.co.za/emaildisclaimer.htm
Re: Scheduled event 'HSM.AUTOMIG' failed
Hi do you run a dsmreconcile before starting the migration?? It should fix any inconsitencies in the HSM filesystem and build you a clean candidate list. Bye Lars On 29.11.2002 15:01:19 ADSM: Dist Stor Manager wrote: all, I have a schedule who run the command 'dsmautomig' every day at 20:00. This schedule fails EVERY DAY, however, I found thousands of migrated files. In the dsmerror.log I find the messages : ANS9094W dsmautomig:no candidates found in file system /oper ANS1512E Scheduled event 'HSM.AUTOMIG' failed.Return code = 100 In the schedule.log, I find the message : ANS1512E Scheduled event 'HSM.AUTOMIG' failed. Return code = 100 Why my schedule (dsmautomig) fails every day and however my files are migrated. Any help will be appreciated Thank you _ Gagne une PS2 ! Envoie un SMS avec le code PS au 61166 (0,35 Hors coût du SMS)
Moving from ADSM 3.1 to TSM 5.1
Hi guys and girls, anyone of you ever migrated from ADSM 3.1 straight to TSM 5.1?? Any hints, tips or possible caveats? Will the upgradedb just work out of the box? Thnx Lars
NSM Upgrade from 3.7 to 4.1
Hi guys and girls, on Friday I upgraded a customers NSM migrating from TSM version 3.7.4 to 4.1.6. The upgrade ran fine so far. The customer is running offline DB2 API backups on Sunday evening at 18:00. DB2 runs across two SP nodes. The first node usually completes backing up at about 21:30 using 4 two 5 tapes on a Magstar 3575. This time after the upgrade the node needs considerably longer (until 00:00) using 7 tapes. The same happend for the second node, running for 2.5 vs 3.5 hours and using 3 vs 5 tapes. However, no tablespaces were enlarged or newly created. Did anyone come across the same problem before? Anyone got an idea what I should look for? Thnx and Bye Lars
Re: OS/2 support on 5.X?
Hi, I have testet exactly this configuration: the latest OS/2-Client which is at 3.7.2.27 running on Warp v4 with fix pack 12 and downloadable from boulder backing up to a TSM server on AIX 5.1.0.0_03. I was running the server with ver 5.1.1.6 as well as 5.1.5.2 and opposite to what was said on the list the client works just fine. I guess the only disadvantage will be that clients ver 3.7.x are out of support. If you can live with that your customer should be fine backing up OS/2. Bye Lars On 18.11.2002 17:08:55 ADSM: Dist Stor Manager wrote: All, A while back I saw a post from someone indicating that the OS/2 clients (3.7 vintage) did not work against a 5.1 server. I have a customer still running a goodly number of OS/2 machines, who would very much like to go to 5.1 (on AIX), but if the 3.7 client/5.X server combination won't work, they're going to have to wait 'til they get off of these OS/2 clients. So...what I'm looking for is some kind of confirmation that this is actually the case. Anyone else tried it who can confirm or deny this? I don't have access to any OS/2 machines that I can use to test. Note that they (and I) are well aware that the 3.7 stuff is out of support, and all the ramifications of that. Any responses appreciated. -Lloyd -- - Lloyd Dieter- Senior Technology Consultant Synergy, Inc. http://www.synergyinc.cc [EMAIL PROTECTED] Main:585-389-1260fax:585-389-1267 -
HSM Problem
Hi Guys and Girls, I experience some weird HSM Problem under AIX using client version 4.1.2.20: I let dsmautomig run manually once at night via crontab. I have stderr redirected to a log file and last night I got a whole lot of this error: 11/19/02 15:33:25 OpenDB:PrepareAccess-2 Cannot open database /sambaarch/.SpaceMan/premigrdb: The process file table is full. Can anyone explain to me what this means and how to rectify it?? THNX Lars
Re: q libv just stopped working?
On 11/09/2002 02:19:23 AM ADSM: Dist Stor Manager wrote: Hi guys, where the heck does one change the barcode label length for the 3583? On the libraries' display? Thanks Lars Got it working. ... When updating the microcode to 2.80.41 on the 3583 library, it sets the barcode length to 'normal' (6 CHARACTER) from 'extended' (8 character). After setting it to 'extended' I should have pulled all my tapes out of the library, inventoried the library, then check all the tapes back in. It took a while but it is now working! Why the cleaning tape? ... Don't know. ... JC -Original Message- Sent: Friday, November 08, 2002 4:45 PM Subject: q libv just stopped working? we just upgraded our IBM3583 library. Things were going OK but all of a sudden q libv just stopped working. Only one volume ( a cleaning tape, is seen in the library ) ... bummer. Suggestions? TSM 4.1.3 on Windows NT4 SP6
Study Guide for TSM
Hi guys and girls, I will be doing my TSM certification exam soon. Any suggestions what to use for preparing myself? Like, for the AIX certification there are certification study guides. Is there anything like that for TSM?? Thanks Lars
Backupsets and DB2
Hi guys and girls, Even if I do not like asking stupid questions, this one bothers me for a while now: Is it possible or even useful to take backup sets of API backup data like the one DB2 creates? If it is not useful to take such backupsets, will the only way to store data offsite be a copy storage pools overflow location? Thanks for helping Lars
Re: what is the problem in the command?
Hi Burak, there's no problem with your command. For safety reasons TSM always wants to read the softlabel on checkin. Only checkout libvol will work with checklabel=barcode. Greetings Lars On 11.10.2002 08:11:14 ADSM: Dist Stor Manager wrote: Hi, I am trying to checkin a volume from BULK on 3583. I was getting no error before. I am using 5.1.1.6 Regards, Burak tsm: TSM01.MBTcheckin libvolume 3583lib 162ACF checklabel=barcode status=pri ANR2020E CHECKIN LIBVOLUME: Invalid parameter - BARCODE. ANS8001I Return code 3.
db2adutl for 64-bit environment
Hi guys and girls, has anyone of you come across this beast: running a 64bit environment we installed the AIX 5.1 version of the TSM client (v. 5.1.1) which contains only the 64bit API. To back up DB2 one uses the db2adutl utility which insists on accessing libApiDS.a that provides the 32bit interface. I worked around this by installing the 32bit API coming with the AIX 4.3 version of the client. Unlike the 64bit version which is inside the BA client fileset, it is installable as a seperate fileset. 64bit environments should be downwards compatible so the db2adutl now runs fine . . . as long as you only run a 32bit DB2 database, right?? What if I would start using a 64bit database. I read there is a way to convert 32bit instances to 64bit ones. Will this also change the behaviour of db2adutl accessing the 64bit API, or is there actually another 64bit db2adutl available?? Any experiences / knowledge would be very appreciated. Thanks Lars
(Silly?) DB2 Backup question
Hi guys and girls, this question is one of the general sort, so forgive me if it has been answered here or in some manual I haven't red yet. Probably, I just need a simpler description. AFAIK, there are three ways to backup a database: 1. Offline filesystem backup 2. Offline backup in the databases quiesced mode via some api 3. Online backup via api The first question I have is about the transaction log archiving. With which of the above does it make sense and why? The second one: does it make sense to run offline AND online backups in a backup cylce, or will one of them be save enough? I read through the TSM Redbook on DB2 UDB implementation. It just explains the ways RDBMS backups can be done not which combinations are useful. As far as I could extract, it is rather useless doing offline backups and log file archiving but no online backups, right?? Thanks in advance Lars
Re: Select for Scratches
Hi Bill, the device class is needed since the output is parsed by an awk script for further processing. Still, you pointed me into the right direction since I remembered we were creating a second device class on the same librarey for testing the other day. So the select shows scratch tapes for both these device classes. I removed the devclass and it works perfectly. Thanks a lot Paul, the device class does not come in here, so this select works w/o problems. Thanks for the fast reply Lars On 19.09.2002 19:33:16 ADSM: Dist Stor Manager wrote: I run something similar every day. I have 1 devclass per library and the command works fine without duplicating the count of scratch tapes. Do you have multiple device classes in the same library? My scratch are not assigned to a specific device class they are scratch for the entire library. If you don't require the deviceclass in the output you could simplify: select library_name, count(*) as volumes from libvolumes where upper(status) like 'SCRATCH%' and library_name='$i' group by library_name You could also improve the scratch query by changing upper(status) like 'SCRATCH%' to status='Scratch' I hope this helps. Bill -Original Message- From: Lars Bebensee [mailto:[EMAIL PROTECTED]] Sent: Thursday, September 19, 2002 5:35 AM To: [EMAIL PROTECTED] Subject: Select for Scratches Hi folks, I am not that good with select queries yes, so I will have to ask one of the SQL gurus here: we wanted to show how many scratch tapes each library has with a query like this: select a.library_name,count(*) as volumes, b.devtype from libvolumes as a, devclasses as b where upper(status) like 'SCRATCH%' and a.library_name ='$i' and a.library_name=b.library_name group by a.library_name,b.devtype $i will be filled with the name of the library before. The funny thing, it shows exactly as twice as many scratches. Is there someone amongst you who can exctract this little monster for me and show me what I am doing wrong. Thanks for helping out Lars
Select for Scratches
Hi folks, I am not that good with select queries yes, so I will have to ask one of the SQL gurus here: we wanted to show how many scratch tapes each library has with a query like this: select a.library_name,count(*) as volumes, b.devtype from libvolumes as a, devclasses as b where upper(status) like 'SCRATCH%' and a.library_name='$i' and a.library_name=b.library_name group by a.library_name,b.devtype $i will be filled with the name of the library before. The funny thing, it shows exactly as twice as many scratches. Is there someone amongst you who can exctract this little monster for me and show me what I am doing wrong. Thanks for helping out Lars
Re: Direct archiving of DB2/SAP R3 logs to TSM
Hi there, the db2 archive logs are still created somewhere on your filesystems. Only after db2 says that it has only commited actions inside the logs it archives the logs to TSM by means of the db2 user exit. You can find out the path of your log files by issueing a db2 get db cfg for DATABASE_NAME. Look for Path to log files. You will find many other interesting things there like the last active log etc. The logs will be restored to this path by the API client that sits between DB2 and TSM meaning you will have to use the various commands db2 offers (like db2adutl, db2 backup|restore database ...). Like Mustafa said, in this case you will have to use db2adutl extract logs LOGNAME or between LOGNAME and LOGNAME Bye Lars On 12.09.2002 06:50:45 ADSM: Dist Stor Manager wrote: Hi TSMers, Please could you assist me. We have set up db2 to archive the logs directly to TSM instead of to a filesystem. The problem I am faced with now is I am not sure how to retrieve these files.do you use TSM or DB2 to retrieve them and where will it put the log files Any ideas. Regards. Lindy Crawford Business Solutions: IT BoE Corporate * +27-31-3642185 ...OLE_Obj... +27-31-3642946
Backupset of HSM data
Hi guys and girls, is it possible to generate backupsets from migrated HSM data?? We are going to rebuild a TSM environment from scratch but want to retain a the HSM data residing on an optical library. Let's suppose we destroy the TSM DB and start with a new clean one. TSM will have forgotten all about the optical volumes as well as the data on it. Backupsets will give me the opportunitiy to restore stand-alone into a newly created HSM-Filesystem. The first option would be copying the data into a disk filesystem large enough to hold all the recalled data (130GB in our case) and later on migrate it back. Apart from that option, what do I do if I haven't got the necessary disk space available? I could store on tape and the only thing that came to my mind was backupsets. I tried to create one using the HSM filespace but it only wrote the directories to tape, not the files. Would this be a case for an export/import? Thanks Lars
Tape Performance on 3580
Hi guys, this is actually not a TSM question but one for the hardware gurus. We are running a backup under AIX using the savevg command and Ultrium 3580 tape drives. With topas I can see a data transfer rate of about 1MB/sec. We attached the drives via Fibre Channel and a SAN Data Gateway. The blocksize of the drive was set to 2048 (since one of the SAN guys told me FC transfers with a blocksize of 2048). I started savevg with -b4 meaning to write 4x512Byte blocks (=2048). With the mentioned speed the drives are seriously crawling. Am I missing out something here? Any clues?? Thanks Lars
Re: INCLEXCL not working as expected on Windows XP client
RTFM Using the Backup-Archive Clients: This is not a bug but the default behaviour of TSM: Tivoli Storage Manager evaluates all EXCLUDE.FS and EXCLUDE.DIR statements first, and removes the excluded filespaces, directories, and files from the list of objects available for processing. You can do something like: INCLUDE C:\Documents and Settings\...\*.* EXCLUDE C:\...\*.* but beware, because your whole directory hierarchy will still be backed up as well as any other drives. On 23.07.2002 10:39:21 ADSM: Dist Stor Manager wrote: We back up some Windows PCs to a TSM 4.1.5.3 server running on AIX 4.3.3. We usually restrict the backups to the C:\My Documents folder by having appropriate INCLEXCL statements in the server's CLIENTOPT set. This has been working just fine for Windows 98 clients. However, with Windows XP clients I have been unable to obtain this desired restricted backup policy. I did install the TSM client 4.2.2.0, following the advice on the ADSM list, so Windows XP support should be OK. I have done many experiments with the INCLEXCL options, and I believe that the following options in the server or client should work: EXCLUDE.DIR C:\* INCLUDE C:\Documents and Settings\...\*.* The second line backs up documents, whereas the first line excludes everything else. However, on Windows XP the first line apparently supercedes the second line, and nothing whatsoever gets backed up !! I think this is a fundamental kind of bug. Can anyone explain why the above example doesn't work as expected ? Even better, how do I achieve the policy of backing up only the Documents and Settings folders ? Thanks a lot, Ole Holm Nielsen Department of Physics Technical University of Denmark
Re: reclamation of unavailable primary tapes
Hi Guys, thanks for the reply. OK, I see TSM will handle more tapes than the library can physically digest. I consulted the TSM Admin Guide on this. The tapes in the overflow location will also age, won't they? Now, what are we going to do about space reclamation of the tapes in the overflow location? Will some operator have to stand by just to exchange tapes? On the other hand, this would be a better solution than checking tapes out of a primary storage pool since no backups have to be deleted in order to supply more scratches. Maintaining a copy storage pool on the 3575 should not pose a problem since the customer has another 3575 in a second location. The DLT will be another story because there is no backup changer or drive, so the created copy storage pool volumes will prove to be quite useless in case of a DR situation. I will have to call the customers attention to this. Thanks anyway. Lars On 12.07.2002 02:03:20 ADSM: Dist Stor Manager wrote: For working disaster recovery your customer will need a *compatible* device to read off-site tapes. In your case if we assume primary tapes are held only on 3575 and DLT autochanger is used only for off-site: - primary pool will use 3575. Library might be small but TSM does have 'MOVE MEDIA' command for that exact purpose. - copypool tapes will be DLT. Whenever set of tapes is built they are checked out and sent off-site. Thus no problems with capacity. - DR site must have DLT drive (not necessarily a library, manual one is just fine) which should be able to read tapes written by autoloader. If autoloader is DLT7000 drive might be either DLT7000 or DLT8000 but if the autoloader is DLT8000 the drive must be the same. If some primary pools are backed up to a copypool on 3575 then DR site must have 3570! Doublecheck which is backed up where. - do not forget to create (and send off-site) DB backups or snapshots after backup stg finishes. Zlatko Krastev IT Consultant
Re: AIX Question - Scripts
Hi, look at your if condition. You should test for something there like if [ -x /bin/mksysb ]; then The same applies to the savevg script. What tape drive do you use? I would check that one before disabling software compression. cya Lars On 12.07.2002 10:22:41 ADSM: Dist Stor Manager wrote: Hi Please could one of you look at my aix script to do a savevg backup and rootvg backupit does not run... What have I done wrong ? Thanks Lindy Crawford Business Solutions: IT BoE Corporate * +27-31-3642185 +27-31-3642946 [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] WARNING: Any unauthorised use or interception of this email is illegal. If this email is not intended for you, you may not copy, distribute nor disclose the contents to anyone. Save for bona fide company matters, BoE Ltd does not accept any responsibility for the opinions expressed in this email. For further details please see: http://www.boe.co.za/emaildisclaimer.htm - sapvg.txt - rootvg.txt
Re: reclamation of unavailable primary tapes (was: w/o subject)
Hi, First, sorry for sending the message without subject, I guess the human brain is not supposed to to more than one thing at the same time (especially mine ;-) ). I already knew the question of why sending primary tapes offsite will pop up. The only thing that came to my mind was that the library is not big enough to hold copies of all the important storage pools. The customer already uses a DLT-Autochanger as copy destination for some of the primary storage pools. We are currently busy redesigning the implementation. I am sure we will not get the customer convinced to buy a bigger library. So what else could possibly be done to prevent removing tapes from a primary pool? The customer's environment consists of: 1x 3575-L18 with 3x3570 (C-Drives) 1x DLT Autochanger 7 Slots with 70GB per tape Client data is about 800GB I am not expecting a detailed draft but hints are very welcome Thanks Lars On 10.07.2002 03:53:31 ADSM: Dist Stor Manager wrote: Lars, I do not know what is the idea behind sending primary tapes off-site (DR ??) but this does not help at all in case tape gets somehow damaged. You will learn it the hard way - when you need to restore from the broken tape. Migration threshold set to anything between 0 and 100% will not force/prevent reclamation. They are just different You already gave the answer: update the volumes to readonly or readwrite, check them in the library and lower reclamation threshold (in any order). BTW: Subjects are very useful to track answers on a mail list with many posts (like this one). Burak, offsite is for copypool volumes only (look at note 3 for UPD VOL). Primary ones have to be unavailable. Zlatko Krastev IT Consultant Please respond to ADSM: Dist Stor Manager [EMAIL PROTECTED] Sent by:ADSM: Dist Stor Manager [EMAIL PROTECTED] To: [EMAIL PROTECTED] cc: Subject: yes, you may change the accesses to readwrite and then reclaim but you should normally use offsite not unavailable access regards, burak [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 09.07.2002 16:23 Please respond to ADSM-L To: [EMAIL PROTECTED] cc: Subject: Hi TSM Guys, I have a customer who moves volumes from a primary storage pool to an offsite location by setting the volumes to UNAVAILABLE. The storage pool's migration threshold is set to 100% for spacerclamation to be prevented. What do I have to consider if I want to manually start space reclamation on the unavailable volumes? Will it be enough to just check-in the volumes, change their access to ReadWrite, and then start the space reclamation by lowering the reclamation threshhold gradually? Thanx a lot Lars
[no subject]
Hi TSM Guys, I have a customer who moves volumes from a primary storage pool to an offsite location by setting the volumes to UNAVAILABLE. The storage pool's migration threshold is set to 100% for spacerclamation to be prevented. What do I have to consider if I want to manually start space reclamation on the unavailable volumes? Will it be enough to just check-in the volumes, change their access to ReadWrite, and then start the space reclamation by lowering the reclamation threshhold gradually? Thanx a lot Lars
DB2 User Exit Problem
Hi TSM guys, good that there is a list like this one. Hope you can help me with a problem on a customers site: The customer runs a DB2 EEE installation across two SP nodes. He uses the DB2 user exit for backing up the DB2 transaction logs. Before backing them up they are copied to a secondary log directory to be able to keep a second backup of the logs. Here is the probelm: I recently upgraded the client code from 3.1.20 to 3.7.2. This meant I had to recompile and link the user exit code. The unmodified sample code that gets delivered with the DB2 installation runs fine for the backing up part. Unfortunately, no secondary logs are created anymore so I guess whoever did the implementation before added this copy functionality to the sample code. I found the old code but after compiling and linking it to the new client libraries it doesn't work the way it used to. Checking the ARCHIVE.LOG, I saw that the log is archived, and it is also copied over to the secondary location. However, the user exit doesn't return anymore and just hangs blocking the next archiving request. Did anyone of you guys use the same way to keep a secondary log backup? If yes, what did you change in the user exit sample code? Many greetings Lars