DB PiT restore with 6.2.1
Hi, We are new to TSM6 and are testing TSM DB restore using TSM 6.2.1 on AIX 6. Restoring to the most current state works fine. Point in time restore seems to work only to the nearest full or incremental backup, e.g. if we have backups: Full 10:00 Incr 11:00 Incr 12:00 and want to restore to PiT 11:30 with dsmserv restore db todate=today totime=11:30:00 it restores to 11:00 only. And when restoring it says (among other ANR*I messages): ANR4645I The restore date reflects the most recent backup available up to the specified TODATE (BTW - no W or E ANR messages during the restore!) So I can't restore to exactly 11:30 in the above example, even if all Logs till 12:00 should be in the backup? Is this the way it is supposed to work or what am I doing wrong? My understanding from the Admin Guide (which under Point-in-time restore in Chap. 25 p. 773 states Applies logs from the overflow directory up to specified point in time as the last action ) was, that a restore to 11:30 in the above example should be possible. TIA Andreas -- -- Andreas Priebe Data Center PROMOS consult Projektmanagement, Organisation und Service GmbH Rungestraße 19 10179 Berlin Telefon: +49(0)30 - 24 31 17 - 613 Mobil: +49(0)151 52 72 89 13 Fax: +49(0)30 - 24 31 17 - 729 andreas.pri...@promos-consult.de www.openpromos.com Bitte denken Sie an die Umwelt, bevor Sie diese E-Mail drucken. - Besuchen Sie uns auf folgenden Veranstaltungen: 8. OpenPromos Anwenderforum 16. - 17.03.2011, Berlin VdW Rheinland Westfalen 10. Forum Wohnungswirtschaft 21. - 22. 06. 2011 in Düsseldorf ImmoCom 2011 08. - 09.09.2011, Berlin 14. SAP-Kongress für die Immobilienwirtschaft 22. - 23.09.2011, Potsdam Verbandstag VdW Rheinland Westfalen 27.-28. 09. in Leverkusen 14. Expo Real 04. - 06.10.2011, München - Für aktuelle Informationen über unsere Neuentwicklungen und Projekte bestellen Sie unseren PROMOS-Newsletter! http://www.openpromos.com/ger/kontakt/newsletter_anmeldung.php - PROMOS consult Projektmanagement, Organisation und Service GmbH Sitz der Gesellschaft: Berlin Amtsgericht Charlottenburg, HRB 108478 B Geschäftsführer: Dipl.-Ing. Jens Kramer, Dipl.-Inf. Volker Schulz
Re: Reaggregate data in copy storage pool (DR550 with SSAM)
Hi Rajesh, I have a similar setup as of yours. And we used tape copy storage pool to copy the data from primary storage pool. Can you brief what you are trying to do with re-aggregation in the copy storage pool. Due to the large number of small files in the storage pool, we get only about 600 to 1000 MB (cf. q vol) into a 5 GB data file (as seen by ls -l), i.e. about 15 .. 20%. Unfortunately a tape copy storage pool is not an option currently. but we are working on it ;-) Regards, Andreas -- -- Andreas Priebe Director Data Center Data Center PROMOS consult Projektmanagement, Organisation und Service GmbH Rungestraße 19 10179 Berlin Telefon: +49(0)30 - 24 31 17 - 613 Mobil: +49(0)151 52 72 89 13 Fax: +49(0)30 - 24 31 17 - 729 andreas.pri...@promos-consult.de www.openpromos.com Bitte denken Sie an die Umwelt, bevor Sie diese E-Mail drucken. - Besuchen Sie uns auf folgenden Veranstaltungen: 8. OpenPromos Anwenderforum 16. - 17.03.2011, Berlin ImmoCom 2011 08. - 09.09.2011, Berlin 14. SAP-Kongress für die Immobilienwirtschaft 22. - 23.09.2011, Potsdam 14. Expo Real 04. - 06.10.2011, München - Für aktuelle Informationen über unsere Neuentwicklungen und Projekte bestellen Sie unseren PROMOS-Newsletter! http://www.openpromos.com/ger/kontakt/newsletter_anmeldung.php - PROMOS consult Projektmanagement, Organisation und Service GmbH Sitz der Gesellschaft: Berlin Amtsgericht Charlottenburg, HRB 108478 B Geschäftsführer: Dipl.-Ing. Jens Kramer, Dipl.-Inf. Volker Schulz
Reaggregate data in copy storage pool (DR550 with SSAM)
Hi all, best wishes for the New Year! - And a question :-) We have a DR500 - this is a IBM solution containing a pSeries, some FC storage and running AIX with a TSM/SSAM; but I think the same question would arise in a normal TSM. Our DR550 is running SSAM 5.4: tsm: DR550PP q status Storage Management Server for AIX-RS/6000 - Version 5, Release 4, Level 1.0 Data are fed into this DR550 via API from a CommonStore for SAP. As I understand it, the CommonStore sends the data one by one, so the data in the primary pool is not aggregated (verified by select ... from contents). I have defined a copy storage pool (devclass type FILE) and would like the data to be (re-)aggregated in the the copy pool (there are many very small files and we ususally get only about 600 to 1000 MB into a 5 GB data file volume - as seen by q vol). Is this possible? Any ideas? Regards, Andreas -- -- Andreas Priebe Service Leiter Rechenzentrum PROMOS consult Projektmanagement, Organisation und Service GmbH Rungestraße 19 10179 Berlin Telefon +49(0)30 24 31 17 - 613 Mobil +49(0)151 52 72 89 13 Fax +49(0)30 24 31 17 - 729 andreas.pri...@promos-consult.de www.openpromos.com Bitte denken Sie an die Umwelt, bevor Sie diese E-Mail drucken. - Besuchen Sie uns auf folgenden Veranstaltungen: 8. OpenPromos Anwenderforum 16. - 17.03.2011, Berlin 14. SAP-Kongress für die Immobilienwirtschaft 22. - 23.09.2011, Potsdam ImmoCom 2011 08. - 09.09.2011, Berlin 14. Expo Real 04. - 06.10.2011, München - Für aktuelle Informationen über unsere Neuentwicklungen und Projekte bestellen Sie unseren PROMOS-Newsletter! http://www.openpromos.com/ger/kontakt/newsletter_anmeldung.php - PROMOS consult Projektmanagement, Organisation und Service GmbH Sitz der Gesellschaft: Berlin Amtsgericht Charlottenburg, HRB 108478 B Geschäftsführer: Dipl.-Ing. Jens Kramer, Dipl.-Inf. Volker Schulz
ACN5759W with TDP for Exchange when using RSG
Hi, we are trying to restore a mailbox to Exchange 2003 using the recovery storage group (RSG) feature. The RSG is not mounted, the two other SGs are mounted. When we start recovery via the DP-GUII I get a ACN5759W warning: At least one of the databases that is selected for restore is mounted. Do you want to dismount the database(s)? This seems confusing, as the manual says in Appendix A: The Data Protection for Exchange GUI will display a text message to remind you that all mailbox database restores will go to the Recovery Storage Group if a Recovery Storage Group exists. Instead it gives us the ACN5759W - so we are afraid that we destroy our storage group and decided to abort the recovery. Versions used: DP for Exchange is 5.2.1, TSM is 5.4 Any ideas? TIA Andreas -- -- Andreas Priebe Service Leiter Rechenzentrum PROMOS consult Projektmanagement, Organisation und Service GmbH Rungestraße 19 10179 Berlin Telefon +49(0)30 24 31 17 - 613 Mobil +49(0)151 52 72 89 13 Fax +49(0)30 24 31 17 - 729 andreas.pri...@promos-consult.de www.openpromos.com - Besuchen Sie uns auf folgenden Veranstaltungen: EXPO REAL 13. Internationale Fachmesse für Gewerbeimmobilien 04. - 06.10.2010, München 8. OpenPromos Anwenderforum 16. - 17.03.2011, Berlin - Für aktuelle Informationen über unsere Neuentwicklungen und Projekte bestellen Sie unseren PROMOS-Newsletter! http://www.openpromos.com/ger/kontakt/newsletter_anmeldung.php - PROMOS consult Projektmanagement, Organisation und Service GmbH Sitz der Gesellschaft: Berlin Amtsgericht Charlottenburg, HRB 108478 B Geschäftsführer: Dipl.-Ing. Jens Kramer, Dipl.-Inf. Volker Schulz
Re: ACN5759W with TDP for Exchange when using RSG
Hi Del, It seems that Exchange does not think the Recovery Storage Group that you created is actually a Recovery Storage Group. I have seen them created incorrectly before. I recommend removing the Recovery Storage Group, recreate it again using the Exchange Management tools, and then retry DP/Exchange. We have recreated the RSG - ACN5759W remains. Interestingly enough in the GUI it says, that it has found a RSG and will restore into the RSG. If you still see the message, check to make sure the following registry entry is NOT set: HKLM\System\CurrentControlSet\Services\MSExchangeIS\Parameters\System “Recovery SG Override” It is NOT set, so this seems OK. Greetings, Andreas -- -- Andreas Priebe Service Leiter Rechenzentrum PROMOS consult Projektmanagement, Organisation und Service GmbH Rungestraße 19 10179 Berlin Telefon +49(0)30 24 31 17 - 613 Mobil +49(0)151 52 72 89 13 Fax +49(0)30 24 31 17 - 729 andreas.pri...@promos-consult.de www.openpromos.com - Besuchen Sie uns auf folgenden Veranstaltungen: EXPO REAL 13. Internationale Fachmesse für Gewerbeimmobilien 04. - 06.10.2010, München 8. OpenPromos Anwenderforum 16. - 17.03.2011, Berlin - Für aktuelle Informationen über unsere Neuentwicklungen und Projekte bestellen Sie unseren PROMOS-Newsletter! http://www.openpromos.com/ger/kontakt/newsletter_anmeldung.php - PROMOS consult Projektmanagement, Organisation und Service GmbH Sitz der Gesellschaft: Berlin Amtsgericht Charlottenburg, HRB 108478 B Geschäftsführer: Dipl.-Ing. Jens Kramer, Dipl.-Inf. Volker Schulz
Re: ACN5759W with TDP for Exchange when using RSG
Hi, we are trying to restore a mailbox to Exchange 2003 using the recovery storage group (RSG) feature. The RSG is not mounted, the two other SGs are mounted. When we start recovery via the DP-GUII I get a ACN5759W warning: At least one of the databases that is selected for restore is mounted. Do you want to dismount the database(s)? This seems confusing, as the manual says in Appendix A: The Data Protection for Exchange GUI will display a text message to remind you that all mailbox database restores will go to the Recovery Storage Group if a Recovery Storage Group exists. Instead it gives us the ACN5759W - so we are afraid that we destroy our storage group and decided to abort the recovery. Versions used: DP for Exchange is 5.2.1, TSM is 5.4 It was just my fault :-( The manual says explicitly: Make sure that no public folders within the storage group are selected. When you start marking the last needed INCRemental in the GUI, it marks the publice folders from the FULL automatically. Excluding the public folders from restore made the restore start without any ANC messages. We are now awaiting the 200 GB from tape - as of now we got already about 15 GB into the RSG. Greetings, Andreas -- -- Andreas Priebe Service Leiter Rechenzentrum PROMOS consult Projektmanagement, Organisation und Service GmbH Rungestraße 19 10179 Berlin Telefon +49(0)30 24 31 17 - 613 Mobil +49(0)151 52 72 89 13 Fax +49(0)30 24 31 17 - 729 andreas.pri...@promos-consult.de www.openpromos.com - Besuchen Sie uns auf folgenden Veranstaltungen: EXPO REAL 13. Internationale Fachmesse für Gewerbeimmobilien 04. - 06.10.2010, München 8. OpenPromos Anwenderforum 16. - 17.03.2011, Berlin - Für aktuelle Informationen über unsere Neuentwicklungen und Projekte bestellen Sie unseren PROMOS-Newsletter! http://www.openpromos.com/ger/kontakt/newsletter_anmeldung.php - PROMOS consult Projektmanagement, Organisation und Service GmbH Sitz der Gesellschaft: Berlin Amtsgericht Charlottenburg, HRB 108478 B Geschäftsführer: Dipl.-Ing. Jens Kramer, Dipl.-Inf. Volker Schulz
Strange filesize after export/import
Hi, we are just testing the migration of a TSM 5.1.9.5 (AIX 4.3) to a TSM 5.2.7.2 (AIX 5.2). So far everything seems OK, policies and data were exported and imported just fine, but I noticed a discrepancy in the file sizes shown in a select from contents for ALL files (data are from a ContenManager)! This is an example: On the old TSM tsm: ECKWANDselect * from contents where file_name='/MC_ARCHIV_PO4_1/ MC_ARCHIV_PO4_1AKKDBJHB.DMD' VOLUME_NAME: /tivoli1/stgpool/archiv_po4_1_file01 NODE_NAME: ARCHIV_PO TYPE: Bkup FILESPACE_NAME: /OBJRZ/FRN/ADSM/192.168.126.134/MC_ARCHIV_PO4_1 FILE_NAME: /MC_ARCHIV_PO4_1/ MC_ARCHIV_PO4_1AKKDBJHB.DMD AGGREGATED: No FILE_SIZE: 56210 SEGMENT: CACHED: No FILESPACE_ID: 6 FILESPACE_HEXNAME: FILE_HEXNAME: After the import on the new TSM: tsm: TEST-KANDELselect * from contents where file_name='/MC_ARCHIV_PO4_1/ MC_ARCHIV_PO4_1AKKDBJHB.DMD' VOLUME_NAME: /usr/local/tsm/server2/data/apo4_01 NODE_NAME: ARCHIV_PO TYPE: Bkup FILESPACE_NAME: /OBJRZ/FRN/ADSM/192.168.126.134/MC_ARCHIV_PO4_1 FILE_NAME: /MC_ARCHIV_PO4_1/ MC_ARCHIV_PO4_1AKKDBJHB.DMD AGGREGATED: 124/256 FILE_SIZE: 11195180 SEGMENT: CACHED: No FILESPACE_ID: 1 FILESPACE_HEXNAME: FILE_HEXNAME: Note the difference in FILE_SIZE: 56210 vs. 11195180! This seems strange! And what does the difference in the field AGGREGATED mean (No vs. 124/256)? Interestingly enough, I can restore the object from the new TSM and inspect it in a viewer (it's a TIFF) just fine. The dsmc query shows it as: tsm query backup /OBJRZ/FRN/ADSM/192.168.126.134/MC_ARCHIV_PO4_1//MC_ARCHIV_PO4_1/MC_ARCHIV_PO4_1AKKDBJHB.DMD Size Backup DateMgmt Class A/I File ----- --- API 55,858 B 03/25/01 00:11:12MC_ARCHIV_ A /OBJRZ/FRN/ADSM/192.168.126.134/MC_ARCHIV_PO4_1/MC_ARCHIV_PO4_1/MC_ARCHIV_PO4_1AKKDBJHB.DMD Another thing is puzzeling me: the blank in the filename in the select, which is not appearing in the query backup result. Any hints? TIA Andreas
Restoring overwritten archived object
Hi, I have an archived object in TSM (archived by a broken and very old version of TDP for SAP - don't ask why). This software archived two different objects under the same name (SAP___9F10111931_0, see below). I need to restore the older one. Here ist the relevant output from select * from archives ... NODE_NAME: BF4 FILESPACE_NAME: /tdpmux FILESPACE_ID: 6 TYPE: FILE HL_NAME: / LL_NAME: SAP___9F10111931_0 OBJECT_ID: 308358362 ARCHIVE_DATE: 2005-10-11 19:31:29.00 OWNER: DESCRIPTION: [EMAIL PROTECTED] 1.COPY 0 CLASS_NAME: MC_LOG_1 NODE_NAME: BF4 FILESPACE_NAME: /tdpmux FILESPACE_ID: 6 TYPE: FILE HL_NAME: / LL_NAME: SAP___9F10111931_0 OBJECT_ID: 308358364 ARCHIVE_DATE: 2005-10-11 19:31:37.00 OWNER: DESCRIPTION: [EMAIL PROTECTED] 0 CLASS_NAME: MC_LOG_1 When using TDP it will restore the second one, I need the first one (the one archived at 19:31:29). How would I do this with TSM tools (preffered command line, as I habe several such cases)? TIA for any suggestions. Andreas BTW I can't see any of them in the graphical dsm client.
Strange thing labeling UDO in Plasmon G24
Hi, we have a new Plasmon G24 library with two UDO1 drives (no barcode reader) attached (SCSI) to a TSM 5.2.4.5 (64bit) on a p630 running AIX 5.2 ML6. According to the TSM support pages this is a (well?) supported configuration. I observed two things which make me smell trouble: 1. After defining library, drives and paths (see below) the drives are listed as of drive type REMOVABLEFILE. From my experience with IBM MOs I remember that they appeared as device type OPTICAL. What is the expected benaviour? 2. When I try two label a UDO medium (again, details below), I get an ANR8807W: ANR8807W Could not write label U0 on the volume in drive UDO0 (/dev/op0) of library PLAS01 because volume is already labelled ROOT. This is a brand new medium! Trying to overwrite the label fails miserably: ANRD pvr.c(5796): ThreadId51 writeLabel vector not valid for drive UDO0. Callchain of previous message: 0x000100017d94 outDiagf - 0x0001000ecd84 PvrWriteLabel - 0x00010022a3cc ReadWriteLabel - 0x00010022b4bc SingleEEVolumeLabel - 0x00010022c8f0 EELabelVolume - 0x000100241598 ScsiLabelVolume - 0x00010024bd00 MmsLtsLabelVolume - 0x00010012db08 LabelThread - 0x00018078 StartThread - 0x092ef460 _pthread_body - ANR8806E Could not write volume label U0 on the tape in library PLAS01. Any hints for me??? TIA, Andreas === More technical details: 1. def libr plas01 libt=scsi def path kandel plas01 srctype=server desttype=library - device=/dev/lb0 def dr plas01 udo0 element=6001 online=yes def dr plas01 udo1 element=6002 online=yes def path kandel udo0 srctype=server desttype=drive - library=plas01 device=/dev/op0 def path kandel udo1 srctype=server desttype=drive - library=plas01 device=/dev/op1 def devcl UDO01 libr=plas01 devt=worm format=drive q dr Library Name Drive Name Device Type On-Line --- --- [...] PLAS01 UDO0 REMOVABLEF- Yes ILE PLAS01 UDO1 REMOVABLEF- Yes ILE 2. label libvol plas01 search=bulk checkin=private labels=prompt repl 4 repl 5 LABEL=U0 ANR8807W Could not write label U0 on the volume in drive UDO0 (/dev/op0) of library PLAS01 because volume is already labelled ROOT. label libvol plas01 search=bulk checkin=private labels=prompt OVERWRITE=YES repl 9 repl 10 LABEL=U0
Re: Strange thing labeling UDO in Plasmon G24
Hi, just to let you know: My problem with UDO-Drives appearing as drive type REMOVABLEFILE is resolved: IBM note 1161410 says: Use raw devices for Opticals So Ihad to do def path kandel udo0 srctype=server desttype=drive - library=plas01 device=/dev/rop0 After this correction the second problem went away. Greetings, Andreas
Re: LTO Volume
Hi Richard, The Select is tried-and-true, and works well. On one of my TSM servers it gives me an SQL error due to the formmating: ANR2948E The value '381468.0' can not be assigned or converted to SQL data type DECIMAL(6,1). | V.. ZED as PCT UTIL , DEC(EST_CAPACITY_MB,6) AS MB_CAPACITY , C ANS8001I Return code 26. As the original poster mentioned LTO, maybe it would be better to change this part to DEC(EST_CAPACITY_MB,7) (until the next big step in increasing type capacity :-) ) Greetings, Andreas
Re: LTO Volume
Hi Iain tsm: ABZBKUP001SELECT STGPOOL_NAME AS STORAGE POOL , - cont CHAR(VOLUME_NAME,6) AS VOLNAME, PCT_UTILIZED as PCT UTIL, - cont DEC(EST_CAPACITY_MB,7) AS MB_CAPACITY, CHAR(ACCESS,11) AS ACCESS, - cont LEFT(CHAR(LAST_WRITE_DATE),19) AS LAST_WRITE_DATE - cont FROM VOLUMES WHERE STATUS='FILLING' ORDER BY STGPOOL_NAME ANR2948E The value '4096000.0' can not be assigned or converted to SQL data type DECIMAL(7,1). | V.. ED as PCT UTIL, DEC(EST_CAPACITY_MB,7) AS MB_CAPACITY, CH ANS8001I Return code 26. Yours said: value '4096000.0', mine said: value '381468.0', so counting megabytes suggests, maybe in your case it should be at least DEC(EST_CAPACITY_MB,8) ? Regards, Andreas
q assoc vs. select * from associations
Hi, I notice, that q assoc gives me a complete list of associations, while a select * from associations misses those domain/schedule pairs, which are not associated with any node. Any reason for that? TIA Andreas
Re: q assoc vs. select * from associations
Hi Andy [...] associated nodes, then there are no associations for that schedule. But maybe the original developers of the QUERY ASSOCIATION command thought it would be useful to show you which schedules had no associations. [...] That was exacly the reason, why I noticed the difference: I was looking for a (fast) select statement, which shows defined schedules which are not associated with any node (maybe simply forgotten or maybe on purpose, e.g. to assocate them later). So it seems I have to use a (slightly?) more complex statement. Thanks for your answer, Andreas
32bit API on AIX5 needed
Hallo, we have an application (IBM CommonStore for SAP 8.2) which seems to insist on the 32bit libApiDS.a, but we have only the 64bit API available (BAAPI client version is 4.2.3.3 on AIX 5.2). As far as I can see there is no 32bit API for AIX 5. For the moment I have dropped in a libApiDS.a from an AIX4.3.3 installation, which seems to get me further - but that can't be the solution? What about other people running 32bit applications on AIX5? Any ideas welcome, Andreas