none
DMP 2012 SP1 tape backup questions, errors and issues RRS feed

  • Question

  •  I'm using Microsoft Data Protection Manager 2012 SP1 with a Quantum Superloader LTO-4 library (SAS interface) and Windows Server 2008 R2 SP1. This library is included in the list of tape units certificated to work with DMP 2012 SP1 (http://technet.microsoft.com/en-us/library/hh916523.aspx)

    This is the first time I'm using DPM, but I'm very familiarized with other tape backup applications (Backup Exec, HP Data Protector Express...).

    I need help with some issues in tasks that are easily accomplished in other backup applications:

    - All backup Jobs are showing error messages with iD 24052 regarding to no availability of the library. Looking into the monitor section of DPM I can see that for each protected member included in the backup job there are four messages: first of them (in time) indicates that the backp of this protected member is ok. The other three have the 24052 error, and they are generated together for all of the protected members almost at the end of the backup job. I tested the backups and I can restore information correctly, but obviously I could not did the test with all the backed up information, so I'm not totally sure of being completely protected. What do these final and recurrente erros mean? What I'm doing bad?

    - Tape media names are always the same. I have configured only long term protection, but with three levels: daily (four days with one week retention), weekly (four weeks with four weeks retention) and monthly (two months with two months retention). When defining the protection group I can configure names for the tape used each day (day1, day2, week3, month2...), but when the job finished all the tapes in the same level have the same name (for example: day1 in the daily ones), and DPM append an incremental number to the name. This looks like: day1-00000007, day1-00000011...; the same for the weekly and monthly tapes. I need correct and fixed names being asigned to each tape to be able to identify them correctly in the rotation and when I move them inside and outside of the library for off-site protection.

    - I would like to make incremental backups with one week retention, and weekly backups with four weeks retention. I tried using short time protection, but if I select four weeks retention it applies to incremental backups too, and I don't want this. Actually I'm using only long term protection, and I'm getting the retention periods I want (one week for daily backups, four weeks for weekly backups and two months for monthly backups) but I would prefer that the daily jobs were incremental backups, and not complete backups. Is there any way to do this?

    - Is there any way to define more than a protection group for the same protected member? I'd like to make disk backup of some of the protected members for quickly recover, but maintaining them into the tape backup too. If I use short term protection to disk in the main job all of the protected members would be affected, and I can't do that (too much space wasted).

    Wednesday, May 8, 2013 1:34 PM

All replies

  • Hi,

    How many tape drives to you have installed in the Quantum SuperLoader ?   Do you see that tape copy jobs fail ?  Can you supply failed job details.

    1) See if the following blog helps.

    http://blogs.technet.com/b/dpm/archive/2012/02/22/dpm-tape-jobs-take-longer-than-expected-or-fail-with-id-24052.aspx

    2) Tape media names are constructed as per your PG settings and as noted, DPM appends the incremental number.  There is no provision for specifying different days of the week for tape labels unless you manually change it yourself every day.

    3) All long term tape backup are full backups.

    4) No, once a data source is part of a protection group, it cannot be added to a second protection group. If you need different recovery goals for different data sources, use separate protection groups for each data source.


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread. Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.

    Wednesday, May 8, 2013 4:05 PM
    Moderator
  • Hi Mike:

    1) I have read this article before posting here. I tried the PS script in  my system, buit it fails with an SQL Server connection error.

    Library has only one drive, and there is no cleaning problem: unit is new, tapes are new, cleaning tape is new and since I note the errors I make a manual cleaning each day before backup Jobs. Also unit is not showing any alarm. Backup logs have information that I would prefear no to make public, but I'm including a serie of events at the end of this reply. This serie of events repeats every day.

    2) I understand what you're saying about not fixed name for each particular day of the week, but if in the PG settings you can specify diferent first part names (daily1, daily2, daily3 and daily4 for example) for different tapes (four tapes for example) in the same level (daily backups) why DPM it is always naming with the same first part name (daily1) all the tapes using in daily backup jobs? I think that maybe I do something wrong, don't you?

    3) Of course; it makes no sense to use incremental backups for off-site protection using tapes, but if you want to combine incremental for short term and full for long term you are very limited, since DPM force you to do the daily incrementals combined with weekly backups using the same retention period in both. If you want to have weekly backups retained for four weeks you need to do the same with the daily backups. Since each tape job requieres a different tape this would mean 16 tapes only for incremental backups. Is there any other more eficient way to do this?

    4) Serious limitation I think, since different protection groups requieres diferent tapes in DPM.

    5) One last question: I'm thinking about a way to get the backup strategy I need. If I use weekly complete backup Jobs in short term and long term protection, and then I schedule the two weekly jobs same week day at same hour, would the long term backup job disable the short term backup job? This is the way that jobs inside the long term protection can be configured to work, but I don't know if this works too with jobs in diferent protection levels (short, long).

    *****DPM logs*****

    * Each day backup job starts at 01:00. Protected group include three diferent servers and a total of 19 protected members. XXXXX-XXX-XXXX.XXXX.XX.XXXXX is the DPM server and the one the library is conected to.

    1) Log "DPM Backups events":

    - Each day starts with 19 identical and consecutive 33222 informational events, indicating that the backup for the protected member referenced in the event was ok. One event for each protection member of the protection group. These are the first three warning events of yesterday:

    ________________________________________
    Nombre de registro:DPM Backup Events
    Origen:        DPM-BackupEvent
    Fecha:         08/05/2013 2:04:45
    Id. del evento:33222
    Categoría de la tarea:Ninguno
    Nivel:         Información
    Palabras clave:Clásico
    Usuario:       No disponible
    Equipo:        XXXXX-XXX-XXXX.XXXX.XX.XXXXX
    Descripción:
    El trabajo de copia de seguridad para el origen de datos: D:\ en el servidor de producción: YYYYY-YYY-YYYY.YYYY.YY.YYYYY se completó correctamente.
     Punto de recuperación creado a las: 08/05/2013 1:00:46.
     Tipo de copia de seguridad: copia de seguridad tape.
    _________________________________________

    _________________________________________
    Nombre de registro:DPM Backup Events
    Origen:        DPM-BackupEvent
    Fecha:         08/05/2013 2:05:39
    Id. del evento:33222
    Categoría de la tarea:Ninguno
    Nivel:         Información
    Palabras clave:Clásico
    Usuario:       No disponible
    Equipo:        XXXXX-XXX-XXXX.XXXX.XX.XXXXX
    Descripción:
    El trabajo de copia de seguridad para el origen de datos: XXXXX-XXX-XXXX.XXXX.XX.XXXXX\MSDPM2012\model en el servidor de producción: XXXXX-XXX-XXXX.XXXX.XX.XXXXX se completó correctamente.
     Punto de recuperación creado a las: 08/05/2013 2:04:55.
     Tipo de copia de seguridad: copia de seguridad tape.
    _________________________________________

    _________________________________________
    Nombre de registro:DPM Backup Events
    Origen:        DPM-BackupEvent
    Fecha:         08/05/2013 2:30:45
    Id. del evento:33222
    Categoría de la tarea:Ninguno
    Nivel:         Información
    Palabras clave:Clásico
    Usuario:       No disponible
    Equipo:        XXXXX-XXX-XXXX.XXXX.XX.XXXXX
    Descripción:
    El trabajo de copia de seguridad para el origen de datos: D:\ en el servidor de producción: XXXXX-XXX-XXXX.XXXX.XX.XXXXX se completó correctamente.
     Punto de recuperación creado a las: 08/05/2013 2:05:43.
     Tipo de copia de seguridad: copia de seguridad tape.
    _________________________________________

     


    - Last one of this events is logged at 08:20.
    - After last 33222 informational event DPM starts to log 57 (19 x 3) identical and consecutive 33223 warning events, indicating that the backup for the protected member referenced in the event could not be done. Three non consecutive but identical events for each protection member on the protection group. Diferent protected member order from the 33222 infromational events. These are the first five warning events of yesterday:

    _________________________________________
    Nombre de registro:DPM Backup Events
    Origen:        DPM-BackupEvent
    Fecha:         08/05/2013 8:40:04
    Id. del evento:33223
    Categoría de la tarea:Ninguno
    Nivel:         Advertencia
    Palabras clave:Clásico
    Usuario:       No disponible
    Equipo:        XXXXX-XXX-XXXX.XXXX.XX.XXXXX
    Descripción:
    No se pudo realizar la copia de seguridad del origen de datos: D:\ en el servidor de producción: YYYYY-YYY-YYYY.YYYY.YY.YYYYY.
     Error en la copia de seguridad a las: 08/05/2013 8:21:00.
     Tipo de copia de seguridad: copia de seguridad tape.
    _________________________________________

    _________________________________________
    Nombre de registro:DPM Backup Events
    Origen:        DPM-BackupEvent
    Fecha:         08/05/2013 8:40:04
    Id. del evento:33223
    Categoría de la tarea:Ninguno
    Nivel:         Advertencia
    Palabras clave:Clásico
    Usuario:       No disponible
    Equipo:        XXXXX-XXX-XXXX.XXXX.XX.XXXXX
    Descripción:
    No se pudo realizar la copia de seguridad del origen de datos: D:\ en el servidor de producción: XXXXX-XXX-XXXX.XXXX.XX.XXXXX
     Error en la copia de seguridad a las: 08/05/2013 8:21:30.
     Tipo de copia de seguridad: copia de seguridad tape.
    _________________________________________


    _________________________________________
    Nombre de registro:DPM Backup Events
    Origen:        DPM-BackupEvent
    Fecha:         08/05/2013 8:40:05
    Id. del evento:33223
    Categoría de la tarea:Ninguno
    Nivel:         Advertencia
    Palabras clave:Clásico
    Usuario:       No disponible
    Equipo:        XXXXX-XXX-XXXX.XXXX.XX.XXXXX
    Descripción:
    No se pudo realizar la copia de seguridad del origen de datos: XXXXX-XXX-XXXX.XXXX.XX.XXXXX\MSDPM2012\model en el servidor de producción: XXXXX-XXX-XXXX.XXXX.XX.XXXXX.
     Error en la copia de seguridad a las: 08/05/2013 8:22:01.
     Tipo de copia de seguridad: copia de seguridad tape.
    _________________________________________

     

     

    - Last one of this events is logged at 08:50, and there are no more events in all day in this log.

     

     

     

    2) Log "DPM Backups events":

    One daily error talknig about that an automated inventary job (I have not programed it) is not able to find a tape because the tape is in the drive. I suppose that this one is not related with the other erros, but I'm reporting it due to the time coincidence.

    _________________________________________
    Nombre de registro:DPM Alerts
    Origen:        DPM-EM
    Fecha:         08/05/2013 9:02:12
    Id. del evento:3316
    Categoría de la tarea:Ninguno
    Nivel:         Error
    Palabras clave:Clásico
    Usuario:       No disponible
    Equipo:        XXXXX-XXX-XXXX.XXXX.XX.XXXXX
    Descripción:

    Error en el inventario detallado de cintas por el siguiente motivo: (ID: 3316)
    No se pudo realizar la operación porque la ranura 8 de la biblioteca Quantum SuperLoader estaba vacía. Esta cinta se encuentra actualmente en Drive Quantum LTO 4 Tape Drive (ID: 24027)
    _________________________________________

    Wednesday, May 8, 2013 10:14 PM
  • Hi,

    1) For the script, you need to change the instance name from msdpm2010 to msdpm2012

    $instance = '.\msdpm2010' # <---- If DPM Database is on a different location, edit this line accordingly.

    Can you make a custom job filter and see if you have any dataset copy jobs that failed. With a single drive tape library, dataset copy jobs for other tape goals should fail. You may need to add a 2nd tape drive to get some LT tape backups to succeed.

         

    2) With co-location enabled, the tape label will be assigned based on the protection group that first wrote to the new tape.  So even if you have different tape labels for the same goal in each PG, the tape label will be formed when the tape is overwritten and the label will be based on the PG whose data source was first written to tape..

    3) I don't follow you - each tape recovery goal can have different retention ranges.

    5) I don't know why you would want two weekly tape backup jobs, but a long term goal will not cancel a short term goal.  Instead we would make the short term backup, then schedule a dataset copy job for the LT goal 15 minutes later.


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread. Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.

    Wednesday, May 8, 2013 11:52 PM
    Moderator
  • Hi Mark:

    Sorry for the delay in answering. I had some health problems. Answering to you:

    i) I modified the script that as you said, and it worked. The script reported that everything was fine with the drive.

    ii) After reading what you say about the filter I found that all the error messages I was talking about were related to dataset copy jobs, and all the jobs that are working fine are all the tape copy jobs. I only want to do tape copy jobs. I don't know why are dataset copy jobs programmed, but I want to delete them. How can I delete the copy dataset jobs? I looked for them but I could not find them.

    iii) If you select tape for short term protection you can not specify diferent retention periods for the incremental and the full backup jobs.

    v) I don't want to do two full weekly backups, but this is the only option I found. What I want exactly to do is:

    . Buckup of three servers with one only PG (to make the jobs all together in time and media) and one only drive.

    . PG daily incremental backup jobs to tape with a retention of one week. Four tapes asigned to this jobs.

    . PG weekly full off-site backup jobs to tape with a retention of four weeks. Four tapes asigned to this jobs.

    . PG monthly full off-site backup jobs to tape with a retention of two months. Two tapes asigned to this jobs.

    Is this possible using DPM? How must I configure de PG to get it done?

    Thanks again.

    Wednesday, July 3, 2013 3:03 PM
  • Hi,

    Qii) After reading what you say about the filter I found that all the error messages I was talking about were related to dataset copy jobs, and all the jobs that are working fine are all the tape copy jobs. I only want to do tape copy jobs. I don't know why are dataset copy jobs programmed, but I want to delete them. How can I delete the copy dataset jobs? I looked for them but I could not find them.

    Aii) The dataset copy jobs are for making the long term backups from other already written long term backups.  See answer Av) below for solution.

    Qiii) If you select tape for short term protection you can not specify diferent retention periods for the incremental and the full backup jobs.

    Aiii)  The retention period is for the single short tape backup term goal, and there must be a FULL backup weekly.  The incremental backups depend on the FULL, so that is why they cannot be separated.

    Qv) I don't want to do two full weekly backups, but this is the only option I found. What I want exactly to do is:

    . Buckup of three servers with one only PG (to make the jobs all together in time and media) and one only drive.

    . PG daily incremental backup jobs to tape with a retention of one week. Four tapes asigned to this jobs.

    . PG weekly full off-site backup jobs to tape with a retention of four weeks. Four tapes asigned to this jobs.

    . PG monthly full off-site backup jobs to tape with a retention of two months. Two tapes asigned to this jobs.

    Av) This is associated to answer Aiii - you cannot get past two weekly full backups since short term requires one full backup per week.  So just change the retention of the daily full+incremental from one week to four.  Eliminate the weekly, and maintain the monthly.  That will give you the exact same protection and eliminate the dataset copy jobs.


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread. Regards, Mike J. [MSFT] This posting is provided "AS IS" with no warranties, and confers no rights.

    Wednesday, July 3, 2013 9:50 PM
    Moderator