Oracle RMAN redirected restore fails with error ORA-19501: read error on file "<backup piece name>", block number 1 (block size=512)


   Article Number:     501057                                   Article Version: 2     Article Type:    Break Fix 




NetWorker Module for Databases and Applications,NetWorker Module for Databases and Applications 8.2





Redirected restore failing when volume is pre-mounted on first Storage node in the Client definition storage node affinity list before starting restore.   
    If you pre-mounted the volume on the NetWorker server which was first in the storage node affinity list on the target client you received the following error: -   
    RMAN> run {   
    allocate channel ch1 type 'SBT_TAPE';   
    send 'NSR_ENV=(NSR_CLIENT=<client name>,NSR_GROUP=<group name>,NSR_SERVER=<NetWorker server name>)';   
    restore controlfile from '<backup piece name>';   
    release channel ch1;   
    }2> 3> 4> 5> 6>   
    using target database control file instead of recovery catalog   
    allocated channel: ch1   
    channel ch1: SID=403 device type=SBT_TAPE   
    channel ch1: NMDA Oracle v8.2.1   
    sent command to channel: ch1   
    Starting restore at 09-JUN-17   
    channel ch1: restoring control file   
    released channel: ch1   
    RMAN-00571: ===========================================================   
    RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============   
    RMAN-00571: ===========================================================   
    RMAN-03002: failure of restore command at 06/09/2017 15:16:57   
    ORA-27192: skgfcls: sbtclose2 returned error - failed to close file   
    ORA-19511: Error received from media manager layer, error text:   
       We could not read the checksum. (0:3:2)   
    ORA-19870: error while restoring backup piece <backup piece name>   
    ORA-19501: read error on file "<backup piece name>", block number 1 (block size=512)   
    ORA-27190: skgfrd: sbtread2 returned error   
    ORA-19511: Error received from media manager layer, error text:   






Storage node used was Windows, original backup done on Unix Storage node.  Device blocksize used can be different depending on hardware/OS software/NetWorker device configuration.   
    The Client definition for the source client had NetWorker Storage node affinity list with Storage node followed by nsrserverhost entry.   
    The Client definition for the source client had nsrserverhost first in the list and then the other Storage Node (reverse order).   
    NetWorker will decide which storage node to use considering a variety of factors (group, pool, recover storage node, storage node relationships/values) as well as explicitly naming in RMAN script/configuration.   






Different storage node used for restore to original backup.                                                           






We advise Customers to use storage node with same OS (preferably the original Storage node) for this type of restore where they have environments with a mixture of Storage node OS's to avoid potential block size issues when reading volumes, but this is not mandatory.  Within the device definitions we do not recommend specifying specific block sizes but to standardize the settings across all devices (usually "handler default") unless specific environmental considerations apply but this is no guarantee that volumes will be able to be read on different storage node from the original one used for backup.   
    Hardware and software can affect what blocksizes are used when writing/reading backups.   
    From Daemon.log files on NetWorker server we traced which Storage node was used to confirm Storage Node affinity list setting and this Storage node was used going forward.