Recovering a Cylinder Index Segregation LUN - Parallel Upgrade Tool (PUT)

Parallel Upgrade Tool (PUT) Reference

Product
Parallel Upgrade Tool (PUT)
Release Number
3.09
Published
February 2020
Language
English (United States)
Last Update
2020-02-24
dita:mapPath
ows1493317469465.ditamap
dita:ditavalPath
ows1493317469465.ditaval
dita:id
B035-5716
Product Category
Software
Teradata Tools and Utilities
Complete the following steps for CIG LUN Recovery after completing Step 8 - Create Partitions on New LUNs. In the following procedure, the device experiencing the I/O failure (as indicated by the PFD_E_IO designation) is /dev/pfdi.
This procedure is destructive in nature, so only users with a thorough knowledge of Pdisk Filter Driver internals should perform this procedure.
  1. Check for I/O errors by using the get_map option of the Pdisk Filter Driver utility.
    # pfdutil -get_map 
    
    0 pfda sdj 0x800090 PFD_OK 23232 128 15488 23104 1874048 1858560
    1 pfdb sdk 0x8000a0 PFD_OK 23232 128 15488 23104 1874048 1858560
    ...
    8 pfdi sdb 0x800010 PFD_E_IO 23232 128 15488 23104 1874048 1858560
  2. Use get_tvs_map option of the Pdisk Filter Driver utility to identify the AMP and pdisk associated with the pfd device experiencing the I/O failure. The following example identifies AMP 0 and pdisk dsk0 as these entities.
    # pfdutil -get_tvs_map dev=/dev/pfdi 
    
    AMP:0, /dev/pdisk/dsk0, /dev/pfdi
  3. If the AMP associated with the pfd device experiencing the I/O failure is not marked as FATAL and the database is still running, stop the database before proceeding.
  4. Use the del_phys option of the Pdisk Filter Driver utility to delete the pfd device that is experiencing the I/O failure. For example:
    # pfdutil -del_phys dev=/dev/pfdi 
    
    device /dev/pfdi deleted from device map
    
    the /etc/opt/teradata/pfd/pfd.conf file with the replacement LUNs new
    /dev/disk/by-id/ number and use the pfdutil -add_phys dev=/dev/pfdi operation
    to bring it back online.
  5. Use the get_map option of the Pdisk Filter Driver utility to confirm that the pfd device experiencing the I/O failure is offline. For example:
    # pfdutil -get_map 
    
    0 pfda sdj 0x800090 PFD_OK 23232 128 15488 23104 1874048 1858560
    1 pfdb sdk 0x8000a0 PFD_OK 23232 128 15488 23104 1874048 1858560
    ...
    8 pfdi sdb 0x800010 PFD_E_OFFLINE 23232 128 15488 23104 1874048 1858560
  6. In /etc/opt/teradata/pfd/pfd.conf, replace the name of the device experiencing the I/O failure with the by-id entry of the new LUN, as recorded as part of Step 8 - Create Partitions on New LUNs. For example:
    version 2
    clique 0
    /dev/disk/by-id/scsi-3600a0b800050363000000a71511e6e9e, /dev/pfda, 23232, 128, 15488, 1858560, 1874048, 277862464
    /dev/disk/by-id/scsi-3600a0b800050363000000a75511e6ea8, /dev/pfdb, 23232, 128, 15488, 1858560, 1874048, 277862464
    ...
    /dev/disk/by-id/scsi-3600a0b8000503630000009324ea1a414,  new LUN by-id value, 23232, 128, 15488, 1858560, 1874048, 277862464
  7. Use the add_phys option of the Pdisk Filter Driver utility to add the new LUN to the pfd device map. For example:
    # pfdutil -add_phys dev=/dev/pfdi 
    
    device /dev/pfdi added to device map
  8. Run tvsaDeviceInit to initialize the pdisk associated with the pfd device (as determined at step 2). For example:
    # tvsaDeviceInit /dev/pdisk/dsk0
  9. Run # tvsaprofiler:
    # tvsaprofiler -put
  10. Run verify_pdisks to verify that pdisk errors do not result.
  11. Restart the database.
  12. Issue a tpareset command to clear the I/O error condition for the fatal AMP associated with the pfd device. For example:
    # tpareset -yes “clear IO error condition”
  13. In vprocmanager, set the failed AMP online and issue a restart coldwait command. For example:
    # vprocmanager 
      
     _______
        |                                |              |
        |    ___     __     ____         |    ____    __|__    ____
        |   /      |/  \    ____|    ____|    ____|     |      ____|
        |   ---    |       /    |   /    |   /    |     |     /    |
        |   \___   |       \____|   \____|   \____|     |__   \____|
      
        Release 15.10m.00.97 Version 15.10m.00.97
        VprocManager Utility (Sep 98)
      
    Enter a command, HELP or QUIT:
    st
    SYSTEM NAME: localhost                                      14/11/20 13:23:56
                              DBS LOGICAL CONFIGURATION
                              -------------------------
                                                                       Rcv        
                                                                       Jrnl/      
    Vproc  Rel.   Node    Can   Crash Vproc   Config   Config Cluster/ Host  TVS  
    Number Vproc# ID      Move  Count State   Status   Type   Host No. Type  Vproc
    ------ ------ ------  ----- ----- ------- -------- ------ -------- ----- -----
        0     1     1-01  Yes     0   FATAL   Down      AMP       0    On    28670
        1*    2     1-01  Yes     0   ONLINE  Online    AMP       0    On    28670
        2     3     1-01  Yes     0   ONLINE  Online    AMP       1    On    28670
        3     4     1-01  Yes     0   ONLINE  Online    AMP       1    On    28670
        4     5     1-01  Yes     0   ONLINE  Online    AMP       2    On    28671
        5     6     1-01  Yes     0   ONLINE  Online    AMP       2    On    28671
        6     7     1-01  Yes     0   ONLINE  Online    AMP       3    On    28671
        7     8     1-01  Yes     0   ONLINE  Online    AMP       3    On    28671
    26623    12     1-01  No      0   ONLINE  N/A       RSG       0    N/A   N/A 
    22528    11     1-01  No      0   ONLINE  N/A       GTW       1    COP   N/A 
    28670    13     1-01  Yes     0   ONLINE  N/A       TVS       0    N/A   N/A 
    28671    14     1-01  Yes     0   ONLINE  N/A       TVS       0    N/A   N/A 
    30718     9     1-01  Yes     0   ONLINE  Online    PE        1    COP   N/A 
    30719    10     1-01  Yes     0   ONLINE  Online    PE        1    COP   N/A 
    --------------------------------------------------------------------------------
    *   DBS Control AMP
      DBS State: Logons are enabled - The system is quiescent
      DBS RestartKind: COLD
                              PDE PHYSICAL CONFIGURATION
                              --------------------------
    Node    Node    Clique      Memory
    ID      State   Number CPUs  (MB)  CHANs LANs AMPs Node Name
    ------- ------- ------ ---- ------ ----- ---- ---- ---------------------------
      1-01  ONLINE     0     2   31971    0     1    7  localhost                       
    ------------------------------------------------------------------------------
                            
      PDE State: RUN/STARTED
    Enter a command, HELP or QUIT:
    set 0 online
    Vproc 0 will begin recovery in the background via the Recovery Control Task
    Enter a command, HELP or QUIT:
    restart nodump coldwait
     
    Are you sure you want to Restart the DBS now (Y/N)?
    y
     
    The DBS will be restarted.
        Dump        : NO
        RestartKind : COLDWAIT
        Reason      : System restarted by VprocManager. 
    Exiting VprocManager...
  14. Verify the PDE state is RUN/STARTED and the DBS state is quiescent:
    sdll8988:~ # pdestate -a 
    
    PDE state is RUN/STARTED.
    BS state is 5: Logons are enabled - The system is quiescent
  15. Verify all DBS vprocs are fully online.
    # vprocmanager  
      
     _______
        |                                |              |
        |    ___     __     ____         |    ____    __|__    ____
        |   /      |/  \    ____|    ____|    ____|     |      ____|
        |   ---    |       /    |   /    |   /    |     |     /    |
        |   \___   |       \____|   \____|   \____|     |__   \____|
      
        Release 15.10m.00.97 Version 15.10m.00.97
        VprocManager Utility (Sep 98)
      
    Enter a command, HELP or QUIT:
    st not
    SYSTEM NAME: localhost                                      14/11/20 13:26:32
    All DBS vprocs are fully online.
    All PDE nodes are fully online.
    Enter a command, HELP or QUIT:
    quit
    Exiting VprocManager...
  16. Verify the status of all the device status is PFD_OK. For example:
    # pfdutil -get_map 
    
    0 pfda sdj 0x800090 PFD_OK 23232 128 15488 23104 1874048 1858560
    1 pfdb sdk 0x8000a0 PFD_OK 23232 128 15488 23104 1874048 1858560
    ...
    8 pfdi sdb 0x800010 PFD_OK 23232 128 15488 23104 1874048 1858560