Rebuilding AMPs from Fallback - Teradata Database on VMware

Teradata Database on VMware Developer Tier Getting Started Guide

Product
Teradata Database on VMware
Release Number
Deployer 3.01
Published
September 2017
Language
English (United States)
Last Update
2018-04-25
dita:mapPath
cmx1499274497615.ditamap
dita:ditavalPath
TDVMWare_phase4a-4b_DEV.ditaval.ditaval
dita:id
B035-5938
lifecycle
previous
Product Category
Cloud
Before you begin, assess if you can recover from fallback. If more than a single AMP in the same fallback cluster is down, you cannot use this procedure.
  1. SSH into one of the nodes.
  2. Run vprocmanager.
  3. Run the command status.
  4. Look for AMPs marked as FATAL** and check the column Cluster/Host No. If two AMPs with the same Cluster/Host No. are Fatal, then recovery using fallback is impossible. In the following example, both AMPs in cluster 0, 3, 6, 9 are Fatal and are not recoverable.
    SYSTEM NAME: MPP160                                         17/01/12 14:40:21
                              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** Online    AMP       0    On    28671
        1     2     1-01  Yes     0   FATAL** Online    AMP       1    On    28671
        2     3     1-01  Yes     0   FATAL** Online    AMP       2    On    28671
        3     4     1-01  Yes     0   FATAL** Online    AMP       3    On    28671
        4*    5     1-01  Yes     0   ONLINE  Online    AMP       4    On    28670
        5     6     1-01  Yes     0   ONLINE  Online    AMP       5    On    28670
        6     7     1-01  Yes     0   ONLINE  Online    AMP       6    On    28670
        7     8     1-01  Yes     0   ONLINE  Online    AMP       7    On    28670
        8     9     1-01  Yes     0   ONLINE  Online    AMP       8    On    28669
        9    10     1-01  Yes     0   ONLINE  Online    AMP       9    On    28669
       10    11     1-01  Yes     0   ONLINE  Online    AMP      10    On    28669
       11    12     1-01  Yes     0   ONLINE  Online    AMP      11    On    28669
       12     1     1-02  Yes     0   FATAL** Online    AMP       0    On    28668
       13     2     1-02  Yes     0   FATAL** Online    AMP       3    On    28668
       14     3     1-02  Yes     0   FATAL** Online    AMP       6    On    28668
       15     4     1-02  Yes     0   FATAL** Online    AMP       9    On    28668
  5. If recovery is possible in your Vsphere client, delete the failed Pdisks that corresponds to the fatal AMPs.
  6. Run tdc.deploy.exe with the storage step using the deployed system's Node.IT.Properties and Common.IT.Properties This re-creates the removed Pdisks.
    A copy of these two configs are located in /root/tvme on any of the nodes.
  7. When the command finishes, run tdc-rebuild as the root user.
  8. Enter your DBC account and password, then follow the prompts until the system is rebuilt.