Data loss does occur if shutdown / init 0 is performed on Teradata Database instance with Local Storage

Teradata Database on AWS
Teradata Employee

Data loss does occur if shutdown / init 0 is performed on Teradata Database instance with Local Storage

Teradata Database on Amazon Web Services (AWS) provides 2 types of storage options.

  • Amazon EC2 Instance Local Storage: Solid State Drive (SSD) or High Density Drive (HDD) local media.
  • Amazon Elastic Block Store (Amazon EBS): Remote storage on SSD volumes that are attached to the instance.

Local Storage:

Teradata Database instances with local storage option provide better performance when compared to EBS storage option, but the storage is temporary or ephemeral. Local storage uses SSD or HDD local media. Local storage is ideal for test and development, where you do not need to keep data.

EBS Storage:

With EBS storage, the root volume for a Teradata Database on AWS instance and additional EBS volumes that you attach to the instance persist by default. Data in EBS storage remains until "Delete on Termination" flag is set on attached volumnes and Teradata Database instance is terminated.

On Teradata Database instance with Local storage, data loss does occur for following events.

  • Shutdown (init 0 or shutdown -h / -P or halt or poweroff)
  • Node hang
  • Machine check (hardware fault such as a double bit memory error)
  • Stopping an instance through Amazon EC2 Console
  • Terminating an instance through Amazon EC2 Console

Data loss does not occur with events, such as reboot (init 6 or shutdown -r) or node panic (oops event).


Accepted Solutions
Teradata Employee

Re: Data loss does occur if shutdown / init 0 is performed on Teradata Database instance with Local Storage

Here are examples for both EBS and local storage.

1. 'stop' DBS instance with 'init 0' and manually 'start' – Amazon EBS storage 

   1) Shutdown database with 'tpareset -x'

   2) 'init 0' from linux prompt

   3) The instance state automatically changed to 'stop'.

   4) Start instance : 'Instance State' > 'Start'

   5) Verified instance state running with '2/2 checks passed' state.

   6) Logon database and verified the DBS objects that created before stop instance still available.

2. 'reboot DBS instance' – Amazon EBS storage

   1) Shutdown database with 'tpareset -x'

   2) Reboot instance : 'Instance State' > 'reboot'

      - Message displayed on SecureCRT : "The system is going down for reboot NOW!"   

   3) Verified instance state running with '2/2 checks passed' state.

   6) Logon database and verified the DBS objects that created before stop instance still available.

3. 'stop/start' DBS instance – Amazon local storage

   1) Shutdown database with 'tpareset -x'

   2) 'init 0' from linux prompt

   3) The instance state automatically changed to 'stop'.

   4) Start instance : 'Instance State' > 'Start'

   5) Verified instance state running with '2/2 checks passed' state.

   6) Database startup failure due to assigned pdisks are gone. verify_pdisks failure.

      ip-172-31-18-120:~ # verify_pdisks

      Error:  Cannot open /dev/pdisk/dsk1.

      Error:  Cannot open /dev/pdisk/dsk3.

      Error:  Cannot open /dev/pdisk/dsk0.

      Error:  Cannot open /dev/pdisk/dsk2.

      *****4 error(s) found during pdisk verification*****

1 ACCEPTED SOLUTION
1 REPLY
Teradata Employee

Re: Data loss does occur if shutdown / init 0 is performed on Teradata Database instance with Local Storage

Here are examples for both EBS and local storage.

1. 'stop' DBS instance with 'init 0' and manually 'start' – Amazon EBS storage 

   1) Shutdown database with 'tpareset -x'

   2) 'init 0' from linux prompt

   3) The instance state automatically changed to 'stop'.

   4) Start instance : 'Instance State' > 'Start'

   5) Verified instance state running with '2/2 checks passed' state.

   6) Logon database and verified the DBS objects that created before stop instance still available.

2. 'reboot DBS instance' – Amazon EBS storage

   1) Shutdown database with 'tpareset -x'

   2) Reboot instance : 'Instance State' > 'reboot'

      - Message displayed on SecureCRT : "The system is going down for reboot NOW!"   

   3) Verified instance state running with '2/2 checks passed' state.

   6) Logon database and verified the DBS objects that created before stop instance still available.

3. 'stop/start' DBS instance – Amazon local storage

   1) Shutdown database with 'tpareset -x'

   2) 'init 0' from linux prompt

   3) The instance state automatically changed to 'stop'.

   4) Start instance : 'Instance State' > 'Start'

   5) Verified instance state running with '2/2 checks passed' state.

   6) Database startup failure due to assigned pdisks are gone. verify_pdisks failure.

      ip-172-31-18-120:~ # verify_pdisks

      Error:  Cannot open /dev/pdisk/dsk1.

      Error:  Cannot open /dev/pdisk/dsk3.

      Error:  Cannot open /dev/pdisk/dsk0.

      Error:  Cannot open /dev/pdisk/dsk2.

      *****4 error(s) found during pdisk verification*****