0% found this document useful (0 votes)
98 views31 pages

Backup and Recovery of Sap Systems On Aws For Linux Maxdb and db2 v1.6

Uploaded by

sabir
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
98 views31 pages

Backup and Recovery of Sap Systems On Aws For Linux Maxdb and db2 v1.6

Uploaded by

sabir
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 31

 

Backup  and  Restore  of  SAP  Systems  on  


Amazon  Web  Services  Infrastructure  
For  MaxDB  and  DB2  LUW  Databases  on  Linux  

 
 
 
 
Authors:   Amazon  Web  Services  
sap-­‐on-­‐[email protected]    
 
Protera  Technologies  
http://www.protera.biz    
 
Version:     1.6  –  March  2012  
   
Table  of  Contents  
Prerequisite  Documents  ...........................................................................................................................  4  
SAP  on  Amazon  Web  Services  ..............................................................................................................  4  
SAP  on  MaxDB  ......................................................................................................................................  4  
SAP  on  DB2  UDB  ...................................................................................................................................  4  
Scope  of  this  Document  ...........................................................................................................................  5  
Components  for  SAP  Backup  and  Restore  on  AWS  infrastructure  ...........................................................  5  
Amazon  Elastic  Compute  Cloud  (EC2)  ..............................................................................................  5  
Amazon  Simple  Storage  Service  (Amazon  S3)  ..................................................................................  5  
Amazon  Elastic  Block  Storage  (EBS)  .................................................................................................  6  
Amazon  Virtual  Private  Cloud  (VPC)  .................................................................................................  6  
Storage  layout  of  SAP  systems  on  EBS  volumes  .......................................................................................  6  
Backup  and  Restore  procedures  using  AWS  infrastructure  .....................................................................  8  
SAP  on  MaxDB  backups  using  AWS  Infrastructure  ...............................................................................  8  
SAP  on  DB2  backups  using  AWS  Infrastructure  ..................................................................................  10  
Restore  ...............................................................................................................................................  11  
Common  backup  and  restore  operations  on  Amazon  EC2  instances  and  EBS  volumes  ........................  11  
Backup:  creating  a  new  EBS  volume  with  an  empty  file  system  ........................................................  11  
Backup:  creating  an  EBS  snapshot  onto  Amazon  S3  of  an  EBS  volume   .............................................  12  
Backup:  dismounting  file  system(s)  and  detaching  an  EBS  Volume  ...................................................  12  
Backup:  creating  a  full  offline  Amazon  EC2  Amazon  Machine  Image  (AMI)  ......................................  12  
Detailed  steps  to  create  the  Amazon  Machine  Image  (AMI)  .........................................................  13  
Examples  for  backing  up  SAP  System  components  using  AWS  infrastructure  .......................................  14  
Example  1:  database  backup  to  an  EBS  backup  file  system  ...............................................................  14  
Example  1a:  full  online  data  and  log  backup  for  MaxDB  ................................................................  15  
Create  MaxDB  backup  templates  ...............................................................................................  15  
Back  up  the  database  using  Database  Studio  .............................................................................  16  
Back  up  the  database  transaction  logs  using  Database  Studio  ..................................................  16  
Back  up  the  database  using  DBMCLI  ..........................................................................................  16  
Back  up  the  database  log  using  DBMCLI  ....................................................................................  16  

Page  2  of  31  


 
Schedule  hourly  automatic  log  backup  using  Database  Studio  ..................................................  16  
Create  a  snapshot  to  send  the  backup  to  Amazon  S3  ................................................................  17  
Example  1b:  full  online  backup  for  DB2  LUW  .................................................................................  17  
Steps  to  enable  rollforward  recovery  and  set  up  DB2  logfile  management:  .............................  17  
Option  1:  back  up  the  database  directly  from  the  SAP  system  ..................................................  18  
Option  2:  back  up  the  database  using  the  CLI  ............................................................................  18  
DB2  transaction  log  file  management  ........................................................................................  19  
Create  a  snapshot  to  send  the  backup  to  Amazon  S3  ................................................................  19  
Example  2:  OS  backup  to  S3  using  the  TAR  command  .......................................................................  20  
Examples  for  restoring  SAP  systems  using  AWS  infrastructure  .............................................................  22  
Example  1:  restore  OS  from  Amazon  S3  using  TAR  ............................................................................  22  
Create  an  Amazon  instance  using  the  “golden  backup”  AMI  of  the  original  system  .....................  22  
Recreate  the  file  systems  required  for  the  recovery  ......................................................................  22  
Restore  the  OS  ...............................................................................................................................  23  
Reboot  the  instance  .......................................................................................................................  24  
Example  2:  Restoring  and  Recovering  the  Database  from  Amazon  S3  ..............................................  24  
Example  2a:  MaxDB  restore  and  recovery  from  Amazon  S3  ..........................................................  25  
Example  2b:  DB2  LUW  restore  and  recovery  from  Amazon  S3  ......................................................  29  

   
   

Page  3  of  31  


 
Prerequisite  Documents  
The  following  information  should  be  read  carefully  before  continuing  with  this  guide.  Especially,  this  
document  cannot  serve  as  a  replacement  for  the  MaxDB  and  DB2  Backup  and  Restore  information  
resources  that  are  referenced  below.      

SAP  on  Amazon  Web  Services  


• SAP  on  AWS  Implementation  Guide  http://aws.amazon.com/sap  
Information  on  deploying  SAP  systems  on  AWS  Infrastructure  
• SAP  note  1588667  –  SAP  on  Amazon  Web  Services  (AWS)  
Entry  SAP  note  for  Amazon  Web  Services  
• SAP  note  1677381  –  Backup  and  Restore  guidelines  
Errata  and  feedback  reference  for  this  guide  

SAP  on  MaxDB  


• MaxDB  Backup  and  Restore  
SAP  SDN  page  on  MaxDB  Backup  and  Restore  
• SAP  Library  –  Database  Studio  –  Backing  up  Databases:  Overview  
SAP  Library  documentation  on  MaxDB  Backup  and  Restore  
• SAP  note  1377148  -­‐-­‐  FAQ:  SAP  MaxDB  backup/recovery  
Answers  to  Frequently  Asked  Questions  on  SAP  MaxDB  backup  and  recovery  
• SAP  note  767598  -­‐-­‐  Available  SAP  MaxDB  documentation  
General  overview  of  SAP  MaxDB  documentation  

SAP  on  DB2  UDB  


• Database  Administration  Guide  "SAP  on  IBM  DB2  for  Linux,  Unix  and  Windows"  
This  document  provides  specific  information  about  the  administration  of  IBM  DB2  for  Linux,  
UNIX,  and  Windows  (in  the  following  referred  to  as  DB2)  in  an  SAP  environment.  In  addition,  it  
provides  references  to  additional  documentation  and  guidelines  as  well  as  recommendations  
from  SAP  that  are  only  available  in  this  document.  It  also  helps  you  to  plan,  install,  and  maintain  
SAP  systems  and  the  database.  
• IBM  DB2  Universal  Database  for  UNIX  and  Windows  -­‐  New  Log  File  Management  
This  documentation  describes  the  concepts  of  the  new  DB2  log  file  management  and  how  you  
can  migrate  from  the  existing  SAP  DB2  log  file  management  solution  to  the  DB2  V8.2  log  file  
management  solution.  
• A  Practical  Guide  to  Backup  and  Recovery  of  IBM  DB2  for  Linux,  UNIX  and  Windows  in  SAP  
Environments  
Introduction  to  the  basics  of  backup  and  recovery  for  DB2  databases.  This  guide  explains  the  
architecture  for  backup  and  recovery,  the  most  relevant  commands  and  discusses  the  DB2  log  
file  management.  In  addition,  the  integration  of  the  DB2  backup  and  recovery  architecture  into  
SAP  NetWeaver  is  described.  

Page  4  of  31  


 
 

Scope  of  this  Document    


This  document  will  cover  backing  up  and  restoring  of  a  SAP  ECC  6.0  system  on  MaxDB  or  DB2  on  SLES  in  
the  Amazon  AWS  environment.  Steps  for  performing  both  OS  and  Database  backups  for  MaxDB  and  DB2  
are  provided.  

Amazon  Elastic  Block  Store  (EBS)  snapshots  are  point  in  time  images  of  volumes  which  are  persisted  to  
Amazon  S3.  These  snapshots  can  be  used  as  the  starting  point  for  new  Amazon  EBS  volumes,  and  
protect  data  for  long-­‐term  durability.  The  same  snapshot  can  be  used  to  instantiate  as  many  volumes  as  
desired.    

Database  and  OS  backups  can  be  accomplished  via  provided  OS  and  DBMS  backup  tools  and  in  addition  
utilizing  Amazon  EBS  snapshots  to  secure  these  backups.  Complete  offline  system  backups  to  a  so  called  
Amazon  machine  Image  (AMI)  will  also  be  described.  

Components  for  SAP  Backup  and  Restore  on  AWS  infrastructure  


Apart  from  an  SAP  system  on  a  platform  that  is  supported  by  SAP  on  Amazon  Web  Services  (AWS)  
infrastructure,  the  following  AWS  products  are  required  to  perform  the  backup  and  restore  operations  
that  are  described  in  this  document.  

• Amazon  Elastic  Compute  Cloud  (Amazon  EC2)  


• Amazon  Elastic  Block  Storage  (EBS)  
• Amazon  Simple  Storage  Service  (Amazon  S3)  
• Indirectly:  Amazon  Virtual  Private  Cloud  (Amazon  VPC)  

Amazon  Elastic  Compute  Cloud  (EC2)  


Amazon  Elastic  Compute  Cloud  (Amazon  EC2)  is  a  web  service  that  provides  resizable  compute  capacity  
in  the  cloud.  Amazon  EC2  presents  a  virtual  computing  environment,  allowing  one  to  use  web  service  
interfaces  to  launch  instances  with  a  variety  of  pre-­‐imaged  operating  systems,  load  them  with  an  own  
custom  application  environment  and  manage  network’s  access  permissions.  A  customized  image  can  be  
persisted  as  own  Amazon  Machine  Image  (AMI),  and  can  be  redeployed  to  as  many  or  few  instances  as  
desired.  

Amazon  EC2  provides  different  instance  types  to  meet  different  computing  needs.  The  specific  Amazon  
EC2  instance  types  that  are  currently  supported  for  SAP  application  deployments  are  listed  in  SAP  note  
1588667.  

Further  information  on  Amazon  EC2  can  be  found  at  http://aws.amazon.com/ec2/.  

Amazon  Simple  Storage  Service  (Amazon  S3)  


Amazon  S3  is  storage  for  the  Internet.  Amazon  S3  provides  a  simple  web  services  interface  that  can  be  
used  to  store  and  retrieve  any  amount  of  data,  at  any  time,  from  anywhere  on  the  web.  The  service  is  

Page  5  of  31  


 
also  designed  to  provide  99.999999999%  durability  and  99.99%  availability  over  a  given  year.  These  and  
other  Amazon  S3  properties  make  it  the  ideal  storage  for  enterprise  backups.  

Further  information  on  Amazon  S3  can  be  found  at  http://aws.amazon.com/s3/.  

Amazon  Elastic  Block  Storage  (EBS)  


Amazon  Elastic  Block  Store  (EBS)  provides  block  level  storage  volumes  for  use  with  Amazon  EC2  
instances.  EBS  volumes  are  off-­‐instance  storage  that  persists  independently  from  the  life  of  an  EC2  
instance.  EBS  provides  highly  available,  highly  reliable  storage  volumes  that  can  be  attached  to  a  running  
Amazon  EC2  instance  and  exposed  as  a  device  within  the  instance.  

EBS  also  provides  the  ability  to  create  point-­‐in-­‐time  snapshots  of  volumes,  which  are  persisted  to  
Amazon  S3.  These  snapshots  can  be  used  as  the  starting  point  for  new  EBS  volumes,  and  protect  data  
for  long-­‐term  durability.  The  same  snapshot  can  be  used  to  instantiate  as  many  volumes  as  required.  

Further  information  on  EBS  can  be  found  at  http://aws.amazon.com/ebs/.  

Amazon  Virtual  Private  Cloud  (VPC)  


Amazon  Virtual  Private  Cloud  (Amazon  VPC)  lets  one  provision  a  private,  isolated  section  of  the  Amazon  
Web  Services  (AWS)  Cloud,  where  AWS  resources  like  Amazon  EC2  instances  and  EBS  volumes  can  be  
launched  in  a  self-­‐defined  virtual  network.  With  Amazon  VPC,  a  virtual  network  can  be  defined  that  
closely  resembles,  and  can  be  connected  securely  to,  the  own  traditional  on-­‐premise  network.  

Amazon  VPC  is  not  a  direct  prerequisite  for  the  backup-­‐  and  restore-­‐operations  that  are  described  in  this  
guide.  However,  SAP  systems  themselves  are  only  supported  on  Amazon  Web  Services  infrastructure  
when  deployed  within  an  Amazon  VPC.  

Further  information  on  Amazon  VPC  can  be  found  at  http://aws.amazon.com/vpc/.  

Storage  layout  of  SAP  systems  on  EBS  volumes  


Foremost,  please  refer  to  the  White  Paper  for  SAP  Systems  on  Amazon  AWS  as  a  baseline  guide  on  
building  SAP  systems  on  AWS.  The  most  recent  version  of  this  document  can  accessed  at  
http://media.amazonwebservices.com/Operating%20SAP%20Solutions%20on%20AWS%20White%20Pa
per.pdf.  

It  is  recommended  to  separate  OS,  SAP,  DBMS,  DB  Data  and  DB  transaction  log  components  onto  
different  EBS  volumes.  In  addition,  separate  file  systems  will  be  used  to  store  the  following  types  of  
backups  on  MaxDB:  

1. OS,  SAP  and  DBMS  system  backups  (binaries,  profiles,  etc.)  


2. Full  and/or  incremental  database  data  backups  
3. Database  transaction  log  backups  

This  translates  to  the  recommended  file  system  layout  for  MaxDB  on  linux  as  shown  in  Table  1.  

Page  6  of  31  


 
Table  1:  A  recommended  file  system  layout  for  an  SAP  system  on  MaxDB  on  linux  

  Linux  File  System   Description   Tag  

/   OS  root  directory   OS-­‐EXE  

/sapmnt   SAP  system,  shared OS-­‐EXE  

/sapdb   MaxDB  system OS-­‐EXE  

/usr/sap   SAP  system,  local OS-­‐EXE  

/sapdb/<SID>/sapdata   Database  Data  files DB-­‐DATA  

/sapdb/<SID>/saplog   Database  Transaction  Logs   DB-­‐LOG  

/os_exe_backups   OS,  SAP  and  MaxDB  system  backups   OS-­‐EXE-­‐BACKUPS  


/db_data_backups   MaxDB  database  backups   DB-­‐DATA-­‐BACKUPS  
/db_log_backups   MaxDB  transaction  log  backups   DB-­‐LOG-­‐BACKUPS  
 

Note  that  file  systems  printed  in  bold  belong  to  the  usual  SAP  system,  and  the  other  file  systems  
are  additional  ones  to  store  backups.  The  tags  will  be  used  in  the  following  sections  to  
distinguish  the  file  system  groups.  

As  DB2  can  be  configured  to  have  a  separate  file  system  to  archive  its  transaction  logs  
automatically,  a  file  system  for  transaction  log  backups  like  we  have  on  MaxDB  is  not  required  
for  DB2.  Table  2  shows  how  a  recommended  file  system  layout  could  look  like  for  DB2  on  Linux.  

Table  2:  A  recommended  file  system  layout  for  an  SAP  system  on  DB2  on  Linux  

Linux  File  System   Description   Tags  

/   OS  root  directory   OS-­‐EXE  

/sapmnt   SAP  system,  shared OS-­‐EXE  

/db2   DB2  LUW  binaries,  configuration  and   DB2-­‐EXE  


trace  files    

/usr/sap   SAP  system,  local OS-­‐EXE  

/db2/<SID>/sapdata1   Database  data  files DB2-­‐DATA  

:  

/db2/<SID>/sapdata<n>  

Page  7  of  31  


 
/db2/<SID>/log_dir   Active  database  transaction    logs   DB2-­‐LOG  

/db2/<SID>/log_archive   Archived  database  transaction  logs   DB2-­‐LOG-­‐ARCHIVE  


/os_exe_backups   OS,  SAP  and  DB2  system  backups   OS-­‐EXE-­‐BACKUPS  
/db2_data_backups   DB2  Database  Data  Backups   DB2-­‐DATA-­‐BACKUPS  
 

Note  that  the  file  system  for  MaxDB  transaction  log  backups  /db_log_backups  in  Table  1  
been  replaced  by  /db2/<SID>/log_archive  for  DB2  (Table  2).  The  content  of  this  DB2  file  
systems  is  managed  by  the  DB2  RDBMS  itself.  

Backup  and  Restore  procedures  using  AWS  infrastructure  


Apart  from  creating  a  full  offline  Amazon  EC2  machine  image,  the  following  generalized  procedure  will  
be  used  to  create  more  fine-­‐grained  and  online  backups:  

1. Create  a  classical  backup  to  a  separate  staging  file  system  (on  EBS  storage)  
2. Create    an  EBS  snapshot  of  the  staging  file  system  

An  EBS  snapshot  is  automatically  persisted  onto  highly  available  Amazon  S3  storage.  Multiple  snapshots  
of  a  file  system  will  be  stored  incrementally,  which  means  that  only  changed  blocks  with  respect  to  the  
previous  snapshot  will  be  stored  to  Amazon  S3.    

The  following  sections  will  apply  the  above  described  general  procedure  to  the  different  backup  types  
for  SAP  systems  on  MaxDB  or  DB2.  

SAP  on  MaxDB  backups  using  AWS  Infrastructure  


Figure  1  illustrates  the  previously  described  backup  procedure  as  applied  to  an  SAP  system  installed  on  
MaxDB  on  Linux.    

Page  8  of  31  


 
 
Figure  1:  SAP  on  MaxDB,  overview  of  backup  types  and  procedures  using  AWS  infrastructure  

In  the  figure,  each  file  system  tag  listed  in  Table  1  is  represented  by  an  EBS  volume  symbol.  The  
following  3  types  of  backup  sequences  are  displayed:  

1. OS-­‐EXE:  backup  of  the  OS-­‐EXE  file  systems  to  the  OS-­‐EXE-­‐BACKUPS  file  system,  using  an  OS  
specific  copy  program  like  tar  on  Linux.  Subsequent  persistence  into  Amazon  S3  by  creating  a  
snapshot  of  the  EBS  volume  that  holds  the  OS-­‐EXE-­‐BACKUPS  file  system.  
2. DB-­‐DATA:  MaxDB  COMPLETE  or  INCREMENTAL  backup  to  the  DB-­‐DATA-­‐BACKUPS  file  system.  
Subsequent  persistence  onto  Amazon  S3  by  creating  an  EBS  snapshot  of  the  volume  that  holds  
the  DB-­‐DATA-­‐BACKUPS  file  system.  
NOTE:  DIRECT  SNAPSHOTS  OF  DB-­‐DATA  AND  DB-­‐LOG  VOLUMES  ARE  ALSO  POSSIBLE  USING  MAXDB  I/O  
SUSPEND/RESUME  FUNCTIONALITY,  BUT  THESE  METHODS  ARE  NOT  WITHIN  THE  SCOPE  OF  THIS  GUIDE.  
3. DB-­‐LOG:  MaxDB  LOG  backup  to  the  DB-­‐LOG-­‐BACKUPS  file  system.  Subsequent  persistence  into  
Amazon  S3  by  creating  an  EBS  snapshot  of  the  volume  that  holds  the  DB-­‐LOG-­‐BACKUPS  file  
system.  
NOTE:  MAXDB  LOG  BACKUPS  CAN  ALSO  BE  AUTOMATED  USING  ITS  AUTOSAVE  LOG  MECHANISM.  THIS  WILL  BE  
DESCRIBED  LATER  IN  THIS  GUIDE.  

Page  9  of  31  


 
SAP  on  DB2  backups  using  AWS  Infrastructure  
Similar  as  for  MaxDB,  Figure  2  displays  the  data  types  and  its  associated  backups  of  an  SAP  system  on  
DB2  on  Linux.    

 
Figure  2:  SAP  on  DB2,  overview  of  backup  types  and  procedures  using  AWS  infrastructure  

As  mentioned  before  in  the  section  “Storage  layout  of  SAP  systems  on  EBS  volumes”,  DB2  functionality  
is  used  to  manage  transaction  log  archiving  itself  to  the  DB2-­‐LOG-­‐ARCHIVE  file  system,  which  also  
resides  on  an  EBS  volume  (Figure  2).  Direct  EBS  snapshots  can  be  made  of  this  file  system,  so  that  its  
contents  are  regularly  persisted  onto  Amazon  S3  storage.  It  is  recommended  to  create  an  EBS  snapshot  
of  the  associated  volume  each  time  a  new  transaction  log  has  been  archived  into  the  DB2-­‐LOG-­‐ARCHIVE  
file  system.    

Note  that  EBS  snapshots  are  written  incrementally  to  Amazon  S3,  differential  to  the  previous  snapshot,  
which  means  that  previously  snapshotted  data  are  not  stored  over  and  over  again,  just  once.  This  makes  
the  snapshot  procedure  fast  and  cost  efficient.  

Recapitulating  for  DB2,  the  following  backup  procedures  can  be  distinguished:  

Page  10  of  31  


 
1. OS-­‐EXE:  backup  of  the  OS-­‐EXE  file  systems  to  the  OS-­‐EXE-­‐BACKUPS  file  system,  using  an  OS  
specific  copy  program  like  tar  on  Linux.  Subsequent  persistence  onto  Amazon  S3  by  creating  a  
snapshot  of  the  EBS  volume  that  holds  the  OS-­‐EXE-­‐BACKUPS  file  system.  
2. DB2-­‐DATA:  DB2  full  and/or  incremental  backups  to  the  DB-­‐DATA-­‐BACKUPS  file  system.  
Subsequent  persistence  onto  Amazon  S3  by  creating  an  EBS  snapshot  of  the  volume  that  holds  
the  DB-­‐DATA-­‐BACKUPS  file  system.  
NOTE:  DIRECT  SNAPSHOTS  OF  DB2-­‐DATA  AND  DB2-­‐LOG  VOLUMES  ARE  ALSO  POSSIBLE  LEVERAGING  DB2  I/O  
SUSPEND/RESUME  FUNCTIONALITY,  BUT  THESE  METHODS  ARE  NOT  WITHIN  THE  SCOPE  OF  THIS  GUIDE.  
3. DB2-­‐LOG-­‐ARCHIVE:  direct  persistence  onto  Amazon  S3  by  creating  an  EBS  snapshot  of  the  
volume  that  holds  the  DB2-­‐LOG-­‐ARCHIVE  file  system,  each  time  a  new  DB2  transaction  log  has  
been  archived  by  the  DBMS  into  this  file  system.  

Restore  
For  each  backup  type,  its  last  backup  can  usually  directly  be  restored  from  its  associated  staging  file  
system.  If  the  file  staging  system  is  not  accessible  anymore  or  an  older  backup  is  required,  a  new  EBS  
volume  can  be  created  out  of  a  snapshot  that  was  created  in  the  past  of  the  parent  EBS  volume.  The  
new  EBS  volume  can  then  be  attached  and  mounted  onto  the  Amazon  EC2  instance  where  restore  and  
(database)  recovery  is  taking  place.    

Common  backup  and  restore  operations  on  Amazon  EC2  instances  and  EBS  
volumes  
This  section  briefly  documents  some  common  operations  on  Amazon  EC2  instances  and  EBS  volumes  
that  are  used  for  backup  and  restore  purposes.  The  operations  are  described  as  if  they  would  be  
performed  from  the  graphical  Amazon  Management  Console,  available  at  

 https://console.aws.amazon.com/ec2    

However,  all  operations  can  be  fully  automated  using  the  Amazon  EC2  web  service  API  and/or  
Command  Line  Tools.  For  more  information  please  visit:    

http://aws.amazon.com/documentation/ec2/  

Backup:  creating  a  new  EBS  volume  with  an  empty  file  system  
1) Create  a  new  EBS  volume  
a. Log  in  to  AWS  EC2  Management  Console    
https://console.aws.amazon.com/ec2  
1. On  Volumes,  click  on  Create  Volume  
2. Type  the  size  
3. Select  the  same  availability  zone  as  the  AWS  instance  to  be  attached  
b. Select  the  volume  
c. Click  on  Attach  Volume  
1. Select  the  instance  
2. Choose  a  free  device  name  (write  this  name  down)  

Page  11  of  31  


 
2) Create  the  file  system  
a. Log  in  to  the  instance  and  create  the  file  system  

pvcreate /dev/sdX
vgcreate vgbackup /dev/sdX
lvcreate -L <SIZE> -n backups vgbackup
mkfs.ext3 /dev/vgbackup/backups
mkdir -p /backups
mount /dev/vgbackup/backups /backups
NOTE:  THE  COMMAND  HAS  BEEN  PROVIDED  AS  AN  EXAMPLE;  YOU  CAN  USE  LVM2  OR  DIRECT  
PARTITIONS  TO  STORE  BACKUPS.  THE  SELECTION  OF  THE  MOUNT  POINT  (/BACKUPS)  IS  ARBITRARY  

Backup:  creating  an  EBS  snapshot  onto  Amazon  S3  of  an  EBS  volume  
1) Make  sure  that  the  EBS  volume  is  not  written  to.  
NOTE:  IF  POSSIBLE,  THE  FILE  SYSTEM(S)  ON  THE  EBS  VOLUME  CAN  BE  DISMOUNTED  TO  ENSURE  THAT  NO  WRITE  
I/O  IS  OCCURRING.  WRITE  I/O  TO  A  FILE  SYSTEM  BEING  SNAPPED  CAN  CAUSE  INCONSISTENCIES  ON  THE  SNAPSHOT  
COPY.  
2) Log  in  to  AWS  EC2  Management  Console  https://console.aws.amazon.com/ec2  
3) Go  to  volumes  and  select  the  volume  to  snapshot  
4) Click  on  Create  Snapshot  
5) Type  the  name  and  description.  
NOTE:  CHOOSE  A  UNIQUE  BUT  EASILY  IDENTIFIABLE  NAME  THAT  INCLUDES  A  TIMESTAMP  AND/OR  SEQUENCE  
NUMBER  
6) Click  on  Yes  Create  
7) You  can  monitor  the  progress  on  Snapshot  navigation  menu  

Backup:  dismounting  file  system(s)  and  detaching  an  EBS  Volume  


a. Log  in  to  the  EC2  instance,  dismount  and  remove  the  volume  

umount /backups
vgchange vgbackup –a n
vgexport vgbackup
 
b. Log  in  to  AWS  EC2  Management  Console  
https://console.aws.amazon.com/ec2  
c. Go  to  Volumes  and  select  the  volume  to  remove  
d. Click  on  Detach  Volume  
e. Click  on  Yes,  Detach  on  the  popup  window  
f. When  detached,  click  on  Delete  volume  

Backup:  creating  a  full  offline  Amazon  EC2  Amazon  Machine  Image  (AMI)  
If  the  SAP  system  can  be  shut  down  for  a  period  of  time,  a  full  image  of  the  system  can  be  created.  The  
Amazon  Machine  Image  (AMI)  offline  backup  creates  a  snapshot  of  each  EBS  volume  and  stores  it  onto  
Amazon  S3  storage.  The  AMI  can  be  used  as  a  golden  image,  to  spin  up  new  instances  in  case  of:  

Page  12  of  31  


 
1. Recovery  from  a  complete  source  system  loss  (DR)  
To  restore  the  root  volume  and  bare  structure  of  the  system.  Typically,  the  database  file  systems  
will  be  outdated  too  much  to  be  rolled  forward,  so  these  should  be  restored  separately  as  
described  in  subsequent  sections.  
2. Set  up  of  new  systems      

A  full  AMI  backup  should  be  performed  (at  least)  after  the  SAP  system  is  installed,  but  best  after  each  
low-­‐level  change  of  the  OS,  SAP  or  DBMS,  like  for  instance:  

• Changing  the  file  system  layout  


• Upgrading  OS,  SAP  or  DBMS  binaries  
• Installing  new  additional  software  dependencies  

Briefly,  the  steps  to  create  a  full  offline  AMI  are  as  follows:  

1. Stop  the  SAP  and  database  instances  


2. Select  the  instance  and  create  the  AMI  from  the  AWS  Management  Console  
The  operating  system  will  be  stopped  and  started  again  automatically  during  the  process.  
3. Monitor  the  AMI  creation  until  successful  completion  
4. Start  the  database  and  SAP  instances  again  

The  detailed  steps  will  be  described  next.  

Detailed  steps  to  create  the  Amazon  Machine  Image  (AMI)  


1) Log  on  into  the  OS  and  shutdown  SAP  

su - <sidadm>
stopsap all
 

2) Make  sure  that  the  SAP  and  database  instances  are  shut  down  completely  by  monitoring  the  
processes  and  logs.  If  the  database  was  not  able  to  shut  down  due  to  still  active  connections,  
then  issue  the  following  commands:  
 
For  MaxDB  

su – <sidadm>
dbmcli db_offline
 

For  DB2  

su – db2<sid>
db2stop force
 
3) Log  in  to  AWS  EC2  Management  Console  
https://console.aws.amazon.com/ec2    

Page  13  of  31  


 
4) Click  ‘Instances‘  
5) Click  on  the  Instance  you  wish  to  create  a  AMI  of  
6) Right-­‐mouse  click  ‘Create  Image  (EBS  AMI)’  
7) Type  the  AMI  name  and  description  
8) Click  on  ‘Create  This  Image‘  
9) A  snapshot  will  be  created  in  parallel  for  each  EBS  volume  in  the  Amazon  EC2  instance  
Snapshot  and  AMI  creation  can  be  monitored  within  the  ‘Snapshots‘  or  ‘AMIs‘  section.  
As  the  Amazon  EC2  instance  will  be  restarted  automatically  after  EBS  AMI  creation  has  finished,  
you  can  also  monitor  in  parallel  to  see  when  the  system  comes  back  up  again,  by  for  instance  
pinging  its  IP  address.  
10) Log  on  into  the  instance  and  startup  SAP  

su - <sidadm>
startsap all
 

The  AMI  should  now  be  available  for  Amazon  EC2  instance  deployment  in  the  ‘AMIs’  section.  

Examples  for  backing  up  SAP  System  components  using  AWS  infrastructure  
The  following  sections  merely  provide  basic  examples  for  backing  up  MaxDB  and  DB2  LUW  databases  on  
AWS  infrastructure.  Please  consult  the  documentation  referenced  in  section  “Prerequisite  Documents”  
at  the  beginning  of  this  guide  as  a  complete  reference.    

In  addition,  specific  references  will  be  provided  within  the  sections  wherever  suitable.  

Example  1:  database  backup  to  an  EBS  backup  file  system    
To  recapitulate,  the  following  general  procedure  will  be  followed:  

1) Online  Database  backups  are  performed  to  EBS  volume  dedicated  for  backups      
2) Transaction  log  backups  are  performed  to  EBS  volume  dedicated  to  offline  (archived)  DB  logs  
3) The  EBS  volumes  for  DB  backups  and  offline/archived  logs  are  snapshot  on  a  recurring  basis  to  
Amazon  S3.    The  incremental  snapshots  ensure  point  in  time  database  recovery  in  case  of  
disaster.  
4) Snapshots  are  tagged  with  descriptions  of  backup  type  and  time.  

Backups  are  created  using  the  DBMS  tools  listed  in  Table  3.  

Table  3:  DBMS  tools  used  for  backup  

DBMS   Backup  Tools  

MaxDB   MaxDB  Database  Studio  and  dbmcli  (Command  


Line  Interface)
DB2  LUW   DB2  standard  commands  

Page  14  of  31  


 
Example  1a:  full  online  data  and  log  backup  for  MaxDB  
Make  sure  the  DB-­‐DATA-­‐BACKUPS  and  DB-­‐LOG-­‐BACKUPS  file  systems  have  enough  storage  space,  and  
ensure  proper  read/write  permissions  for  these  file  systems.  The  file  system  names  listed  in  Table  1  will  
be  used  in  this  example.    If  required,  please  follow  steps  described  in  the  section  “Backup:  creating  a  
new  EBS  volume  with  an  empty  file  system”.  

Apart  from  the  general  references  mentioned  in  section  “Prerequisite  Documents”,  specific  MaxDB  
Backup  /  Recovery  examples  can  also  be  found  at  “SAP  MaxDB  HowTo”  on  the  SAP  Community  Network  
(SCN).  

Create  MaxDB  backup  templates  


a. Start  the  SAP  MaxDB  Database  Studio    
If  not  installed,  it  can  be  downloaded  from  
http://www.sdn.sap.com/irj/scn/maxdb-­‐downloads  
and  installed  on  any  computer  for  remote  management  
b. Add  the  server  and  database  into  the  landscape  (if  required)    
Go  to:  My  Landscape  /  Servers  /  (right  click)  /Add/”Server/Database”)  
c. Type  the  server  name  or  ip:  on  “Server  Name:”  and  click  next  
d. Select  the  Databases  and  click  on  Finish  
e. Log  in  to  the  database  using  the  CONTROL  user    
(My  Landscape  /  Servers  /  <servername/ip>/<DB>/(right  click)  /Login)  
f. (Right  click)  on  the  database  and  click  on    Administration  
g. Go  to  the  Backup  Tab  
h. Expand  Templates  
i. Create  a  Backup  template  for  FULL  Backup  
j. Choose  New…  
Name     <Template  Name>  can  be  FULL  

Backup  Type   COMPLETE  DATA  


Device  Type   FILE  
Backup  Tool   NONE  
Device/File   /db_data_backups/<SID>_FULL  
Compressed   Unselect  

k. Click  Ok  
l. Create  a  Backup  template  for  LOG  Backup  
Name     <Template  Name>  can  be  LOG  
 
Backup  Type     LOG  
Device  Type     FILE  
Backup  Tool     NONE  
Device/File     /db_log_backups/<SID>_LOG  
Compressed   Unselect  
 

Page  15  of  31  


 
Back  up  the  database  using  Database  Studio  
1) Start  the  SAP  MaxDB  Database  Studio  
2) Log  in  to  the  database  using  the  CONTROL  user    
(My  Landscape  /  Servers  /  <servername/ip>/<DB>/  (right  click)  /Login)  
3) (Right  click)  on  the  database  and  click  on    Administration  
4) Go  to  the  Backup  Tab  
5) Expand  Templates  
6) Right  click  on  the  template  “FULL”  and  click  on  Backup  

Back  up  the  database  transaction  logs  using  Database  Studio  


1) Start  the  SAP  MaxDB  Database  Studio  
2) Log  in  to  the  database  using  the  CONTROL  user    
(My  Landscape  /  Servers  /  <servername/ip>/<DB>/  (right  click)    /Login)  
3) (Right  click)  on  the  database  and  click  on    Administration  
4) Go  to  the  Backup  Tab  
5) Expand  Templates  
6) Right  click  on  the  template  “LOG”    and  click  on  Backup  

Back  up  the  database  using  DBMCLI  


1) Log  in  to  the  Amazon  EC2  instance  
2) Execute  the  following  commands  

su - <sid>adm
#start an utility session
dbmcli –d <SID> -U c -uUTL
#start backup using the template FULL
backup_start FULL
 
3) Wait  until  the  backup  has  completed  

Back  up  the  database  log  using  DBMCLI  


1) Log  in  to  the  Amazon  EC2  instance  
2) Execute  the  following  commands  

su - <sid>adm
#start an utility session
dbmcli –d <SID> -U c -uUTL
#start backup using the template FULL
backup_start LOG

3) Wait  until  the  backup  has  completed

Schedule  hourly  automatic  log  backup  using  Database  Studio  


1) Start  the  SAP  MaxDB  Database  Studio  
2) Log  in  to  the  database  using  the  CONTROL  user    
(My  Landscape  /  Servers  /  <servername/ip>/<DB>/  (right  click)  /Login)  

Page  16  of  31  


 
3) (Righ  click)  on  the  database  and  click  on      
Administration  Tasks/Automatic  Log  Backup  
4) Select  the  template  “LOG”  from  the  list  
5) Select  “Create  a  log  backup  every      <XX>    minutes”  
6) Type  60  on  the  <XX>  field  
7) Click  on  Activate  

The  EBS  snapshot  creation  of  the  volume  that  holds  the  DB-­‐LOG-­‐BACKUPS  file  system  can  now  be  
automated  through  a  script,  time-­‐synchronized  with  the  automatic  log  backup.  

Create  a  snapshot  to  send  the  backup  to  Amazon  S3  


Please  refer  to  the  section  “Backup:  creating  an  EBS  snapshot  onto  Amazon  S3  of  an  EBS  volume”  for  the  
detailed  steps.    

It  is  recommended  to  tag  snapshots  with  a  description  like    

• “<SID>_MAXDB_DATA_BACKUPS_<YYYY-MM-DD>”  for  the  database  data  backups  


• “<SID>_MAXDB_LOG_BACKUPS_<YYYY-MM-DD-HH-MM-SS>”  for  the  database  log  backups  

Example  1b:  full  online  backup  for  DB2  LUW  


Make  sure  that  you  have  enabled  your  DB2  database  for  rollforward  recovery.  Rollforward  recovery  
mode  enables  you  to  recover  from  a  database  backup  to  the  most  recent  point  in  time  using  archived  
database  log  files  and  is  a  prerequisite  for  taking  DB2  online  backups.  If  you  have  not  configured  your  
DB2  database  for  rollforward  recovery  mode,  you  cannot  take  online  backups.  Instead  you  will  need  to  
shutdown  the  SAP  system  and  DB2  to  take  offline  backups.  For  production  systems,  your  database  must  
be  in  rollforward  recovery  mode.  For  more  details  refer  to  the  „Database  Administration  Guide  SAP  on  
IBM  DB2  for  Linux,  UNIX,  and  Windows”  

Steps  to  enable  rollforward  recovery  and  set  up  DB2  logfile  management:  
1) Stop  SAP  and  DB2  
2) Enable  rollforward  recovery  by    updating  the  database  configuration  parameter  
LOGARCHMETH1:  

su – db2<sid>
db2 update db cfg for <SID> using logarchmeth1
DISK:/db2/<SID>/log_archive
 
The  database  will  now  be  placed  in  backup  pending  state.  A  full  database  backup  must  be  taken.  

From  this  point  onwards  DB2  will  automatically  archive  log  files  from  the  /db2/<SID>/log_dir  
filesystem  to  the  /db2/<SID>/log_archive  filesystem.  

3) Take  a  full  database  offline  backup  

db2 backup database <sid> to "/backups" compress


 

Page  17  of  31  


 
4) Configure  a  retention  period  for  your  database  backups  and  the  corresponding  database  log  
files.  For  example  if  you  want  to  keep  at  least  4  database  backups  and  you  want  to  remove  
surplus  backups  older  than  30  days  and  all  corresponding  database  log  files,  configure  DB2  in  
the  following  way:  

db2 update db cfg for <sid> using NUM_DB_BACKUP 4


db2 update db cfg for <sid> using REC_HIS_RETENTN 30
db2 update db cfg for <sid> using AUTO_DEL_REC_OBJ ON
 

5) Start  SAP  

Make  sure  the  DB2-­‐DATA-­‐BACKUPS  file  system  has  enough  storage  space,  and  proper  read/write  
permissions.  The  file  system  names  listed  in  Table  2  will  be  used  for  the  examples.      

If  required,  please  follow  steps  described  in  the  section  “Backup:  creating  a  new  EBS  volume  with  an  
empty  file  system”.  

Please  remember  to  use  the  references  mentioned  in  section  “Prerequisite  Documents”  as  primary  
documentation.  In  addition  the  SAP  on  DB2  UDB  for  Unix  and  Windows  (DB6)  forum  on  the  SAP  
Community  Network  (SCN)  can  be  consulted  for  other  questions.  The  following  sections  merely  provide  
examples  and  should  not  be  used  as  a  general  reference.  

Two  options  to  back  up  the  DB2  LUW  database  will  be  shown  in  the  following  sections:  

• Option  1:  back  up  the  database  directly  from  the  SAP  system  
• Option  2:  back  up  the  database  using  the  CLI  

Option  1:  back  up  the  database  directly  from  the  SAP  system    
1) Log  in  to  the  SAP  system  with  an  admin  user  
2) Execute  the  transaction  /nDBACOCKPIT  
3) On  the  left  screen  panel,  navigate  to    
Jobs/DBA  Planning  Calendar  
4) In  the  calendar  area,  select  any  cell  representing  time  older  than  current  time  and  click  on  ‘Add’  
button.  
5) On  the  pop-­‐up  window,  select  ‘Database  Backup  to  Device’  action,  choose  ‘Online’  backup  
mode  with  ‘Include  Logs’  option  and  enter  /db2_data_backups  in  the  ‘Device/Directory’  field  
6) Click  on  ‘Execute  Immediately’  button  

Option  2:  back  up  the  database  using  the  CLI  


1) Log  in  to  the  Amazon  EC2  instance  
2) Execute  the  following  commands:  

su – db2<sid>
#start online compressed backup including logs
db2 backup database <sid> online to "/db2_data_backups" compress \
include logs

Page  18  of  31  


 
 
3) Wait  until  the  backup  has  completed.  
You  will  see  the  following  message  when  the  backup  has  completed:  

Backup successful. The timestamp for this


backup image is : 20111220221428
 

This  message  is  shown  in  the  following  log  file,  located  in  the  backup  directory:    

<SID>.0.db2<sid>.NODE0000.CATN0000.<datetime stamp>.001
 
4) Backup  the  database  manager  configuration  required  to  rebuild  the  database:  

su – db2<sid>
cd /db2_data_backups
db2cfexp <SID>_cfg_backup.txt BACKUP
 
5) Backup  the  DB2  recovery  history  file:  

su – db2<sid>
cp /db2/<SID>/db2<sid>/NODE0000/SQL00001/db2rhist.asc \
/db2_data_backups
 

DB2  transaction  log  file  management  


It  is  recommended  to  configure  the  DB2  transaction  log  management  as  mentioned  in  the  example  
above  and  described  in  “IBM  DB2  Universal  Database  for  UNIX  and  Windows  -­‐  New  Log  File  
Management”,  section  2.2.1.  

As  from  DB2  UDB  V9.5,  automatic  log  file  retention  management  can  be  configured  in  addition.  This  is  
described  in  the  example  above  and  in  the  Database  Administration  Guide  "SAP  on  IBM  DB2  for  Linux,  
Unix  and  Windows",  section  “DB2  V9.5  and  Higher  Only:  Automatic  Log  File  and  Backup  Retention”.  

The  EBS  volume  holding  DB2-­‐LOG-­‐ARCHIVE  file  system  should  be  sent  to  Amazon  S3  on  a  regular  basis  
by  creating  a  direct  snapshot,  optimally  each  time  after  a  transaction  log  was  written  into  the  DB2-­‐LOG-­‐
ARCHIVE  file  system.  The  snapshot  can  be  taken  directly  without  dismounting  the  file  system.  

Create  a  snapshot  to  send  the  backup  to  Amazon  S3  


Please  refer  to  the  section  “Backup:  creating  an  EBS  snapshot  onto  Amazon  S3  of  an  EBS  volume”  for  the  
detailed  steps.    

It  is  recommended  to  tag  snapshots  with  a  description  like    

• “<SID>_DB2_DATA_BACKUPS_<YYYY-MM-DD>”  for  the  database  data  backups  


• “<SID>_DB2_LOG_ARCHIVE_<YYYY-MM-DD-HH-MM-SS>”  for  the  database  archive  log  
backups  

Page  19  of  31  


 
Example  2:  OS  backup  to  S3  using  the  TAR  command  
This  chapter  provides  an  example  for  creating  a  low-­‐level  operating  system  backup  that  can  be  used  in  
case  of  a  full  system  loss  (DR).  

NOTE:  THIS  OS  BACKUP  CANNOT  BE  USED  TO  RESTORE  THE  DATABASE,  AS  DATABASE  DATA  AND  TRANSACTION  LOGS  ARE  
SPECIFICALLY  EXCLUDED  FROM  THE  BACKUP.  AFTER  RESTORING  THIS  BACKUP,  RESTORE  AND  RECOVERY  OF  THE  DATABASE  
SHOULD  FOLLOW.  

The  example  procedure  is  as  follows:  

1) Ensure  you  have  enough  space  on  the  file  system  /os_exe_backups  for  the  OS  backup.    
NOTE:  DATABASE  DATA,  TRANSACTION  LOGS  AND  MOUNTED  BACKUP  FILE  SYSTEMS  WILL  BE  EXCLUDED  FROM  THIS  
BACKUP  
 
2) Start  the  OS  backup  
a. Logon  into  the  OS  start  a  tar  backup  on  the  /os_exe_backups  file  system  
NOTE:  THE  USE  OF  COMPRESSION  AFFECTS  THE  CPU  UTILIZATION  AND  BACKUP  TIME.  CHOOSE  TO  USE  
COMPRESSION  OR  NOT  BY  ADDING  OR  REMOVING  THE  “-­‐-­‐GZIP”  PARAMETER  OF  THE  TAR  COMMAND    
 
For  database  type  of  MaxDB  use  this  script  as  an  example:  

export exclusion_file=/os_exe_backups/backup-exclude-dirs.txt
export backup_file=/os_exe_backups/backup.tar.gz
export log_file=/os_exe_backups/backup.stdout
export error_log_file=/os_exe_backups/backup.stderr

#Create the exclusion file


#Exclude OS directories
>$exclusion_file
echo "/tmp" >>$exclusion_file
echo "/proc" >>$exclusion_file
echo "/sys" >>$exclusion_file
echo "/dev" >>$exclusion_file
#Exclude database files
echo "/sapdb/<SID>/sapdata1" >>$exclusion_file
echo "/sapdb/<SID>/saplog" >>$exclusion_file
#Exclude the backup directories
echo "/os_exe_backups" >>$exclusion_file
echo "/db_data_backups" >>$exclusion_file
echo "/db_log_backups" >>$exclusion_file

#Run the backup


cd /os_exe_backups
tar -v --gzip -cf $backup_file / --exclude-from=$exclusion_file >
$log_file 2> $error_log_file

#to monitor
#tail -f $log_file
#tail -f $error_log_file
NOTE:  THE  COMMAND  HAS  BEEN  PROVIDED  AS  AN  EXAMPLE,  PLEASE  TEST  AND  CHANGE  AS  REQUIRED.  

For  database  type  of  DB2  use  this  script  as  an  example:  

Page  20  of  31  


 
export exclusion_file=/os-exe-backups/exclude-dirs.txt
export backup_file=/os-exe-backups/backup.tar.gz
export log_file=/os-exe-backups/backup.stdout
export error_log_file=/os-exe-backups/backup.stderr

#Create the exclusion file


#Exclude OS directories
>$exclusion_file
echo "/tmp" >>$exclusion_file
echo "/proc" >>$exclusion_file
echo "/sys" >>$exclusion_file
echo "/dev" >>$exclusion_file

#Exclude database-related filesystems


echo "/db2/<SID>/sapdata1" >>$exclusion_file
:
echo "/db2/<SID>/sapdata<n>" >>$exclusion_file
echo "/db2/<SID>/log_dir" >>$exclusion_file
echo "/db2/<SID>/log_archive" >>$exclusion_file

#Exclude the backup directories


echo "/os-exe-backups" >>$exclusion_file
echo "/db2-data-backups" >>$exclusion_file

#Run the backup


cd /os-exe-backups
tar -v --gzip -cf $backup_file / --exclude-
from=$exclusion_file > $log_file 2> $error_log_file

#to monitor
#tail -f $log_file
#tail -f $error_log_file
 
NOTE:  THE  COMMANDS  HAVE  BEEN  PROVIDED  AS  AN  EXAMPLE,  PLEASE  TEST  AND  CHANGE  AS  REQUIRED.  

To  send  the  backup  to  Amazon  S3,  create  an  EBS  snapshot  of  the  EBS  volume  that  holds  the  
/os_exe_backups  file  system.  You  can  refer  to  the  section  “Backup:  creating  an  EBS  snapshot  onto  
Amazon  S3  of  an  EBS  volume”  for  the  required  steps.    

It  is  recommended  to  tag  the  snapshot  with  a  description  like    


“<EC2-INSTANCE-ID>_OS_EXE_BACKUPS_<YYYY-MM-DD>”,  
where  <EC2-INSTANCE-ID>  can  be  retrieved  from  the  EC2  metadata  web  service  through    

“GET http://169.254.169.254/latest/meta-data/instance-id”

Linux  tools  like  curl  or  wget  can  be  used  to  issue  the  above  HTTP  command  from  a  local  shell  on  the  
Amazon  EC2  instance.    

More  information  on  using  Amazon  EC2  instance  metadata  is  available  at  
http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/AESDG-­‐chapter-­‐instancedata.html  

Page  21  of  31  


 
Examples  for  restoring  SAP  systems  using  AWS  infrastructure  
Example  1:  restore  OS  from  Amazon  S3  using  TAR  
 

Create  an  Amazon  instance  using  the  “golden  backup”  AMI  of  the  original  system  
1) Log  in  to  the  AWS  EC2  Management  Console  
https://console.aws.amazon.com/ec2  
2) Within  AMIs  select  your  “golden  backup”  AMI  and  click  Launch  
The  “golden  backup”  AMI  should  be  created  before,  as  described  in  the  previous  section    
“Backup:  creating  a  full  offline  Amazon  EC2  Amazon  Machine  Image  (AMI)”    
a. On  Instance  Details  
1. Choose  an  instance  type  that  is  supported  for  SAP  
Check  SAP  note  1588667  –  SAP  on  Amazon  Web  Services  (AWS)  
2. Select  the  VPC  and  Subnet  
3. Click  on  Continue  
4. Type  the  IP  Address  (Select  the  same  IP  address  of  the  server  you  want  to  recover)  
5. Check  the  Termination  Protection  
6. Click  on  Continue  
7. On  Name,  type  the  name  of  the  instance  
b. On  Create  Key  Pair  
1. Select  the  existing  Key  Pairs  from  Choose  From  your  existing  Key  Pairs  
2. Click  on  Continue  
c. On  CONFIGURE  FIREWALL  
1. Choose  an  existing  SG  or  Create  a  new  one  
NOTE:  DEPENDING  ON  THE  SAP  INSTANCE  TYPE,  THE  PORTS  REQUIRED  WOULD  BE  TCP/22,  
TCP/3200,  TCP/3300,  TCP/3600.    
d. On  REVIEW  
1. Click  on  Launch  

Recreate  the  file  systems  required  for  the  recovery  


NOTE:  IT  IS  REQUIRED  TO  HAVE  EBS  SIZES  AT  LEAST  THE  SAME  SIZE  OR  LARGER  THAN  THE  ORIGINAL  SYSTEM.  ALSO  THE  
NAMES  OF  THE  MOUNT  POINTS  MUST  BE  IDENTICAL.  

1) To  create  a  new  EBS  volume  


a. On  Volumes,  click  on  Create  Volumes  
b. Type  the  Size  of  the  Volume  
c. Select  the  same  Availability  Zone  as  the  Instance  
d. Click  on  Yes,  Create  
2) To  attach  the  EBS  volume  an  instance  
a. On  Volumes,  select  the  volume  
b. Click  on  Attach  Volume  
c. Select  the  AWS  instance  on  Instances  

Page  22  of  31  


 
d. Type  the  Device  the  EBS  will  be  presented  on  the  Instance  
(write  it  down)  
e. Click  on  Yes,  Attach  
3) To  create  a  file  system  using  LVM  
a. Log  in  to  the  Amazon  EC2  instance  

For  the  MaxDB  database  type,  use  this  script  as  an  example:  

#To Create swap


pvcreate /dev/sdX
vgcreate vgswap /dev/sdX
lvcreate -L <size>M -n swap vgswap
mkswap /dev/vgswap/swap
swapon –a
#
pvcreate /dev/sdY
vgcreate vgsapdb /dev/sdY
lvcreate -L <size>M -n sapdb vgsapdb
mkfs.ext3 /dev/vgsapdb/sapdb
mkdir /sapdb
mount /dev/vgsapdb/sapdb /sapdb

#Add the mount points into the /etc/fstab


/dev/vgsapdb/sapdb /sapdb ext3 acl,user_xattr 1 1
/dev/vgswap/swap swap swap defaults 0 0
NOTE:  THESE  COMMANDS  HAVE  BEEN  PROVIDED  AS  AN  EXAMPLE,  YOU  CAN  USE  LVM2  OR  LINUX  PARTITION.  
THIS  SCENARIO  WAS  TESTED  WITH  LVM2  
 
For  the  DB2  LUW  database  type,  use  this  script  as  an  example:  

#To Create swap


pvcreate /dev/sdX
vgcreate vgswap /dev/sdX
lvcreate -L <size>M -n swap vgswap
mkswap /dev/vgswap/swap
swapon –a
#
pvcreate /dev/sdY
vgcreate vgdb2 /dev/sdY
lvcreate -L <size>M -n db2 vgdb2
mkfs.ext3 /dev/vgdb2/db2
mkdir /db2
mount /dev/vgdb2/db2 /db2

#Add the mount points into the /etc/fstab


/dev/vgdb2/db2 /db2 ext3 acl,user_xattr 1 1
/dev/vgswap/swap swap swap defaults 0 0
NOTE:  THESE  COMMANDS  HAVE  BEEN  PROVIDED  AS  AN  EXAMPLE,  YOU  CAN  USE  LVM  OR  LINUX  PARTITION.  THIS  
SCENARIO  WAS  TESTED  WITH  LVM  
 

Restore  the  OS  


1) Create  an  EBS  volume  based  on  an  Amazon  S3  OS  backup  snapshot  

Page  23  of  31  


 
a. On  Volumes,  click  on  Create  Volumes  
b. Type  the  Size  of  the  Volume  
c. Select  the  same  Availability  Zone  as  the  Instance  
d. On  Snapshot  select  the  latest  OS  backup  
e. Click  on  Yes,  Create  

2) Attach  the  volume  to  the  Instance  


3) Mount  the  file  system  
a. Log  in  to  the  Amazon  EC2  instance  
b. Mount  the  /os_exe_backup  file  system  

Example:  
vgscan
vgimport vgbackup
vgchange vgbackup -a y
mkdir /os_exe_backups
mount /dev/vgbackup/os_exe /os_exe_backups
 
4) Restore  the  OS  
 
Example:  
#Backup the current fstab
cp /etc/fstab /etc/fstab.<timestamp>
#Restore the OS
cd /
tar -zxvf /os_exe_backups/backup.tar.gz

#Rewrite the lvm information


vgcfgbackup

#restore the original fstab


cp /etc/fstab.<timestamp> /etc/fstab
#Update the /etc/hosts file and update the
IP if necessary
vi /etc/hosts

Reboot  the  instance  


 
shutdown –r –y 0
 

The  system  is  now  ready  for  database  restore  and  recovery.  

Example  2:  Restoring  and  Recovering  the  Database  from  Amazon  S3    
In  general,  restore  and  recovery  of  databases  requires  careful  planning  and  preparation  before  
execution.  Take  time  for  root-­‐cause  analysis  to  be  better  able  to  identify  the  most  efficient  recovery  
strategy  before  executing.  

Page  24  of  31  


 
For  example,  data  only  need  to  be  restored  from  Amazon  S3  if  recovery  is  not  possible  anymore  from  
data  and  backups  that  are  already  available  on  the  system.  If  a  database  needs  to  be  recovered  to  the  
latest  possible  point  in  time,  always  make  sure  to  not  overwrite  and  destroy  the  latest  database  
transaction  logs.  As  these  are  typically  not  yet  archived  and/or  backed  up  to  Amazon  S3,  the  latest  
transactions  contained  in  these  logs  could  then  get  lost  forever!  

This  guide  does  not  intend  to  replace  the  original  backup  and  restore  documentation  of  the  database  
vendors  that  was  referenced  in  section  “Prerequisite  Documents”.  As  restore  and  recovery  scenarios  are  
most  diverse  and  dependent  on  the  environment  and  failure  cause,  it  is  strongly  recommended  to  
follow  the  original  documentation  in  case  of  a  real  failure.  

The  following  sections  describe  some  basic  examples  of  MaxDB  and  DB2  database  restore  and  recovery,  
to  get  an  initial  idea  how  that  works  on  Amazon  infrastructure.  

Example  2a:  MaxDB  restore  and  recovery  from  Amazon  S3    


1) If  this  is  a  DR,  make  sure  you  have  restored  the  OS  as  described  in  section  “Example  1:  restore  OS  
from  Amazon  S3  using  TAR”  
2) If  required,  mount  the  Amazon  S3  Database  backup  on  the  Amazon  EC2    instance  
NOTE:  NORMALLY,  THE  MOST  RECENT  BACKUP  SHOULD  ALREADY  BE  AVAILABLE  ON  THE  AMAZON  EC2  INSTANCE  

a. Only  if  required,  create  a  new  EBS  volume  based  on  an  Amazon  S3  snapshot  backup  
1. On  Volumes,  click  on  Create  Volumes  
2. Type  the  Size  of  the  Volume  
3. Select  the  same  Availability  Zone  as  the  Instance  
4. On  Snapshot  select  the  latest  OS  backup  
5. Click  on  Yes,  Create  

b. Attach  the  volume  to  the  Instance  


c. Mount  the  file  system  
1. Log  in  to  the  EC2  instance    
2. Mount  the  /db_data_backups  file  system  

Example  if  using  LVM2:  


vgscan
vgimport vgbackup
vgchange vgbackup -a y
mkdir /db_data_backups
mount /dev/vgbackup/db_data /db_data_backups
 
3) If  required,  repeat  the  previous  step,  but  now  for  the  /db_log_backups  file  system  
NOTE:  NORMALLY  THE  MOST  RECENT  BACKUP  SHOULD  ALREADY  BE  AVAILABLE  ON  THE  EC2  INSTANCE  
 
4) Restore  the  database  using  Database  Studio  
a. Logon  into  the  instance  as  root  
b. Start  x_server  

Page  25  of  31  


 
c. Logon  into  the  database  using  the  CONTROL  user  
 (My  Landscape  /  Servers  /  <servername/ip>/<DB>/(right  click)  /Login)  
d. Set  the  database  in  admin  mode:  (Right  click)  on  the  database  and  click  on      
Administration  Tasks/Set  State/Admin  
e. Start  the  recovery:  (Right  click)  on  the  database  and  click  on    
Administration  Tasks/Recovery…  
f. On  the  Recovery  of  Database  window,  click  on  Recover  a  medium  
g. Select  FULL  from  the  template  list  and  then  Next  
h. Click  on  Start  
i. Click  Ok  on  the  in  the  Confirmation  Initialization  database  window  
j. Wait  until  the  restore  has  completed.  
5) Apply  transaction  logs  using  Database  Studio  
a. First  restore  the  database  using  dbmcli  or  Database  Studio  without  restarting  the  database  
b. Logon  into  the  database  using  the  CONTROL  user  (My  Landscape  /  Servers  /  
<servername/ip>/<DB>/(right  click)  /Login)  
c. Start  the  recovery:  (Right  click)  on  the  database  and  click  on    “Administration  
Tasks/Recovery…”  
d. You  can  select  if  you  want  to  restore  until  a  specific  time  (Point  in  time  recovery),  in  this  
example  we  will  restore  until  the  last  available  log,  so  do  not  select  “Recover  until  a  specific  
time.”  Select  Recover  a  medium  and  then  Click  Next    
NOTE:  CREATE  AN  EBS  VOLUME  FROM  AN  AMAZONS3  LOG  BACKUP  SNAPSHOT  –  FOLLOWING  THE  LOG  SEQUENCE  
–  AND  MOUNT  IT  TO  THE  /DB_LOG_BACKUPS  IF  IT  IS  REQUIRED  TO  APPLY  OLDER  LOGS  AS  THE  ONES  AVAILABLE  ON  
THE  SYSTEM.  HOWEVER,  BEFORE  DOING  SO,  MAKE  SURE  TO  HAVE  A  SNAPSHOT  AVAILABLE  OF  THE  LATEST  CONTENT  
OF  /DB_LOG_BACKUPS  FILE  SYSTEM,  AS  THAT  MIGHT  BE  REQUIRED  LATER  IN  THE  RECOVERY  PROCESS.  
e. Select  LOG    from  the  template  list  and  then  Next  
f. Select  the  Log  File  Number  and  Then  Next  (In  this  case  put  the  next  log  after  the  backup)  and  
click  Next  
g. Click  Start  
h. The  recovery  session  will  try  to  recover  all  consecutive  logs  until  it  fails  with  a  missing  log.  You  
can  choose  to  restore  more  logs  and  continue…    
i. Choose  Start  Database  and  then  Continue    
6) Restore  the  database  using  DBMCLI  
a. Log  in  to  the  Amazon  EC2  instance  as  root  
b. Execute  the  following  example  commands  

#Fix permissions
chown sdb:sdba /sapdb/<SID>/sapdata
chown sdb:sdba /sapdb/<SID>/saplog

#logon as <sid>adm
su - <sid>adm

#start the db in ADMIN mode


dbmcli -U c db_admin

Page  26  of  31  


 
#Logon a recovery session
dbmcli -U c

#List the backup history, and last logs


#take notes of the next log number

backup_history_open
backup_history_list -r last -c
label,action,pages,firstlog,lastlog,media

#If possible, check if the backup is


accessed as expected
recover_check FULL data

#Restore the database

db_connect
recover_start FULL data

c. Wait  until  the  restore  is  completed  


To  monitor  the  restore  session,  use  the  following  command  
Example:  

#logon as <sid>adm
su - <sid>adm

#start restore session session


dbmcli -U c -uUTL -d <SID>
recover_state

Monitor  the  “Pages  Transferred”  and  “Pages  Left”  


7) Restore  the  database  logs  using  CLI  
a. Before  restoring  logs,  you  should  have  had  restored  a  database  without  restarting,  (see  step  1)  
or  step  3)).  The  DB  should  be  in  ADMIN  mode  
b. Logon  into  the  instance  as  root  
c. Execute  the  following  

su - <sid>adm

#Logon a recovery session


dbmcli -U c

#List the backup history, and last logs


#take notes of the next log number

backup_history_open
backup_history_list -r last -c
label,action,pages,firstlog,lastlog,media

service_connect

Page  27  of  31  


 
#Restore the Logs where <XXX> is the next log sequence.
db_connect
recover_start LOG log <XXX>

#If the recovery ends with -8020 error code and you still
have logs to recover that are not listed in the backup
history, you can continue with the following commands,
where <YYY> is the next log to recover, recover log by log
until you restore the latest available log.

recover_replace LOG /backuplog/<SID>_LOG.<YYY>

#Use the following command to review the status of the


database after the restore

db_restartinfo
#If the consistent=1 the database can start

d. Use  the  following  commands  to  start  the  recovery  if  you  want  to  restore  in  point  in  time  
recovery  

su - <sid>adm

#Logon a recovery session


dbmcli -U c

#List the backup history, and last logs


#take notes of the next log number

backup_history_open
backup_history_list -r last -c
label,action,pages,firstlog,lastlog,media

service_connect

#Restore the Logs where <XXX> is the next log sequence.

db_connect
recover_start LOG log <XXX> UNTIL <date> <time>
#If the recovery ends with -8020 error code and you still
have logs to recover that are not listed in the backup
history, you can continue with the following commands,
where <YYY> is the next log to recover, recover log by log
until you restore the latest available log.

recover_replace LOG /backuplog/<SID>_LOG.<YYY>

#Note: the database will be put on ONLINE automatically


after the DB is recovered until the time specified

8) Start  the  SAP  instance  

Page  28  of  31  


 
 
su - <sidadm>
startsap all
 
At  this  point  the  SAP  instance  should  be  started  with  no  issues  
 
9) After  restoring  the  database,  you  can  remove  the  EBS  if  required  
a. Logon  into  the  Amazon  EC2    instance  
b. dismount  and  remove  the  volume  
Example:  

umount /db_data_backups
vgchange vgbackup –a n
vgexport vgbackup
 
c. Logon  in  to  the  AWS  EC2  Management  Console  
https://console.aws.amazon.com/ec2/    
d. Go  to  Volumes  and  select  the  volume  to  remove  
e. Click  on  Detach  Volume  
f. Click  on  Yes,  Detach  on  the  popup  window  
g. When  detached,  click  on  Delete  Volume  

Example  2b:  DB2  LUW  restore  and  recovery  from  Amazon  S3  
1) If  this  is  an  DR,  make  sure  you  have  restored  the  OS  
2) Mount  the  Amazon  S3  Database  backup  on  the  instance  

a. Create  an  EBS  volume  based  on  an  Amazon  S3  DB  backup  snapshot  
1. On  Volumes,  click  on  Create  Volumes  
2. Type  the  Size  of  the  Volume  
3. Select  the  same  Availability  Zone  as  the  Instance  
4. On  Snapshot  select  the  latest  DB  backup  
5. Click  on  Yes,  Create  

b. Attach  the  volume  to  the  Instance  


c. Mount  the  file  system  
1. Logon  to  the  instance  using  putty,  or  any  other  ssh  client  
2. Mount  the  backup  filesystem  

Example:  
vgscan
vgimport vgbackup
vgchange vgbackup -a y
mkdir /backups
mount /dev/vgbackup/backups /db2_data_backups
 

Page  29  of  31  


 
3) Mount  the  Amazon  S3  Database  logs  on  the  instance  
NOTE:  THIS  IS  STEP  ONLY  REQUIRED  IF  THE  AVAILABLE  ARCHIVED  LOGS  ARE  NOT  SUFFICIENT  FOR  RECOVERY,  OR  IF  THEY  
ARE  NOT  AVAILABLE  ANYMORE  

a. Create  an  EBS  volume  based  on  an  AmazonS3  log  backup  snapshot  
1. On  Volumes,  click  on  Create  Volumes  
2. Type  the  Size  of  the  Volume  
3. Select  the  same  Availability  Zone  as  the  Instance  
4. On  Snapshot  select  the  latest  log  backup  
5. Click  on  Yes,  Create  

b. Attach  the  volume  to  the  Instance  


c. Mount  the  FS  
NOTE:  FOR  SAFETY  REASONS,  THE  FILE  SYSTEM  WILL  BE  MOUNTED  TO  ANOTHER  MOUNT  POINT  AS  THE  
ARCHIVE  LOG  FILE  SYSTEM  ITSELF.  
1. Logon  to  the  instance  using  putty,  or  any  other  ssh  client  
2. Mount  the  backup  filesystem  

Example:  
vgscan
vgimport vgbackuplog
vgchange vgbackuplog -a y
mkdir /backuplogs
mount /dev/vgbackuplog/backuplogs /backuplogs
 

4) Restore  and  recover  the  database  using  CLI  


a. Logon  into  the  instance  as  root  
b. Execute  the  following  commands  

# Log on as db2<sid>
su – db2<sid>

# If required, copy additional DB logs to the default log archive location


# For example:
cp –rp /backuplogs/* /db2/<SID>/log_archive

# Start database manager


db2start

# Start the db recovery


db2 recover db <sid> using history file (/db2_data_backups/db2rhist.asc)

# Once DB recovery is done, restore configuration stored outside of DB


cd /db2_data_backups
db2cfimp <SID>_cfg_backup.txt
 

5) Start  the  SAP  system  


 

Page  30  of  31  


 
su - <sidadm>
startsap all
 
 

Page  31  of  31  


 

You might also like