Neste post veremos como aplicar Release Update no Oracle Grid Infrastructure usando a automação do Fleet Patching and Provisioning. O processo consiste em provisionar uma nova Working Copy do Grid com os Release Updates já aplicados no Grid Home e realizar o switch do Grid Home atual para o novo. Dessa forma, faremos a atualização usando a estratégia Out-Of-Place Patching com downtime mínimo de poucos segundos.
Neste exemplo estou usando o servidor lab01 que tem o Grid Infrastructure 19c com Release Update 11 e farei a atualização para o Release Update 12. Neste ambiente há uma instância orcl rodando na versão 19c Release Update 3.
FPP Target (lab01): Versão do Grid Home:
[grid@lab01 ~]$ $ORACLE_HOME/OPatch/opatch lspatches 32585572;DBWLM RELEASE UPDATE 19.0.0.0.0 (32585572) 32584670;TOMCAT RELEASE UPDATE 19.0.0.0.0 (32584670) 32579761;OCW RELEASE UPDATE 19.11.0.0.0 (32579761) 32576499;ACFS RELEASE UPDATE 19.11.0.0.0 (32576499) 32545013;Database Release Update : 19.11.0.0.210420 (32545013) OPatch succeeded.
FPP Server: Consultando Gold Images disponíveis
[grid@fppserver01 ~]$ rhpctl query image fppserver01.dibiei.com: Audit ID: 144 Image name: gi19cR11 Image name: db19cR11 Image name: db12cR1 Image name: gi19cR12 Image name: db19cR3
Vamos usar a Gold Image gi19cR12:
[grid@fppserver01 ~]$ rhpctl query image -image gi19cR12 fppserver01.dibiei.com: Audit ID: 145 Image name: gi19cR12 Owner: grid@fppserver Site: fppserver Access control: USER:grid@fppserver Access control: ROLE:OTHER Access control: ROLE:GH_IMG_PUBLISH Access control: ROLE:GH_IMG_ADMIN Access control: ROLE:GH_IMG_VISIBILITY Parent Image: Software home path: /fpp_images/images/igi19cR1294719/.ACFS/snaps/igi19cR12/swhome Image state: PUBLISHED Image size: 15189 Megabytes Image Type: ORACLEGISOFTWARE Image Version: 19.0.0.0.0:19.12.0.0.0 Groups configured in the image: OSDBA=asmdba,OSASM=asmadmin,OSBACKUP=asmadmin,OSDG=asmadmin,OSKM=asmadmin,OSRAC=asmadmin Image platform: Linux_AMD64 Interim patches installed: 32918050,32916816,32915586,32904851,32585572 Contains a non-rolling patch: FALSE Complete: TRUE
Para atualizar o Grid Infrastructure, precisamos fazer o deploy de uma Working Copy usando a opção softwareonly, no qual um novo Grid Home será instalado no servidor, mas o Oracle Restart e ASM não serão configurados.
O comando utilizado neste exemplo:
rhpctl add workingcopy -image gi19cR12 -workingcopy wc_lab01_gi19cR12 -path /u01/app/product/19.12.0.0/grid/ -oraclebase /u01/app/grid -targetnode lab01 -softwareonly -root
Provisionando a Working Copy no lab01:
[grid@fppserver01 ~]$ rhpctl add workingcopy -image gi19cR12 -workingcopy wc_lab01_gi19cR12 -path /u01/app/product/19.12.0.0/grid/ -oraclebase /u01/app/grid -targetnode lab01 -softwareonly -root Enter user "root" password: fppserver01.dibiei.com: Audit ID: 139 fppserver01.dibiei.com: Storing metadata in repository for working copy "wc_lab01_gi19cR12" ... fppserver01.dibiei.com: Creating snapshot "tmpgi19cR12wc_lab01_gi19cR12" ... fppserver01.dibiei.com: Changing the home ownership to user grid... fppserver01.dibiei.com: Copying software contents to Local File System ... fppserver01.dibiei.com: Starting clone operation... fppserver01.dibiei.com: Using inventory file /etc/oraInst.loc to clone ... lab01: lab01: lab01: [INFO] [INS-32183] Use of clone.pl is deprecated in this release. Clone operation is equivalent to performing a Software Only installation from the image. lab01: You must use /u01/app/product/19.12.0.0/grid/gridSetup.sh script available to perform the Software Only install. For more details on image based installation, refer to help documentation. lab01: lab01: Starting Oracle Universal Installer... lab01: lab01: [WARNING] [INS-32029] The Installer has detected that the Oracle Base location is not empty. lab01: ACTION: Oracle recommends that the Oracle Base location is empty. lab01: You can find the log of this install session at: lab01: /u01/app/oraInventory/logs/cloneActions2021-08-04_11-26-58PM.log lab01: .................................................. 5% Done. lab01: .................................................. 10% Done. lab01: .................................................. 15% Done. lab01: .................................................. 20% Done. lab01: .................................................. 25% Done. lab01: .................................................. 30% Done. lab01: .................................................. 35% Done. lab01: .................................................. 40% Done. lab01: .................................................. 45% Done. lab01: .................................................. 50% Done. lab01: .................................................. 55% Done. lab01: .................................................. 60% Done. lab01: .................................................. 65% Done. lab01: .................................................. 70% Done. lab01: .................................................. 75% Done. lab01: .................................................. 80% Done. lab01: .................................................. 85% Done. lab01: .......... lab01: Copy files in progress. lab01: lab01: Copy files successful. lab01: lab01: Link binaries in progress. lab01: .......... lab01: Link binaries successful. lab01: lab01: Setup files in progress. lab01: .......... lab01: Setup files successful. lab01: lab01: Setup Inventory in progress. lab01: lab01: Setup Inventory successful. lab01: .......... lab01: Finish Setup successful. lab01: The cloning of wc_lab01_gi19cR12 was successful. lab01: Please check '/u01/app/oraInventory/logs/cloneActions2021-08-04_11-26-58PM.log' for more details. lab01: lab01: Setup Oracle Base in progress. lab01: lab01: Setup Oracle Base successful. lab01: .................................................. 95% Done. lab01: lab01: As a root user, execute the following script(s): lab01: 1. /u01/app/product/19.12.0.0/grid/root.sh lab01: lab01: lab01: lab01: .................................................. 100% Done. fppserver01.dibiei.com: Successfully executed clone operation. fppserver01.dibiei.com: Executing root script on nodes [lab01]. lab01: Check /u01/app/product/19.12.0.0/grid/install/root_lab01_2021-08-04_23-27-59-970690222.log for the output of root script fppserver01.dibiei.com: Successfully executed root script on nodes [lab01]. fppserver01.dibiei.com: Oracle home provisioned. fppserver01.dibiei.com: Working copy creation completed.
Agora faremos o move do Oracle Grid Infrastructure (Oracle Restart e ASM) para o novo Grid Home.
Usamos o comando rhpctl move gihome, abaixo todas as opções possíveis:
[grid@fppserver01 ~]$ rhpctl move gihome -h
Moves the Oracle Grid Infrastructure from the source working copy or source home path to the destination working copy.
Usage: rhpctl move gihome -destwc <destination_workingcopy_name>
{{-sourcewc <source_workingcopy_name> |
-sourcehome <oracle_home_path>}
[-targetnode <node_name>]
[-ignorewcpatches]
[-nonrolling]
[-keepplacement]
[-auto -dbhomes <mapping_of_Oracle_homes>
[-dblist <db_name_list> |
-excludedblist <db_name_list>]
[-nodatapatch]
[-disconnect]
[-stopoption <stop_option>]
[-drain_timeout <session_drain_time>]
[-dbsinparallel <number_of_instances>]
[-raconetimeout <timeout>]]
[-batches <list_of_batches> |
-smartmove
[-saf <availability>]
[-separate]]
[-eval]
[-schedule
{ <timer_value> |
NOW }]
[[-aupath <gold_image_path>
[-agpath <read_write_path>]] |
[-tgip
[-nodriverupdate]]]
[-ignoremissingpatches <patch_name1>
[,<patch_name2>...]] |
-continue |
-revert |
-abort |
-forcecomplete}
[-root |
-cred <cred_name> |
-sudouser <sudo_username> -sudopath <path_to_sudo_binary> |
-auth <plugin_name>
[-arg1 <name1>:<value1>
[-arg2 <name2>:<value2>...]]]
[-cleanpids]
[-useractiondata <user_action_data>]
[ -image <image_name>]
[-smtpfrom "<address>"]
[-smtpto "<addresses>"]
[-iso_repo <iso_repository>]
[-patchmgrloc <patch_mgr_loc>]
-destwc <destination_workingcopy_name> Name of the destination working copy to which the Oracle Grid Infrastructure needs to be moved
-sourcewc <source_workingcopy_name> Name of the source working copy from which the Oracle Grid Infrastructure needs to be moved
-sourcehome <oracle_home_path> Source Oracle home path
-targetnode <node_name> Name of a node in a remote cluster with no Rapid Home Provisioning Client
-ignorewcpatches Ignores if the patched working copy is missing some patches which are present in the source path or working copy
-nonrolling Move the Oracle home in non-rolling mode
-keepplacement Ensure that services of administrator-managed Oracle RAC or Oracle RAC One databases are running on the same instances before and after the move operation
-auto Automatically patch databases along with patching Grid Infrastructure
-dbhomes <mapping_of_Oracle_homes> Mapping of source and destination working copies in the format: "<source_wc1>=<dest_wc1>,...,<source_oracle_home_path>=<dest_wcN>"
-dblist <db_name_list> Patch only the specified databases
-excludedblist <db_name_list> Patch all databases except the specified databases
-nodatapatch Indicates that 'datapatch' should not be executed for databases being moved
-disconnect Disconnect all sessions before stopping or relocating services
-stopoption <stop_option> Stop option for database: ABORT, IMMEDIATE, NORMAL, TRANSACTIONAL, TRANSACTIONAL_LOCAL
-drain_timeout <session drain time> Service drain timeout specified in seconds
-dbsinparallel <number_of_instances> Number of database instances that can be started in parallel on a given node.
-raconetimeout <timeout> RAC One Node database relocation timeout in minutes.
-batches <list_of_batches> List of batches of nodes in the format: "(Ba),...,(Bz)"
-smartmove Auto-generate the list of batches of nodes and move databases by restarting instances batch after batch
-saf <availability> Service availability factor, the minimum percentage of instances on which a service must remain running during the move
-separate Process batches separately. Move operation pauses for user intervention before continuing with the next batch
-eval Evaluate without executing the command. For Smart Move, print the auto-generated batches of nodes and sequence of moves
-aupath <gold_image_path> Read-only gold image path for the persistent Oracle home path
-agpath <read_write_path> Read/write path for the persistent Oracle home path
-tgip Perform transparent move of Oracle Grid Infrastructure home
-nodriverupdate skip patching the drivers if the patch contains a driver patch
-continue Continue restarting CRS stack on the next batch of nodes
-revert Revert to source Oracle home or working copy in case of batch-move or smartmove
-abort Abort the ongoing 'move' operation
-forcecomplete Mark move as completed after it has been manually completed
-root Use root credentials to access the remote node
-sudouser <username> perform super user operations as sudo user name
-sudopath <sudo_binary_path> location of sudo binary
-cred <cred_name> Credential name to associate the user/password credentials to access a remote node
-auth <plugin_name> [<plugin_args>] Use an authentication plugin to access the remote node
-cleanpids cleans up processes during Oracle Grid Infrastructure move operation
-useractiondata <user_action_data> Value to be passed to useractiondata parameter of useraction script
-schedule { <timer_value> | NOW } Preferred time to execute the operation. If an absolute timer_value is specified, it should be in ISO-8601 format. For example: 2016-12-21T19:13:17+05. If offset is specified as the timer value, it should be in the format +dd:mm:yy:hh:mm:ss. For example: +02:22:22. If NOW is specified, job will be scheduled immediately.
-ignoremissingpatches Proceed with the move/upgrade even though the specified patches, which are present in the source path or working copy, might be missing from the destination path or working copy
-ignoredbstarterror Ignore any database startup errors during Oracle Grid Infrastructure patching.
Iremos usar as seguintes opções:
- -destwc: O nome da Working Copy para qual estamos fazendo o move (a que ficará ativa no final do processo).
- -sourcehome: Indica o caminho do Grid Home atual. Essa opção é usada quando o Grid Home atual foi instalado manualmente (não foi provisionado como uma Working Copy pelo Fleet Patching and Provisioning).
Comando:
rhpctl move gihome -path /u01/app/product/19.11.0.0/grid -destwc wc_lab01_gi19cR12 -root
Executando o move do Grid Home 19.11 para 19.12:
[grid@fppserver01 ~]$ rhpctl move gihome -sourcehome /u01/app/product/19.11.0.0/grid -destwc wc_lab01_gi19cR12 -root Enter user "root" password: fppserver01.dibiei.com: Audit ID: 140 PRCG-1242 : failed to execute a remote operation on the Rapid Home Provisioning Client PRGO-1284 : Rapid Home Provisioning Client "ORACLERESTART_lab01" was never started. [grid@fppserver01 ~]$ rhpctl move gihome -sourcehome /u01/app/product/19.11.0.0/grid -destwc wc_lab01_gi19cR12 -targetnode lab01 -root Enter user "root" password: fppserver01.dibiei.com: Audit ID: 141 fppserver01.dibiei.com: verifying versions of Oracle homes ... fppserver01.dibiei.com: verifying owners of Oracle homes ... fppserver01.dibiei.com: verifying groups of Oracle homes ... fppserver01.dibiei.com: starting to move the Oracle Grid Infrastructure home from home location "/u01/app/product/19.11.0.0/grid" to "/u01/app/product/19.12.0.0/grid" on client cluster "ORACLERESTART_lab01" fppserver01.dibiei.com: Executing user action dummy on nodes "lab01"... lab01: Wed Aug 4 23:33:03 -03 2021 fppserver01.dibiei.com: Done executing user action dummy... fppserver01.dibiei.com: retrieving status of databases ... fppserver01.dibiei.com: retrieving status of services of databases ... fppserver01.dibiei.com: relocating services of databases ... fppserver01.dibiei.com: stopping services of databases ... fppserver01.dibiei.com: stopping instances of databases ... fppserver01.dibiei.com: Executing prepatch and postpatch on nodes: "lab01". fppserver01.dibiei.com: Executing root script on nodes [lab01]. lab01: Using configuration parameter file: /u01/app/product/19.12.0.0/grid/crs/install/crsconfig_params lab01: The log of current session can be found at: lab01: /u01/app/grid/crsdata/lab01/crsconfig/haunlock__2021-08-04_11-33-39PM.log lab01: 2021/08/04 23:33:40 CLSRSC-347: Successfully unlock /u01/app/product/19.12.0.0/grid fppserver01.dibiei.com: Successfully executed root script on nodes [lab01]. fppserver01.dibiei.com: Executing root script on nodes [lab01]. lab01: Using configuration parameter file: /u01/app/product/19.12.0.0/grid/crs/install/crsconfig_params lab01: The log of current session can be found at: lab01: /u01/app/grid/crsdata/lab01/crsconfig/roothas_2021-08-04_11-33-41PM.log lab01: 2021/08/04 23:33:54 CLSRSC-329: Replacing Clusterware entries in file 'oracle-ohasd.service' fppserver01.dibiei.com: Successfully executed root script on nodes [lab01]. fppserver01.dibiei.com: starting instances of databases "orcl" ... fppserver01.dibiei.com: Updating inventory on nodes: lab01. lab01: Starting Oracle Universal Installer... lab01: lab01: Checking swap space: must be greater than 500 MB. Actual 2040 MB Passed lab01: The inventory pointer is located at /etc/oraInst.loc lab01: You can find the log of this install session at: lab01: /u01/app/oraInventory/logs/UpdateNodeList2021-08-04_11-35-09PM.log lab01: 'UpdateNodeList' was successful. fppserver01.dibiei.com: Updated inventory on nodes: lab01. fppserver01.dibiei.com: Updating inventory on nodes: lab01. lab01: Starting Oracle Universal Installer... lab01: lab01: Checking swap space: must be greater than 500 MB. Actual 2040 MB Passed lab01: The inventory pointer is located at /etc/oraInst.loc lab01: You can find the log of this install session at: lab01: /u01/app/oraInventory/logs/UpdateNodeList2021-08-04_11-35-12PM.log lab01: 'UpdateNodeList' was successful. fppserver01.dibiei.com: Updated inventory on nodes: lab01. fppserver01.dibiei.com: Completed the 'move gihome' operation on cluster ORACLERESTART_lab01.
Operação concluída com sucesso.
Verificando o downtime do ASM (alert log da instância ASM) :
2021-08-04T23:33:42.011184-03:00
Shutting down ORACLE instance (immediate) (OS id: 31284)
2021-08-04T23:34:52.551878-03:00
SUCCESS: ALTER DISKGROUP ALL MOUNT /* asm agent call crs *//* {0:0:2} */
Avaliando o horário em que o shutdown da instância iniciou e o horário em que todos os diskgroups foram montados após incialização da instância ASM no novo Grid Home, o downtime foi de apenas 70 segundos (isso é realmente ótimo). Esse é o diferencial da estratégia Out-Of-Place Patching, em um cenário com aplicação de patch In-Place, esse tempo poderia levar 15~30 minutos.
Verificando que o serviços estão rodando no novo Grid Home:
[grid@lab01 ~]$ . oraenv <<< +ASM ORACLE_SID = [+ASM] ? The Oracle base remains unchanged with value /u01/app/grid [grid@lab01 ~]$ crsctl query has releasepatch Oracle Clusterware release patch level is [3998055650] and the complete list of patches [32585572 32904851 32915586 32916816 32918050 ] have been applied on the local node. The release patch string is [19.12.0.0.0].
Verificando downtime da instância de banco de dados (alert log da instância):
2021-08-04T23:33:08.232314-03:00
Shutting down ORACLE instance (immediate) (OS id: 30958)
2021-08-04T23:35:09.092623-03:00
Completed: ALTER DATABASE OPEN /* db agent *//* {0:0:140} */
Como a instância RDBMS é baixada antes da instância ASM e é iniciada depois, o downtime do banco de dados foi de 2 minutos.
Conclusão
Neste post foi demonstrado um cenário de atualização do Grid Infrastructuture em uma instalação Standalone Server para um novo Release Update usando a automação do Fleet Patching and Provisioning, que eleva os benefícios da estratégia OOP (Out-Of-Place) para reduzir o downtime e simplificar o rollback em caso de falha da operação.