Enable adding new database to an existing VMs #226
Conversation
….sh script was previosly run. (google#63) * Add the add_database functionality * Add the add_database functionality * Fixed the undefined oracle-home var error * Fxing more issues with undefined vars * Fxing more issues with undefined vars * Fxing more issues with undefined vars * Fixed the add_database.sh skipping db creation * Fixed misc. issues with the db crreation and handling the listener * Fixed misc. issues with the db crreation and handling the listener * Fixed misc. issues with the db crreation and handling the listener * Fixed misc. issues with the db crreation and handling the listener * Fixing issue with missing sw location * Another attempt to fix listener issues * 1. Issues are caught as early as possible 2. Error messages are clear and explain why the operation can't proceed 3. Users understand the resource requirements before starting the time-consuming DB creation process * 1. Adding an explicit conversion step that converts the string values to floating-point numbers using the float filter in Ansible 2. Creating new variables for the floating-point values: sga_size_gb_float and pga_size_gb_float 3. Using these floating-point variables in the calculation * Fixing issues with memory check * Fixing issues with memory checks * Improve memory calculation precision in db-create role - Change memory check command from 'free -g' to 'free -m' for more precise measurements - Convert memory values to GB with decimal precision using division and rounding - Minor formatting adjustment to commented verbosity line * Fixing memory check * Fixing memory check * Fixing memory check * Fixing memory check * Fixing memory check * Fixing memory check * Fixing memory check * Fixing backup script generation code * Remove default swlib_unzip_path fallback in config-db.yml * Configure Oracle control file autobackup format for RECO disk group * Update Oracle home detection to use 'oracle' binary instead of 'sqlplus' * Fixing backup config issue * Fixing how we handle the + prefix in the --ora-reco-destination * Refactored the oracle home validaiton logic to be in a different file * Code dediplication * Improved docs for adding a new db * Changed the multiline shell command to a single line as requested. * Updated the documentation to reflect that we check for the 'oracle' binary instead of 'sqlplus'. * Docs improvments and code cleanup/deduplication * 1. Removed the redundant default setting for swlib_unzip_path in roles/db-create/tasks/main.yml and replaced it with a comment explaining that it's already defined in group_vars/all.yml. 2. Removed all inline default fallbacks (swlib_unzip_path | default('/tmp/oracle_swlib')) in the following places: - The template destination path for the register_db_listener.sh.j2 template - The shell command that executes the register_db script - The file path when removing the registration script * Improved the way we look for the oracle home * Improved the way we look for the oracle home * Made the parameter for setting Oracle Home mandatory and only way to determine the Oracle home * Fixing handling of the input parameter backup_dest properly * Be optional for the install-oracle.sh script, using a best-practice default path when not provided: : * Be optional for the install-oracle.sh script, using a best-practice default path when not provided: : * explicitly converts memory_pct to an integer using the | int filter in two places * The fix adds explicit integer conversion to both ansible_memory_mb.real.total and memory_pct variables before performing the calculation. This ensures that both values are integers before the floor division operation (//), preventing the error where Ansible was attempting to use floor division between a string and an integer.
|
Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA). View this failed invocation of the CLA check for more information. For the most up to date status, view the checks section at the bottom of the pull request. |
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: gorjant-pythian The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
gorjant-pythian
force-pushed
the
rel/oratk-58-v2
branch
from
9bf2628 to
0af1fdc
Compare
There was a problem hiding this comment.
Thanks for all the coding work here! Adding so much code and copying shell logic has me worried about technical debt we might be taking on. Can I suggest something much simpler: we can create a new database using the existing install-oracle.sh with the --config-db argument. There are rough edges: the default memory_pct of 45 doesn't work well with DBs. We don't convert the data type in the calculation. We unnecessarily run datapatch. But it does work in my tests. THe new DB gets registered with the OCR and the listener too.
$ bash install-oracle.sh --instance-ssh-key '~/.ssh/ansible' --ora-swlib-bucket gs://gcp-oracle-software --ora-swlib-type gcs --ora-swlib-path /swlib --ora-version 18 --backup-dest /u01/backups --config-db --ora-db-name new --ora-db-container false --cluster-type RAC --cluster-config cluster.json --ora-reco-diskgroup DATA -- --extra-vars "sga_target=528m"
https://gist.github.com/mfielding/26731bca921cf2c5406ce4f1481ed773
.ansible/.lock
Outdated
.cursor/rules/coding-rules.mdc
Outdated
| ) | ||
| ) | ||
|
|
||
| LOGGING_{{ listener_name }} = ON |
There was a problem hiding this comment.
Does this differ from the default?
Please add a trailing newline.
|
@mfielding My idea is to have a cleaner new code for adding new databases, which would not be mixed with the install-oracle.sh. I tried initially doing that, but there were a number of complexities in dealing with checks to make sure we skipped GI and RDBMS installations/patching, etc. But if we are going with the parameter --db-config, that might make things simpler since we would skip RDBMS/GI tasks if this parameter is specified. So the plan would be to do these changes tot he install-oracle.sh execution path:
Probably it would be better to do this in a new branch. |
Consider a more intelligent memory sizing algorithm
gorjant-pythian
force-pushed
the
rel/oratk-58-v2
branch
from
785c8cb to
253f87a
Compare
|
I have updated the code to cover the suggestions in this PR as well. |
Summary list of changes being introduced: