-
Notifications
You must be signed in to change notification settings - Fork 1.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
High-level Design of Storage Monitoring Daemon #1481
Conversation
…DB schema per review.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@assrinivasan Can you draw a class inheritance diagram to show StorageDevices, SSdBase, SsdUtil, EmmcUtil, etc
@assrinivasan please add the PR link to the description of this PR |
done. |
Done in latest commit |
doc/storagemond/storagemond-hld.md
Outdated
|
||
1. **Planned cold, fast, and warm reboot scenario** | ||
|
||
- Prior to invoking an OS-level reboot, the latest FSIO Read and Write metrics are captured from the `/proc/diskstats` file and stored into the `fsio-rw-stats.json` by executing the `fsio-rw-sync` script from the respective reboot script (cold, soft, or warm). |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@assrinivasan can we keep fsio-rw-sync
as a systemd service that is invoked during system reboot/shutdown?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Addressed in latest commit
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
After discussion we have decided to remove the sync script and instead adopt the following strategy:
cold/soft-reboot: When reboot script sends SIGTERM to pmon which in turn sends SIGTERM to stormond, we catch that signal and sync the STATE_DB values to the JSON file.
warm-reboot: We add the STORAGE_INFO|
key to the backup_database() function in the fast-reboot script so that the values survive the system reboot.
I will make above changes to HLD.
…e script. Changed CONFIG_DB interval types.
doc/storagemond/storagemond-hld.md
Outdated
| **Event** | **State_DB** | **JSON** | **PROCFS STATUS** | **JSON SYNCED WITH `STATE_DB`?** | **STORMON RESTARTED** | | ||
| ---------------------- | ------------ | --------- | ----------------------- | ----------------------------- | --------------------- | | ||
| | | | | | | | ||
| Init | CLEARED | CLEARED | CLEARED, Initial Values | YES | YES | |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@assrinivasan What is Init? first boot or any boot up?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@assrinivasan CLEARED or EMPTY?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
First boot - will make this change
EMPTY - will make this change
…#18657) #### Why I did it This is part of a larger feature: [SONiC Storage Monitoring Daemon](sonic-net/SONiC#1481) -- this commit adds the option to configure the daemon's polling interval and fsstats file sync interval (in seconds) of the daemon via config_db by introducing YANG models. #### How I did it Gives userside the option to dynamically a new table 'STORMOND' with key INTERVALS and fields 'daemon_polling_interval' with default value of '3600' seconds and 'fsstats_sync_interval' with default value of '86400' seconds as defined in the YANG model. #### How to verify it Flash image onto a DUT and add the aforementioned table to the CONFIG_DB. Verify that `stormond` has picked up your config intervals.
@prgeor can you please double check if the 1 open code PR can be merged today? Otherwise, we will move to backlog for future release. |
Hi @zhangyanzhao -- all PRs have been merged. |
This PR is intended to harden the high-level design of a newly-proposed Storage monitoring daemon.
[List of changes]
Signed-off-by: Ashwin Srinivasan assrinivasan@microsoft.com