swuferhong opened a new issue, #3538:
URL: https://github.com/apache/fluss/issues/3538

   ### Search before asking
   
   - [x] I searched in the [issues](https://github.com/apache/fluss/issues) and 
found nothing similar.
   
   
   ### Fluss version
   
   0.9.0 (latest release)
   
   ### Please describe the bug 🐞
   
   When `TabletServers` are close to full disk usage, a `TabletServer` shutdown 
can trigger leader migration. For log tables, leader migration usually does not 
need to download much new data, so the existing data disk write protection 
limits the impact.
   
   For primary-key tables, however, making a replica leader may initialize KV 
state by downloading remote KV snapshots. If this happens while the target 
`TabletServers` is already above the data disk write-limit ratio, the snapshot 
download can fill the disk and make the cluster harder to recover.
   
   This can turn a single `TabletServer` shutdown under high disk usage into 
broader disk pressure across the cluster, because newly elected primary-key 
table leaders may consume additional local disk space during recovery.
   
   ### Solution
   
   _No response_
   
   ### Are you willing to submit a PR?
   
   - [ ] I'm willing to submit a PR!


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to