gsadmin/rustfs
1
0
Fork 0
mirror of https://github.com/rustfs/rustfs.git synced 2026-09-03 02:38:12 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(heal): resume remote rebuilds after target restart (#6941)

Browse Source 
* fix(heal): retry unavailable recreate targets

* fix(heal): refresh put-file epochs after target restart

* test(e2e): harden heal restart evidence

Co-Authored-By: heihutu <heihutu@gmail.com>

* test(e2e): cancel competing heal before restart

Co-Authored-By: heihutu <heihutu@gmail.com>

---------

Co-authored-by: houseme <housemecn@gmail.com>
Co-authored-by: heihutu <heihutu@gmail.com>
This commit is contained in:
Henry Guo
2026-08-31 19:39:47 +08:00
committed by GitHub
parent 896781a52b
commit 61821a6f3e
11 changed files with 1169 additions and 58 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/ecstore/src/set_disk/ops/heal.rs
+26 -10
View File
@@ -36,6 +36,20 @@ const EVENT_HEAL_OBJECT_RENAME: &str = "heal_object_rename";
const HEAL_RENAME_INCOMPLETE: &str = "heal rename incomplete";
const READ_REPAIR_DATA_PHASE_TIMEOUT: std::time::Duration = std::time::Duration::from_secs(60 * 60);
fn heal_drive_state_for_error(error: &DiskError) -> DriveState {
match error {
DiskError::DiskNotFound | DiskError::RemoteClientUnavailable(_) => DriveState::Offline,
DiskError::FaultyDisk | DiskError::FaultyRemoteDisk => DriveState::Faulty,
DiskError::FileNotFound
| DiskError::FileVersionNotFound
| DiskError::VolumeNotFound
| DiskError::PartMissingOrCorrupt
| DiskError::OutdatedXLMeta => DriveState::Missing,
DiskError::FileCorrupt => DriveState::Corrupt,
_ => DriveState::Unknown(error.to_string()),
}
}
#[cfg(test)]
static HEAL_RENAME_FAILURES: std::sync::Mutex<Vec<(String, String, usize)>> = std::sync::Mutex::new(Vec::new());
@@ -892,16 +906,7 @@ impl SetDisks {
}
let drive_state = match reason {
Some(err) => match err {
DiskError::DiskNotFound => DriveState::Offline.to_string(),
DiskError::FileNotFound
| DiskError::FileVersionNotFound
| DiskError::VolumeNotFound
| DiskError::PartMissingOrCorrupt
| DiskError::OutdatedXLMeta => DriveState::Missing.to_string(),
DiskError::FileCorrupt => DriveState::Corrupt.to_string(),
_ => DriveState::Unknown(err.to_string()).to_string(),
},
Some(err) => heal_drive_state_for_error(&err).to_string(),
None => DriveState::Ok.to_string(),
};
result.before.drives.push(HealDriveInfo {
@@ -2673,6 +2678,17 @@ mod heal_result_report_tests {
assert!(!super::metadata_less_part_file("xl.meta"));
}
#[test]
fn unavailable_heal_errors_use_stable_drive_states() {
for error in [DiskError::FaultyDisk, DiskError::FaultyRemoteDisk] {
assert_eq!(super::heal_drive_state_for_error(&error).to_string(), DriveState::Faulty.to_string());
}
assert_eq!(
super::heal_drive_state_for_error(&DiskError::RemoteClientUnavailable("peer restarting".to_string())).to_string(),
DriveState::Offline.to_string()
);
}
#[test]
fn read_repair_commit_fingerprint_tracks_commit_identity_only() {
let data_dir = Uuid::parse_str("11111111-1111-1111-1111-111111111111").expect("data dir should parse");