Template
1
0
Fork 0
mirror of https://codeberg.org/forgejo/forgejo synced 2024-11-27 12:16:10 +01:00
forgejo/modules/queue/base_levelqueue.go
wxiaoguang 84c8ab9fd1
Help to recover from corrupted levelqueue (#24912)
gitea.com experienced the corrupted LevelQueue bug again.

I think the problem is clear now: if the keys in LevelDB went
out-of-sync, the LevelQueue itself doesn't have the ability to recover,
eg:

* LevelQueue.Len() reports 100
* LevelQueue.LPop() reports ErrNotFound = errors.New("no key found")

So it needs to dive into the LevelDB to remove all keys to recover the
corrupted LevelQueue.

More comments are in TestCorruptedLevelQueue.
2023-05-29 10:52:32 +08:00

84 lines
2.2 KiB
Go

// Copyright 2023 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package queue
import (
"context"
"sync/atomic"
"code.gitea.io/gitea/modules/nosql"
"code.gitea.io/gitea/modules/queue/lqinternal"
"gitea.com/lunny/levelqueue"
"github.com/syndtr/goleveldb/leveldb"
)
type baseLevelQueue struct {
internal atomic.Pointer[levelqueue.Queue]
conn string
cfg *BaseConfig
db *leveldb.DB
}
var _ baseQueue = (*baseLevelQueue)(nil)
func newBaseLevelQueueGeneric(cfg *BaseConfig, unique bool) (baseQueue, error) {
if unique {
return newBaseLevelQueueUnique(cfg)
}
return newBaseLevelQueueSimple(cfg)
}
func newBaseLevelQueueSimple(cfg *BaseConfig) (baseQueue, error) {
conn, db, err := prepareLevelDB(cfg)
if err != nil {
return nil, err
}
q := &baseLevelQueue{conn: conn, cfg: cfg, db: db}
lq, err := levelqueue.NewQueue(db, []byte(cfg.QueueFullName), false)
if err != nil {
return nil, err
}
q.internal.Store(lq)
return q, nil
}
func (q *baseLevelQueue) PushItem(ctx context.Context, data []byte) error {
c := baseLevelQueueCommon(q.cfg, nil, func() baseLevelQueuePushPoper { return q.internal.Load() })
return c.PushItem(ctx, data)
}
func (q *baseLevelQueue) PopItem(ctx context.Context) ([]byte, error) {
c := baseLevelQueueCommon(q.cfg, nil, func() baseLevelQueuePushPoper { return q.internal.Load() })
return c.PopItem(ctx)
}
func (q *baseLevelQueue) HasItem(ctx context.Context, data []byte) (bool, error) {
return false, nil
}
func (q *baseLevelQueue) Len(ctx context.Context) (int, error) {
return int(q.internal.Load().Len()), nil
}
func (q *baseLevelQueue) Close() error {
err := q.internal.Load().Close()
_ = nosql.GetManager().CloseLevelDB(q.conn)
q.db = nil // the db is not managed by us, it's managed by the nosql manager
return err
}
func (q *baseLevelQueue) RemoveAll(ctx context.Context) error {
lqinternal.RemoveLevelQueueKeys(q.db, []byte(q.cfg.QueueFullName))
lq, err := levelqueue.NewQueue(q.db, []byte(q.cfg.QueueFullName), false)
if err != nil {
return err
}
old := q.internal.Load()
q.internal.Store(lq)
_ = old.Close() // Not ideal for concurrency. Luckily, the levelqueue only sets its db=nil because it doesn't manage the db, so far so good
return nil
}