2019-10-01 16:40:17 +03:00
|
|
|
// Copyright 2019 The Gitea Authors. All rights reserved.
|
2022-11-27 21:20:29 +03:00
|
|
|
// SPDX-License-Identifier: MIT
|
2019-10-01 16:40:17 +03:00
|
|
|
|
|
|
|
package mirror
|
|
|
|
|
|
|
|
import (
|
2019-12-15 12:51:28 +03:00
|
|
|
"context"
|
2019-10-01 16:40:17 +03:00
|
|
|
"fmt"
|
|
|
|
|
2021-12-10 04:27:50 +03:00
|
|
|
repo_model "code.gitea.io/gitea/models/repo"
|
2020-01-14 06:38:04 +03:00
|
|
|
"code.gitea.io/gitea/modules/graceful"
|
2019-10-01 16:40:17 +03:00
|
|
|
"code.gitea.io/gitea/modules/log"
|
2022-07-08 22:45:12 +03:00
|
|
|
mirror_module "code.gitea.io/gitea/modules/mirror"
|
2021-10-17 14:43:25 +03:00
|
|
|
"code.gitea.io/gitea/modules/queue"
|
2019-10-01 16:40:17 +03:00
|
|
|
"code.gitea.io/gitea/modules/setting"
|
|
|
|
)
|
|
|
|
|
2021-10-17 14:43:25 +03:00
|
|
|
// doMirrorSync causes this request to mirror itself
|
2022-07-08 22:45:12 +03:00
|
|
|
func doMirrorSync(ctx context.Context, req *mirror_module.SyncRequest) {
|
2022-03-02 10:43:11 +03:00
|
|
|
if req.ReferenceID == 0 {
|
|
|
|
log.Warn("Skipping mirror sync request, no mirror ID was specified")
|
|
|
|
return
|
|
|
|
}
|
2021-10-17 14:43:25 +03:00
|
|
|
switch req.Type {
|
2022-07-08 22:45:12 +03:00
|
|
|
case mirror_module.PushMirrorType:
|
2022-03-02 10:43:11 +03:00
|
|
|
_ = SyncPushMirror(ctx, req.ReferenceID)
|
2022-07-08 22:45:12 +03:00
|
|
|
case mirror_module.PullMirrorType:
|
2022-03-02 10:43:11 +03:00
|
|
|
_ = SyncPullMirror(ctx, req.ReferenceID)
|
2021-10-17 14:43:25 +03:00
|
|
|
default:
|
2022-03-02 10:43:11 +03:00
|
|
|
log.Error("Unknown Request type in queue: %v for MirrorID[%d]", req.Type, req.ReferenceID)
|
2021-10-17 14:43:25 +03:00
|
|
|
}
|
|
|
|
}
|
2019-10-01 16:40:17 +03:00
|
|
|
|
2021-11-23 06:09:35 +03:00
|
|
|
var errLimit = fmt.Errorf("reached limit")
|
|
|
|
|
2019-10-01 16:40:17 +03:00
|
|
|
// Update checks and updates mirror repositories.
|
2021-11-23 06:09:35 +03:00
|
|
|
func Update(ctx context.Context, pullLimit, pushLimit int) error {
|
2021-09-07 18:49:36 +03:00
|
|
|
if !setting.Mirror.Enabled {
|
|
|
|
log.Warn("Mirror feature disabled, but cron job enabled: skip update")
|
|
|
|
return nil
|
|
|
|
}
|
2019-10-01 16:40:17 +03:00
|
|
|
log.Trace("Doing: Update")
|
2021-06-14 20:20:43 +03:00
|
|
|
|
2023-07-05 06:41:32 +03:00
|
|
|
handler := func(idx int, bean any) error {
|
2022-02-08 17:02:32 +03:00
|
|
|
var repo *repo_model.Repository
|
2022-07-08 22:45:12 +03:00
|
|
|
var mirrorType mirror_module.SyncType
|
|
|
|
var referenceID int64
|
|
|
|
|
2021-12-10 04:27:50 +03:00
|
|
|
if m, ok := bean.(*repo_model.Mirror); ok {
|
2022-06-15 18:58:44 +03:00
|
|
|
if m.GetRepository() == nil {
|
2021-06-14 20:20:43 +03:00
|
|
|
log.Error("Disconnected mirror found: %d", m.ID)
|
|
|
|
return nil
|
|
|
|
}
|
2022-02-08 17:02:32 +03:00
|
|
|
repo = m.Repo
|
2022-07-08 22:45:12 +03:00
|
|
|
mirrorType = mirror_module.PullMirrorType
|
|
|
|
referenceID = m.RepoID
|
2021-12-10 04:27:50 +03:00
|
|
|
} else if m, ok := bean.(*repo_model.PushMirror); ok {
|
2022-06-15 18:58:44 +03:00
|
|
|
if m.GetRepository() == nil {
|
2021-06-14 20:20:43 +03:00
|
|
|
log.Error("Disconnected push-mirror found: %d", m.ID)
|
|
|
|
return nil
|
|
|
|
}
|
2022-02-08 17:02:32 +03:00
|
|
|
repo = m.Repo
|
2022-07-08 22:45:12 +03:00
|
|
|
mirrorType = mirror_module.PushMirrorType
|
|
|
|
referenceID = m.ID
|
2021-06-14 20:20:43 +03:00
|
|
|
} else {
|
|
|
|
log.Error("Unknown bean: %v", bean)
|
2019-10-01 16:40:17 +03:00
|
|
|
return nil
|
|
|
|
}
|
2021-06-14 20:20:43 +03:00
|
|
|
|
2021-11-23 06:09:35 +03:00
|
|
|
// Check we've not been cancelled
|
2019-12-15 12:51:28 +03:00
|
|
|
select {
|
|
|
|
case <-ctx.Done():
|
2021-11-23 06:09:35 +03:00
|
|
|
return fmt.Errorf("aborted")
|
2019-12-15 12:51:28 +03:00
|
|
|
default:
|
|
|
|
}
|
2021-11-23 06:09:35 +03:00
|
|
|
|
|
|
|
// Push to the Queue
|
2022-07-08 22:45:12 +03:00
|
|
|
if err := mirror_module.PushToQueue(mirrorType, referenceID); err != nil {
|
2022-02-08 17:02:32 +03:00
|
|
|
if err == queue.ErrAlreadyInQueue {
|
2022-07-08 22:45:12 +03:00
|
|
|
if mirrorType == mirror_module.PushMirrorType {
|
2022-02-08 17:02:32 +03:00
|
|
|
log.Trace("PushMirrors for %-v already queued for sync", repo)
|
|
|
|
} else {
|
|
|
|
log.Trace("PullMirrors for %-v already queued for sync", repo)
|
|
|
|
}
|
|
|
|
return nil
|
|
|
|
}
|
2021-11-23 06:09:35 +03:00
|
|
|
return err
|
|
|
|
}
|
|
|
|
return nil
|
2021-06-14 20:20:43 +03:00
|
|
|
}
|
|
|
|
|
2022-02-08 17:02:32 +03:00
|
|
|
pullMirrorsRequested := 0
|
2021-11-23 06:09:35 +03:00
|
|
|
if pullLimit != 0 {
|
2023-07-05 06:41:32 +03:00
|
|
|
if err := repo_model.MirrorsIterate(pullLimit, func(idx int, bean any) error {
|
2022-02-28 22:41:06 +03:00
|
|
|
if err := handler(idx, bean); err != nil {
|
|
|
|
return err
|
|
|
|
}
|
|
|
|
pullMirrorsRequested++
|
|
|
|
return nil
|
2021-11-23 06:09:35 +03:00
|
|
|
}); err != nil && err != errLimit {
|
|
|
|
log.Error("MirrorsIterate: %v", err)
|
|
|
|
return err
|
|
|
|
}
|
2021-06-14 20:20:43 +03:00
|
|
|
}
|
2022-02-28 22:41:06 +03:00
|
|
|
|
2022-02-08 17:02:32 +03:00
|
|
|
pushMirrorsRequested := 0
|
2021-11-23 06:09:35 +03:00
|
|
|
if pushLimit != 0 {
|
2023-07-05 06:41:32 +03:00
|
|
|
if err := repo_model.PushMirrorsIterate(ctx, pushLimit, func(idx int, bean any) error {
|
2022-02-28 22:41:06 +03:00
|
|
|
if err := handler(idx, bean); err != nil {
|
|
|
|
return err
|
|
|
|
}
|
|
|
|
pushMirrorsRequested++
|
|
|
|
return nil
|
2021-11-23 06:09:35 +03:00
|
|
|
}); err != nil && err != errLimit {
|
|
|
|
log.Error("PushMirrorsIterate: %v", err)
|
|
|
|
return err
|
|
|
|
}
|
2019-10-01 16:40:17 +03:00
|
|
|
}
|
2022-02-08 17:02:32 +03:00
|
|
|
log.Trace("Finished: Update: %d pull mirrors and %d push mirrors queued", pullMirrorsRequested, pushMirrorsRequested)
|
2020-05-17 02:31:38 +03:00
|
|
|
return nil
|
2019-10-01 16:40:17 +03:00
|
|
|
}
|
|
|
|
|
Rewrite queue (#24505)
# ⚠️ Breaking
Many deprecated queue config options are removed (actually, they should
have been removed in 1.18/1.19).
If you see the fatal message when starting Gitea: "Please update your
app.ini to remove deprecated config options", please follow the error
messages to remove these options from your app.ini.
Example:
```
2023/05/06 19:39:22 [E] Removed queue option: `[indexer].ISSUE_INDEXER_QUEUE_TYPE`. Use new options in `[queue.issue_indexer]`
2023/05/06 19:39:22 [E] Removed queue option: `[indexer].UPDATE_BUFFER_LEN`. Use new options in `[queue.issue_indexer]`
2023/05/06 19:39:22 [F] Please update your app.ini to remove deprecated config options
```
Many options in `[queue]` are are dropped, including:
`WRAP_IF_NECESSARY`, `MAX_ATTEMPTS`, `TIMEOUT`, `WORKERS`,
`BLOCK_TIMEOUT`, `BOOST_TIMEOUT`, `BOOST_WORKERS`, they can be removed
from app.ini.
# The problem
The old queue package has some legacy problems:
* complexity: I doubt few people could tell how it works.
* maintainability: Too many channels and mutex/cond are mixed together,
too many different structs/interfaces depends each other.
* stability: due to the complexity & maintainability, sometimes there
are strange bugs and difficult to debug, and some code doesn't have test
(indeed some code is difficult to test because a lot of things are mixed
together).
* general applicability: although it is called "queue", its behavior is
not a well-known queue.
* scalability: it doesn't seem easy to make it work with a cluster
without breaking its behaviors.
It came from some very old code to "avoid breaking", however, its
technical debt is too heavy now. It's a good time to introduce a better
"queue" package.
# The new queue package
It keeps using old config and concept as much as possible.
* It only contains two major kinds of concepts:
* The "base queue": channel, levelqueue, redis
* They have the same abstraction, the same interface, and they are
tested by the same testing code.
* The "WokerPoolQueue", it uses the "base queue" to provide "worker
pool" function, calls the "handler" to process the data in the base
queue.
* The new code doesn't do "PushBack"
* Think about a queue with many workers, the "PushBack" can't guarantee
the order for re-queued unhandled items, so in new code it just does
"normal push"
* The new code doesn't do "pause/resume"
* The "pause/resume" was designed to handle some handler's failure: eg:
document indexer (elasticsearch) is down
* If a queue is paused for long time, either the producers blocks or the
new items are dropped.
* The new code doesn't do such "pause/resume" trick, it's not a common
queue's behavior and it doesn't help much.
* If there are unhandled items, the "push" function just blocks for a
few seconds and then re-queue them and retry.
* The new code doesn't do "worker booster"
* Gitea's queue's handlers are light functions, the cost is only the
go-routine, so it doesn't make sense to "boost" them.
* The new code only use "max worker number" to limit the concurrent
workers.
* The new "Push" never blocks forever
* Instead of creating more and more blocking goroutines, return an error
is more friendly to the server and to the end user.
There are more details in code comments: eg: the "Flush" problem, the
strange "code.index" hanging problem, the "immediate" queue problem.
Almost ready for review.
TODO:
* [x] add some necessary comments during review
* [x] add some more tests if necessary
* [x] update documents and config options
* [x] test max worker / active worker
* [x] re-run the CI tasks to see whether any test is flaky
* [x] improve the `handleOldLengthConfiguration` to provide more
friendly messages
* [x] fine tune default config values (eg: length?)
## Code coverage:
![image](https://user-images.githubusercontent.com/2114189/236620635-55576955-f95d-4810-b12f-879026a3afdf.png)
2023-05-08 14:49:59 +03:00
|
|
|
func queueHandler(items ...*mirror_module.SyncRequest) []*mirror_module.SyncRequest {
|
|
|
|
for _, req := range items {
|
2021-10-17 14:43:25 +03:00
|
|
|
doMirrorSync(graceful.GetManager().ShutdownContext(), req)
|
2019-10-01 16:40:17 +03:00
|
|
|
}
|
2022-01-23 00:22:14 +03:00
|
|
|
return nil
|
2020-10-25 10:32:25 +03:00
|
|
|
}
|
|
|
|
|
2019-10-01 16:40:17 +03:00
|
|
|
// InitSyncMirrors initializes a go routine to sync the mirrors
|
|
|
|
func InitSyncMirrors() {
|
Rewrite queue (#24505)
# ⚠️ Breaking
Many deprecated queue config options are removed (actually, they should
have been removed in 1.18/1.19).
If you see the fatal message when starting Gitea: "Please update your
app.ini to remove deprecated config options", please follow the error
messages to remove these options from your app.ini.
Example:
```
2023/05/06 19:39:22 [E] Removed queue option: `[indexer].ISSUE_INDEXER_QUEUE_TYPE`. Use new options in `[queue.issue_indexer]`
2023/05/06 19:39:22 [E] Removed queue option: `[indexer].UPDATE_BUFFER_LEN`. Use new options in `[queue.issue_indexer]`
2023/05/06 19:39:22 [F] Please update your app.ini to remove deprecated config options
```
Many options in `[queue]` are are dropped, including:
`WRAP_IF_NECESSARY`, `MAX_ATTEMPTS`, `TIMEOUT`, `WORKERS`,
`BLOCK_TIMEOUT`, `BOOST_TIMEOUT`, `BOOST_WORKERS`, they can be removed
from app.ini.
# The problem
The old queue package has some legacy problems:
* complexity: I doubt few people could tell how it works.
* maintainability: Too many channels and mutex/cond are mixed together,
too many different structs/interfaces depends each other.
* stability: due to the complexity & maintainability, sometimes there
are strange bugs and difficult to debug, and some code doesn't have test
(indeed some code is difficult to test because a lot of things are mixed
together).
* general applicability: although it is called "queue", its behavior is
not a well-known queue.
* scalability: it doesn't seem easy to make it work with a cluster
without breaking its behaviors.
It came from some very old code to "avoid breaking", however, its
technical debt is too heavy now. It's a good time to introduce a better
"queue" package.
# The new queue package
It keeps using old config and concept as much as possible.
* It only contains two major kinds of concepts:
* The "base queue": channel, levelqueue, redis
* They have the same abstraction, the same interface, and they are
tested by the same testing code.
* The "WokerPoolQueue", it uses the "base queue" to provide "worker
pool" function, calls the "handler" to process the data in the base
queue.
* The new code doesn't do "PushBack"
* Think about a queue with many workers, the "PushBack" can't guarantee
the order for re-queued unhandled items, so in new code it just does
"normal push"
* The new code doesn't do "pause/resume"
* The "pause/resume" was designed to handle some handler's failure: eg:
document indexer (elasticsearch) is down
* If a queue is paused for long time, either the producers blocks or the
new items are dropped.
* The new code doesn't do such "pause/resume" trick, it's not a common
queue's behavior and it doesn't help much.
* If there are unhandled items, the "push" function just blocks for a
few seconds and then re-queue them and retry.
* The new code doesn't do "worker booster"
* Gitea's queue's handlers are light functions, the cost is only the
go-routine, so it doesn't make sense to "boost" them.
* The new code only use "max worker number" to limit the concurrent
workers.
* The new "Push" never blocks forever
* Instead of creating more and more blocking goroutines, return an error
is more friendly to the server and to the end user.
There are more details in code comments: eg: the "Flush" problem, the
strange "code.index" hanging problem, the "immediate" queue problem.
Almost ready for review.
TODO:
* [x] add some necessary comments during review
* [x] add some more tests if necessary
* [x] update documents and config options
* [x] test max worker / active worker
* [x] re-run the CI tasks to see whether any test is flaky
* [x] improve the `handleOldLengthConfiguration` to provide more
friendly messages
* [x] fine tune default config values (eg: length?)
## Code coverage:
![image](https://user-images.githubusercontent.com/2114189/236620635-55576955-f95d-4810-b12f-879026a3afdf.png)
2023-05-08 14:49:59 +03:00
|
|
|
mirror_module.StartSyncMirrors(queueHandler)
|
2019-10-01 16:40:17 +03:00
|
|
|
}
|