616 Commits

Author SHA1 Message Date
Xiang Li
a8cc1570d0 raft: support quorum check when raft is leader
If quorum check fails, the leader will step down to follower.
2015-11-24 09:36:37 -08:00
Gyu-Ho Lee
81229dbea9 *: add missing package descriptions
This adds and updates package descriptions in etcd projects.
And also deletes some duplicate LICENSE statements.
2015-11-17 20:54:10 -08:00
Gyu-Ho Lee
e1c108e604 raft: minor typo in progress.go
Fixes a minor typo.
2015-11-17 14:21:35 -08:00
Xiang Li
5d0268aa2e Merge pull request #3877 from bdarnell/campaign-while-leader
raft: no-op instead of panic for Campaigning while leader
2015-11-16 19:59:34 -08:00
Ben Darnell
fbeb58d265 raft: no-op instead of panic for Campaigning while leader
We need to be able to force an election (on one node) after creating a
new group (cockroachdb/cockroach#1384), but it is difficult to ensure
that our call to Campaign does not race with an election that may be
started by raft itself. A redundant call to Campaign should be a no-op
instead of a panic. (But the panic in becomeCandidate remains, because
we don't want to update the term or change the committed index in this
case)
2015-11-16 21:44:14 -05:00
Yicheng Qin
3a65442d7d raft: fix print format for term in one log line
`term` should be printed in decimal representation instead of
hexadecimal one.
2015-11-15 20:26:16 -08:00
Xiang Li
2990249c1d Merge pull request #3856 from xiang90/raft_doc_restart
raft: add doc to make restart clear
2015-11-11 11:15:49 -08:00
Xiang Li
f7f28b9984 raft: add doc to make restart clear, especially for configuration changed case 2015-11-11 11:11:58 -08:00
Xiang Li
6df52614fc raft: add more words about raft protocol 2015-11-11 09:20:25 -08:00
Yicheng Qin
0de52414cd raft: extend wait timeout in TestNodeAdvance
This fixes the failure met in semaphore CI.
2015-11-03 16:57:18 -08:00
Yicheng Qin
bf3057e5bd raft: extend wait timeout in TestMultiNodeAdvance
This fixes the failure met in semaphore CI:

```
--- FAIL: TestMultiNodeAdvance-2 (0.01s)
		multinode_test.go:458: expect Ready after Advance, but there is
		no Ready available
```
2015-10-23 12:08:24 -07:00
Yicheng Qin
01806c3e80 raft: fix malformed example name
It is reported by latest govet:
```
gopath/src/github.com/coreos/etcd/raft/example_test.go:26: Example_Node
has malformed example suffix: Node
```
2015-10-20 16:40:01 -07:00
Gyu-Ho Lee
1716d5858f raft/documentation: clarify progress's subjects.
If I understand correctly, `progress` represents the states of follower. For
me, some comments weren't clear because it was missing the subjects of
`progress`. This adds more clarification on who is doing what. Please let me
know if I misunderstood anything. Thanks,
2015-10-15 19:15:08 -07:00
Cong Ding
362df8e470 raft/doc: fix misuse of `for' loop in docs 2015-10-15 11:13:30 -05:00
Cong Ding
f1f92f0fa3 raft/doc: fix typos 2015-10-15 02:17:34 -05:00
Kenji Kaneda
ebd8cb04c1 raft: fix a description of MemoryStorage.Compact
The parameter name is compactIndex, not i.
2015-10-06 21:49:33 -07:00
Cong Ding
b2edf1d24a raft: fix typo in doc 2015-10-01 11:21:23 -05:00
Yicheng Qin
533e728b64 Merge pull request #3609 from yichengq/raft-snapshot
raft: kill TODO about behavior when snapshot fails
2015-09-29 19:32:31 -07:00
Yicheng Qin
4c82b481a5 raft: improve behavior when snapshot fails
etcd is going to support incremental snapshot, and we design to let it
send at most one snapshot out at first stage. So when one snapshot is in
flight, snapshot request will return error.

When failing to get snapshot when sending MsgSnap, raft prints out
related log and abort sending this message.
2015-09-29 19:15:15 -07:00
Kenji Kaneda
f602767e50 raft: remove an obsolete TODO comment on 4MB maxMsgSize hard coding
The TODO comment was added by 7571b2cd, and it was addressed by d9b5b56c.
2015-09-28 21:31:12 -07:00
Emil Hessman
b9f22cb69b raft: fix Node doc typo 2015-09-21 06:13:33 +02:00
Ben Darnell
b7baaa6bc8 raft: Allow per-group nodeIDs in MultiNode.
This feature is motivated by
https://github.com/cockroachdb/cockroach/blob/master/docs/RFCS/replica_tombstone.md
which requires a change to the way CockroachDB constructs its node IDs.
2015-09-18 15:36:36 -04:00
Jonathan Boulle
7848ac3979 *: add missing license headers 2015-09-15 14:09:01 -07:00
Brandon Philips
68d4ec3e13 raft: improve panic error message
Give a human being some insight into how we might have gotten to this
state based on feedback from #3504.
2015-09-12 12:17:02 -07:00
Dmitry Smirnov
b2f4a5f587 *: fix spelling issues (codespell).
Signed-off-by: Dmitry Smirnov <onlyjob@member.fsf.org>
2015-09-11 10:22:29 +10:00
Xiang Li
ef7cf058a2 *: update gogoproto 2015-09-03 15:32:25 -07:00
Tamir Duberstein
45390b9fb8 *: regenerate proto to use local import path
Using Go-style import paths in protos is not idiomatic. Normally, this
detail would be internal to etcd, but the path from which gogoproto
is imported affects downstream consumers (e.g. cockroachdb).

In cockroach, we want to avoid including `$GOPATH/src` in our protoc
include path for various reasons. This patch puts etcd on the same
convention, which allows this for cockroach.

More information: https://github.com/cockroachdb/cockroach/pull/2339#discussion_r38663417

This commit also regenerates all the protos, which seem to have
drifted a tiny bit.
2015-09-03 13:38:28 -04:00
Ben Darnell
4f20e01f60 raft: Ignore proposals if not a current member.
Fixes another panic in MultiNode.Propose.
2015-08-31 20:31:14 -04:00
Xiang Li
6cbaaa715c Merge pull request #3396 from bdarnell/multinode-propose-panic
raft: Fix a nil-pointer panic in MultiNode.Propose.
2015-08-28 12:34:49 -07:00
Ben Darnell
05924b330a raft: Fix a nil-pointer panic in MultiNode.Propose. 2015-08-28 11:17:59 +02:00
Yicheng Qin
df83af944b Merge pull request #3384 from yichengq/fix-shadow
test: use go vet shadow feature instead of go-nyet
2015-08-27 14:27:57 -07:00
Yicheng Qin
92cd24d5bd *: fix govet shadow check failure 2015-08-27 14:15:30 -07:00
Matt Keller
32372e1d70 raft: Fixed a test misassumption
network_test.go:56: total = 59.22354ms, want > 50ms
59 is > 50, but the equation added 10 to the right side
2015-08-27 15:15:34 -04:00
Cong Ding
c09b667d57 *: fix go vet reported issues 2015-08-22 12:19:02 -05:00
Xiang Li
6b23a8131f *: test gofmt with -s and fix reported issues 2015-08-21 18:52:16 -07:00
Xiang Li
50c1db3fbf raft: downgrade the logging around snapshot to debugf
Snapshot related logging is spamming when leader trying to
sync a failed peer.
2015-08-18 15:43:53 -07:00
es-chow
cc362ccdad raft: set logger to raft so log context such as multinode groupID can be logged 2015-08-12 22:56:00 +08:00
Xiang Li
845c51fedd *: fix typos vaild->valid 2015-08-07 10:57:11 -07:00
Xiang Li
581ef05bab *: resolve proto warnings 2015-06-29 18:39:46 -07:00
Xiang Li
13f44e4b79 *: update generated proto code 2015-06-29 16:45:25 -07:00
Xiang Li
e01d53b853 Merge pull request #2979 from xiang90/fix_sendapp
raft: fix panic in send app
2015-06-29 10:49:04 -07:00
Xiang Li
b4022899eb raft: fix panic in send app
sendApp accesses the storage several times. Perviously, we
assume that the storage will not be modified during the read
opeartions. The assumption is not true since the storage can
be compacted between the read operations. If a compaction
causes a read entries error, we should not painc. Instead, we
can simply retry the sendApp logic until succeed.
2015-06-15 14:23:33 -07:00
Xiang Li
2f0169c3ab raft: fix usage section of doc
We recently added a config struct to start raft. Update
our doc accordingly.
2015-06-15 10:26:10 -07:00
Yicheng Qin
4e79abcfeb Merge pull request #2944 from yichengq/fix-2procs
pkg/testutil: ForceGosched -> WaitSchedule
2015-06-10 14:44:32 -07:00
Yicheng Qin
018fb8e6d9 pkg/testutil: ForceGosched -> WaitSchedule
ForceGosched() performs bad when GOMAXPROCS>1. When GOMAXPROCS=1, it
could promise that other goroutines run long enough
because it always yield the processor to other goroutines. But it cannot
yield processor to goroutine running on other processors. So when
GOMAXPROCS>1, the yield may finish when goroutine on the other
processor just runs for little time.

Here is a test to confirm the case:

```
package main

import (
	"fmt"
	"runtime"
	"testing"
)

func ForceGosched() {
	// possibility enough to sched up to 10 go routines.
	for i := 0; i < 10000; i++ {
		runtime.Gosched()
	}
}

var d int

func loop(c chan struct{}) {
	for {
		select {
		case <-c:
			for i := 0; i < 1000; i++ {
				fmt.Sprintf("come to time %d", i)
			}
			d++
		}
	}
}

func TestLoop(t *testing.T) {
	c := make(chan struct{}, 1)
	go loop(c)
	c <- struct{}{}
	ForceGosched()
	if d != 1 {
		t.Fatal("d is not incremented")
	}
}
```

`go test -v -race` runs well, but `GOMAXPROCS=2 go test -v -race` fails.

Change the functionality to waiting for schedule to happen.
2015-06-10 14:37:41 -07:00
Xiang Li
1279e495f0 raft: make the repeated log message under bad path debug level 2015-06-05 17:29:24 -07:00
Xiang Li
1561b85bf3 raft: drop the raft prefix in logging 2015-06-02 12:50:42 -07:00
Xiang Li
0ca6be31f8 raft: remove wrong invariant
The commit > unstable might not true for follower. The leader only need
to ensure the entry is stored on the majority of nodes to commit an
entry. So the minority of the cluster might receive commit > unstable
append request. This is normal.
2015-05-29 18:48:59 -07:00
Xiang Li
085447ed85 raft: fix raft node start bug
raft node should set initial prev hard state to empty.
Or it will not send the first hard coded state to application
until the state changes again.

This commit fixs the issue. It introduce a small overhead, that
the same tate might send to application twice when restarting.
But this is fine.
2015-05-27 13:32:04 -07:00
Xiang Li
0ad6d7e3ba Merge pull request #2853 from bdarnell/status
raft: MultiNode.Status returns nil for non-existent groups.
2015-05-20 13:07:23 -07:00