Mirroristas/etcd
2
0
Fork 0
mirror of https://github.com/etcd-io/etcd.git synced 2024-09-27 06:25:44 +00:00
Code Issues Packages Projects Releases Wiki Activity
10,927 Commits 15 Branches 491 Tags 148 MiB
Commit Graph

10927 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Anthony Romano
d1a9ccb2b9 clientv3: wait for Get goroutine in TestDialCancel 2017-03-21 16:43:39 -07:00
Gyu-Ho Lee
6511171725 Merge pull request #7561 from gyuho/travis 
travis: always 'go get -u' in 'before_install'
 2017-03-21 14:16:55 -07:00
Gyu-Ho Lee
e127214c6c travis: always 'go get -u' in 'before_install' 
See https://github.com/dominikh/go-tools/issues/76#issuecomment-288189194.

Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-03-21 12:21:55 -07:00
Gyu-Ho Lee
327e255695 Merge pull request #7546 from gyuho/fix-blocking-etcd-process 
*: fix blocking etcd process
 2017-03-21 12:04:53 -07:00
Xiang Li
7698a2a546 Merge pull request #7553 from xiang90/fix_defrag 
backend: add FillPercent option
 2017-03-21 11:16:17 -07:00
Gyu-Ho Lee
2d5f890091 integration: ensure 'StopNotify' on publish error 
Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-03-21 10:29:00 -07:00
Gyu-Ho Lee
17e2e762b1 etcdmain: handle StopNotify when ErrStopped aborted publish 
Fix https://github.com/coreos/etcd/issues/7512.

If a server starts and aborts due to config error,
it is possible to get stuck in ReadyNotify waits.
This adds select case to get notified on stop channel.

Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-03-21 10:22:39 -07:00
Xiang Li
cd70ea33ce Merge pull request #7552 from mitake/ordinary 
e2e, etcdserver: fix wrong usages of ordinal
 2017-03-21 09:42:27 -07:00
Xiang
95870a21eb backend: add FillPercent option 2017-03-21 08:06:03 -07:00
Hitoshi Mitake
5594f695bc e2e, etcdserver: fix wrong usages of ordinal 
They must be "ordinary".
 2017-03-21 23:50:16 +09:00
Anthony Romano
004c1388fb Merge pull request #7541 from heyitsanthony/remove-legacy-range 
etcdserver: remove legacy range/txn
 2017-03-20 19:22:39 -07:00
Gyu-Ho Lee
27550b229a Merge pull request #7545 from gyuho/go1.7-go1.8 
*: use 'io.Seek*' for go1.7+
 2017-03-20 16:31:21 -07:00
Anthony Romano
effa6e0767 etcdserver: remove legacy range/txn 
Needed for 3.0->3.1. Not needed for 3.1->3.2
 2017-03-20 15:17:17 -07:00
Gyu-Ho Lee
aca2abd8fe *: use 'io.Seek*' for go1.7+ 
For https://github.com/coreos/etcd/issues/6174.

Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-03-20 15:15:24 -07:00
Anthony Romano
3a1368d4d2 Merge pull request #7543 from heyitsanthony/fix-timeseries 
*: fix gosimple warning for fmt.Sprintf("%s")
 2017-03-20 15:02:13 -07:00
Anthony Romano
ae7b4ee8ed *: fix gosimple warning for fmt.Sprintf("%s") 2017-03-20 13:26:39 -07:00
Anthony Romano
53ca03b655 Merge pull request #7539 from heyitsanthony/fix-protobuf-help 
ctlv3: have "protobuf" in output help string instead of "proto"
 2017-03-20 11:26:13 -07:00
Anthony Romano
cfdad38f4e Merge pull request #7531 from heyitsanthony/fix-mem-remove-again 
e2e: force endpoint for member removal
 2017-03-20 09:52:30 -07:00
Anthony Romano
432c19de61 ctlv3: have "protobuf" in output help string instead of "proto" 
Fixes #7538
 2017-03-20 09:40:21 -07:00
fanmin shi
fba87558a6 Merge pull request #7529 from fanminshi/fix_closing_embedded_error 
embed: don't return error when closing on embed etcd
 2017-03-17 17:02:22 -07:00
Anthony Romano
21ac657e67 e2e: force endpoint for member removal 
e2e tests use different invocations of etcdctl, so the endpoint used to get
the member list will not necessarily be the same to make the remove call.
Instead, select an endpoint that is not being remove, and connect with that.
 2017-03-17 16:24:54 -07:00
fanmin shi
f292a4c953 embed: don't return error when closing on embed etcd 
FIXES #7019
 2017-03-17 13:41:05 -07:00
Xiang Li
5015480e0c Merge pull request #7517 from jsok/7516-discovery-flags 
embed: Delay setting initial cluster
 2017-03-16 09:17:42 -07:00
Anthony Romano
79f4c196b8 Merge pull request #7518 from heyitsanthony/filepath 
*: replace path.Join on files with filepath.Join
 2017-03-16 08:59:49 -07:00
Anthony Romano
2f1542c06d *: use filepath.Join for files 2017-03-16 07:46:06 -07:00
Jonathan Sokolowski
1a91ed0e99 embed: Clear default initial cluster 
NewConfig() should sets initial cluster from name but we should clear it
in the event that another discovery option has been specified.

Fixes #7516
 2017-03-16 13:59:06 +11:00
Anthony Romano
d78b03fb27 Merge pull request #7515 from tessr/master 
wal: use path/filepath instead of path
 2017-03-15 19:25:08 -07:00
Tess Rinearson
39c733ebe7 wal: use path/filepath instead of path 
Use the path/filepath package instead of the path package. The
path package assumes slash-separated paths, which doesn't work
on Windows. But path/filepath manipulates filename paths in a way
that's compatible across OSes.
 2017-03-15 17:30:23 -07:00
Gyu-Ho Lee
5856c8bce9 Merge pull request #7513 from gyuho/raft-applied-term 
etcdserver: remove possibly compacted entry look-up
 2017-03-15 13:35:36 -07:00
Gyu-Ho Lee
80c10e150f etcdserver: remove possibly compacted entry look-up 
Fix https://github.com/coreos/etcd/issues/7470.

This patch removes unnecessary term look-up in
'createMergedSnapshotMessage', which can trigger panic
if raft entry at etcdProgress.appliedi got compacted
by subsequent 'MsgSnap' messages--if a follower is
being (in this case, network latency spikes), it
could receive subsequent 'MsgSnap' requests from leader.

etcd server-side 'applyAll' routine and raft's Ready
processing routine becomes asynchronous after raft
entries are persisted. And given that raft Ready routine
takes less time to finish, it is possible that second
'MsgSnap' is being handled, while the slow 'applyAll'
is still processing the first(old) 'MsgSnap'. Then raft
Ready routine can compact the log entries at future
index to 'applyAll'. That is how 'createMergedSnapshotMessage'
tried to look up raft term with outdated etcdProgress.appliedi.

Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-03-15 12:46:56 -07:00
fanmin shi
902c676cdb Merge pull request #7397 from fanminshi/fix_SetEndpoints 
clientv3/balancer: update eps if pinAddr is not included in updateAddrs
 2017-03-15 12:16:15 -07:00
fanmin shi
a23609efe6 clientv3: update eps if pinAddr is not included in updateAddrs 
FIXES #7392
 2017-03-15 11:03:25 -07:00
Anthony Romano
a2a6b693f1 Merge pull request #7511 from heyitsanthony/fix-v3client-embed 
v3client: fix doc to use e.Server
 2017-03-15 10:28:41 -07:00
Anthony Romano
dea2516177 v3client: fix doc to use e.Server 
Was passing embed.Etcd instead of etcdserver.EtcdServer.
 2017-03-15 09:17:17 -07:00
Anthony Romano
8f83d11724 Merge pull request #7499 from heyitsanthony/fix-etcdctl-add-member-env 
ctlv3: ensure synced member list before printing env vars on member add
 2017-03-15 08:59:04 -07:00
Anthony Romano
27960911af Merge pull request #7500 from heyitsanthony/fix-balancer-test-leak 
clientv3: synchronize on goroutines in TestBalancerDoNotBlockOnClose
 2017-03-15 08:58:03 -07:00
Anthony Romano
7a6b61cd6f Merge pull request #7504 from heyitsanthony/fix-watch-wait 
clientv3: close open watch channel if substream is closing on reconnect
 2017-03-15 08:57:14 -07:00
Xiang Li
df839f3b7f Merge pull request #7497 from xiang90/fix_candidate 
etcdserver: candidate should wait for applying all configuration changes
 2017-03-14 20:10:02 -07:00
Anthony Romano
3e86779ad5 ctlv3: ensure synced member list before printing env vars on member add 
In cases of multiple endpoints, it's possible member add would get a its
member list from a member that has not yet recognized the membership
update. Instead, confirm that the member list response is from the
member that acked the member add or from a member that has synced
with the cluster following the member add.

Fixes #7498
 2017-03-14 20:01:44 -07:00
Anthony Romano
b36734f1d3 clientv3: synchronize on goroutines in TestBalancerDoNotBlockOnClose 
Was leaking dialers.
 2017-03-14 19:53:33 -07:00
Anthony Romano
18a813a9fe Merge pull request #7496 from heyitsanthony/v3client-doc 
v3client: add example and godoc New
 2017-03-14 19:50:01 -07:00
Anthony Romano
a087325452 clientv3: close open watch channel if substream is closing on reconnect 
If substream is closing but outc is still open while reconnecting, then outc
would only be closed once the watch client would connect or once the watch
client is closed. This was leading to deadlocks in the proxy tests. Instead,
close immediately if the context is canceled.

Fixes #7503
 2017-03-14 17:25:18 -07:00
Xiang
7f0733cf46 etcdserver: candidate should wait for applying all configuration changes 2017-03-14 17:20:20 -07:00
Gyu-Ho Lee
eed4a3f035 Merge pull request #7502 from gyuho/scripts 
test: mask go1.8 gosimple warnings
 2017-03-14 17:00:01 -07:00
Gyu-Ho Lee
a9588952a0 test: mask go1.8 gosimple warnings 2017-03-14 15:10:32 -07:00
fanmin shi
ace3a217b0 Merge pull request #7483 from fanminshi/add_tests_to_mutex 
integration: add TestMutexWaitsOnCurrentHolder test
 2017-03-14 13:01:47 -07:00
fanmin shi
276039e835 integration: add TestMutexWaitsOnCurrentHolder test 
TestMutexWaitsOnCurrentHolder ensures a series of waiters
obtain lock only after the previous lock requests are gone.
 2017-03-14 11:00:07 -07:00
Anthony Romano
01d1a579bc v3client: add example and godoc New 2017-03-14 10:50:41 -07:00
Anthony Romano
781196fa87 Merge pull request #7495 from heyitsanthony/more-cov 
test: add coverage for more packages
 2017-03-14 09:31:01 -07:00
Anthony Romano
e3218e2dd1 test: add coverage for more packages 
Was only getting coverage for packages with test files. Instead, include
packages that don't have test files as well.
 2017-03-14 01:08:07 -07:00