e179225f28
grpcproxy: tear down watch when client context is done
...
If client closes but all watch streams are not canceled, the outstanding
watch will wait until it is canceled, causing watch server to potentially
wait forever to close.
Fixes #7102
2017-01-04 16:23:27 -08:00
154f268031
Merge pull request #7001 from heyitsanthony/etcdctl-doc
...
etcdctl: tighten up output, reorganize README.md
2017-01-04 13:44:49 -08:00
10d3b81c39
Merge pull request #7093 from gyuho/member
...
etcdserver: expose ErrMemberNotEnoughStarted
2017-01-04 12:09:29 -08:00
f9f691ef1f
auth: use quorum get for GetUser/GetRole for mutable operations
...
GetUser would not propagate to the minority node, causing TestCtlV2GetRoleUser to
run CreateUser instead of UpdateUser. Instead, use quorum get to fetch the
current state of auth.
Fixes #7069
2017-01-04 11:55:07 -08:00
729dcd51ce
Merge pull request #7090 from vimalk78/fix-comactor-resume-leadr-change#7040
...
etcdserver: resume compactor only if leader
2017-01-04 10:47:44 -08:00
559a82f66e
Merge pull request #7097 from heyitsanthony/benchmark-verbose
...
benchmark: enable grpc error logging on stderr
2017-01-04 10:32:07 -08:00
40ae83beab
Merge pull request #7099 from overvenus/patch-1
...
docs: fix recovery example in recovery.md
2017-01-04 10:16:48 -08:00
37501e2a5d
Merge pull request #7092 from xiang90/fix_raft
...
raft: use status to test node stop
2017-01-04 09:13:11 -08:00
7aeddf6cd7
docs: fix recovery example in recovery.md
2017-01-04 19:41:15 +08:00
d0f301adb7
etcd-runner:add flags in watcher for hardcoded values
2017-01-04 15:17:53 +05:30
b8444d4d35
benchmark: enable grpc error logging on stderr
...
Lets you see connection errors (e.g., if tls is misconfigured)
2017-01-04 00:26:43 -08:00
5fac6b8d15
etcdserver: resume compactor only if leader
2017-01-04 05:01:14 +05:30
2b5f9e1c6b
etcdserver: expose ErrNotEnoughStartedMembers
...
Fix https://github.com/coreos/etcd/issues/7072 .
2017-01-03 15:23:06 -08:00
fc8cd44c72
raft: use status to test node stop
...
n.Tick() is async. It can be racy when running with n.Stop().
n.Status() is sync and has a feedback mechnism internally. So there wont be
any race between n.Status() and n.Stop() call.
2017-01-03 15:18:48 -08:00
61064a7be3
Merge pull request #7085 from gyuho/raft-example-snapshot
...
raftexample: load snapshot when opening WAL
2017-01-03 10:34:13 -08:00
5cb6dd268b
etcd-runner: make command compliant
2017-01-03 14:43:58 +05:30
0af1679b61
raftexample: load snapshot when opening WAL
...
Fix https://github.com/coreos/etcd/issues/7056 .
Previously we don't load snapshot when replaying WAL.
2016-12-30 17:28:57 -08:00
24601ca24b
Merge pull request #7084 from heyitsanthony/watch-proxy-leak
...
integration: wait for watch proxy to finish on client close
2016-12-30 12:51:31 -08:00
75441390b6
integration: defer clus.Terminate in watch tests
...
Common pattern was defer cancel(), but clus.Terminate() at the end of
the test. This appears to lead to a deadlock that is only released
once the context times out, causing inflated test times.
2016-12-30 12:34:04 -08:00
9b5eb1ae5a
grpcproxy, etcdmain, integration: return done channel with WatchServer
...
Makes it possible to synchronously close the watch server.
Fixes #7078
2016-12-30 12:09:48 -08:00
29e14dde0c
Merge pull request #7081 from gyuho/timeout-rafthttp
...
rafthttp: bump up timeout in pipeline test
2016-12-30 10:14:12 -08:00
cbb6ede69d
Merge pull request #7067 from fanminshi/rework_coverage_unit_integration
...
coverage: rework coverage for unit and integration tests
2016-12-30 10:13:07 -08:00
d25f9feb19
rafthttp: bump up timeout in pipeline test
...
Fix https://github.com/coreos/etcd/issues/6283 .
The timeout is too short. It could take more than 10ms
to send when the buffer gets full after 'pipelineBufSize' of
requests.
2016-12-30 09:46:16 -08:00
74e7614759
testutil: whitelist thread created by go cover
2016-12-29 17:19:27 -08:00
d9a3472894
coverage: rework code coverage for unit and integration tests
2016-12-29 17:19:03 -08:00
0dce29ae57
Merge pull request #7077 from fanminshi/consistent_naming
...
etcdserver: consistent naming in raftReadyHandler
2016-12-29 14:37:46 -08:00
8242049a33
Merge pull request #7076 from fanminshi/fix_e2e_test
...
e2e: unset ETCDCTL_API env var before running e2e tests
2016-12-29 14:37:25 -08:00
734dd75565
Merge pull request #7075 from gyuho/version-pull
...
e2e: poll '/version' in release upgrade tests
2016-12-29 11:29:45 -08:00
2a1bae0c2a
etcdserver: consistent naming in raftReadyHandler
2016-12-29 11:27:16 -08:00
b741452d03
e2e: unset ETCDCTL_API env var before running u2e tests
...
existing ETCDCTL_API env var causes e2e to fail some of its tests. ETCDCTL_API should not be set before e2e tests start.
the tests themselves should set ETCDCTL_API properly.
2016-12-29 11:21:15 -08:00
4e1010c1b9
e2e: poll '/version' in release upgrade tests
...
Fix https://github.com/coreos/etcd/issues/7065 .
2016-12-29 10:52:40 -08:00
67c75606db
Merge pull request #7070 from heyitsanthony/fix-lease-race
...
lease: use atomics for accessing lease expiry
2016-12-28 16:30:08 -08:00
b5cde6b321
lease: use atomics for accessing lease expiry
...
Demote was racing on expiry when LeaseTimeToLive called Remaining. Replace
with intrinsics since the ordering isn't important, but torn writes are
bad.
2016-12-28 15:44:14 -08:00
1643ed5667
Merge pull request #7071 from heyitsanthony/bump-integration-timeout
...
test: bump grpcproxy pass timeout to 15m
2016-12-28 15:41:00 -08:00
f876ccb055
test: bump grpcproxy pass timeout to 15m
...
integration tests have a 15m timeout elsewhere. The lease stress tests
seem to have pushed the running time over 10m on proxy CI, causing
failures from timeout.
2016-12-28 14:56:57 -08:00
12d930b40f
Merge pull request #7068 from heyitsanthony/fix-v2-health
...
v2http: submit QGET in health endpoint if no progress
2016-12-28 14:30:31 -08:00
3519a9784e
Merge pull request #7039 from mitake/benchmark-dialtimeout
...
benchmark: a new option for configuring dial timeout
2016-12-28 13:12:11 -08:00
9690220cd1
Merge pull request #7064 from heyitsanthony/fix-health-perms
...
etcdctl: treat permission denied as healthy endpoint
2016-12-28 13:04:55 -08:00
e2463569e7
v2http: submit QGET in health endpoint if no progress
...
Removing the periodic SYNC calls broke the health endpoint since the
raft index stops updating. Instead, don't bother monitoring the
raft index; issue a QGET directly to get a consensus response.
Fixes #6985
2016-12-28 12:20:56 -08:00
46062efa78
e2e: test cluster-health
2016-12-28 12:20:55 -08:00
e63059ec31
Merge pull request #7030 from crandles/grpc-histograms
...
etcdmain: add '--metrics' option
2016-12-28 12:03:53 -08:00
36b2d3f5eb
etcdmain: add --metrics flag for exposing histogram metrics
...
this adds a new flag, --metrics, that can be used to enable extensive (histogram) metrics.
Fixes #7024
2016-12-28 13:04:52 -05:00
00e00f16bb
ctlv3: consider permission denied error to be healthy for endpoints
...
Relaxes the permission expectations for endpoint health by noting:
* permission denial on linearized reads is always through consensus
* endpoint health means consensus with the cluster through the endpoint
So, there's no need to require permission on a health check key in order
to know whether the endpoint is healthy.
Fixes #7057
2016-12-28 09:13:27 -08:00
b940e0d514
Merge pull request #7042 from petermattis/pmattis/resume-after-heartbeat-resp
...
raft: resume paused followers on receipt of MsgHeartbeatResp
2016-12-27 21:15:53 -08:00
a662ddefbb
benchmark: a new option for configuring dial timeout
...
Current benchmark doesn't have an option for configuring dial timeout
of gRPC. This commit adds --dial-timeout for the purpose. It is useful
for stopping long sticking benchmarks.
2016-12-28 14:07:43 +09:00
407afc69ed
e2e: check etcdctl endpoint health is healthy if denied permission to key
2016-12-27 14:49:52 -08:00
c00084812c
Merge pull request #7054 from gyuho/err
...
etcd-tester: remove unused err var from maxRev
2016-12-27 12:36:48 -08:00
db8b15bf8f
etcd-tester: remove unused err var from maxRev
2016-12-27 12:16:43 -08:00
89b18ff1af
Merge pull request #7015 from fanminshi/fix_lease_expired_too_soon
...
lease: force leader to apply its pending committed index for lease op…
2016-12-27 11:26:15 -08:00
2faf72f47c
etcdserver: rework update committed index logic
2016-12-27 10:11:40 -08:00
Anthony Romano