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
11,851 Commits 15 Branches 491 Tags
4267d368dfaf7b370d6fca491c8dcf451b3de606
Commit Graph

576 Commits

Author SHA1 Message Date
Anthony Romano
318caeee7e clientv3: return CompactRevision wresp when set with Canceled 2017-07-25 12:36:01 -07:00
Anthony Romano
ebcfdd1a3d integration: check Canceled is true in compacted watch response 2017-07-25 12:36:01 -07:00
mangoslicer
887df72d13 clientv3/ordering: kv order caching 2017-07-19 21:40:50 -04:00
Anthony Romano
755270fa6a integration: sync lapi server after puts in TestLeaseTimeToLive 
Linearized read to ensure the keys have committed.
 2017-07-13 14:55:43 -07:00
Xiang Li
acfde8aba0 Merge pull request #8199 from sakshamsharma/clientv3-keep-alive 
clientv3: add keep-alive to connection
 2017-07-08 16:32:35 -07:00
Gyu-Ho Lee
a57405a958 Merge pull request #8153 from gyuho/leadership-transfer 
*: expose Leadership Transfer API to clients
 2017-07-06 13:00:08 -07:00
Saksham Sharma
2a30a754e9 clientv3: add keep-alive to connection 
this makes the grpc client connection use a keep-alive.
 2017-07-06 12:55:52 -07:00
Anthony Romano
d48e59e389 Merge pull request #8201 from arthurkiller/master 
concurrency: fix typo in Serializable godoc
 2017-07-06 11:03:55 -07:00
lza@arthur
4df1970188 concurrency: fix typo in Serializable godoc 2017-07-06 12:57:55 +08:00
Gyu-Ho Lee
581a83dfd9 clientv3/*: add 'MoveLeader' method to 'Maintenance' 
Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-06-23 13:07:17 -07:00
Hui Kang
7fffd8b827 concurrency: comment the default stm isolation level is serializable snapshot 
Default stm isolation level is serializable snapshot isolation, which
is different than snapshot isolation (SI)

Signed-off-by: Hui Kang <kangh@us.ibm.com>
 2017-06-22 22:24:17 -04:00
Anthony Romano
9e574afb84 clientv3/integration: wait for leader before trying to count watches 
Fixes #8044
 2017-06-22 15:02:41 -07:00
Anthony Romano
b10ea20113 namespace: support nested txns 2017-06-21 14:33:16 -07:00
Anthony Romano
f400010028 clientv3/integration: test clientv3 nested txns 2017-06-21 14:33:15 -07:00
Anthony Romano
f8dbcd86ec clientv3: support nested Txns with OpTxn 2017-06-21 14:33:15 -07:00
Anthony Romano
8f34d0c8b6 clientv3/integration: test compare on range 2017-06-16 12:13:27 -07:00
Anthony Romano
7ff6e62c56 namespace: prefix comparison range_end 2017-06-16 12:13:27 -07:00
Anthony Romano
fcf1abd23b clientv3: compare helper functions to set range/prefix 2017-06-16 12:13:27 -07:00
Gyu-Ho Lee
5e059fd8dc *: use metadata Incoming/OutgoingContext 
Fix https://github.com/coreos/etcd/issues/7888.

Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-06-15 16:41:23 -07:00
Anthony Romano
f5b96991a1 concurrency: fetch current lock holder when creating waitlist key 
The uncontended path for a mutex would fetch the minimum
revision key on the prefix after creating its entry in
the wait list. This fetch can be rolled into the txn for
creating the wait key, eliminating a round-trip for immediately
acquiring the lock.
 2017-06-15 11:29:34 -07:00
Anthony Romano
13d9438cf9 clientv3: clarify Watch close conditions 
The "too slow" comment is rather vague. If the server closes
the watch for being too slow (it doesn't seem to any more), the
watch client should gracefully resume instead of forcing the
user to handle it.

Also removed the 'opts' comment since it wasn't being maintained.
 2017-06-15 09:34:00 -07:00
fanmin shi
762b2c625c clientv3: change watchGrpcStream Close() to close() 
private struct shouldn't have public method.
 2017-06-08 12:11:06 -07:00
fanmin shi
4dff7aaa2a clientv3: change keepAlive Close() to close() 
keepAlive is a private struct that belongs to clientv3 pkg and shouldn't expose a public Close() method.
 2017-06-08 11:53:59 -07:00
Anthony Romano
887db5a3db *: fix go tool vet -all -shadow errors 2017-06-03 21:32:36 -07:00
Anthony Romano
ea8561c35c clientv3: support unixs:// scheme 
For using TLS without giving a TLSConfig to the client.
 2017-05-31 15:51:48 -07:00
Anthony Romano
1b48d6e5df clientv3/integration: test dialing to TLS without a TLS config times out 
etcdctl was getting ctx errors from timing out trying to issue RPCs to
a TLS endpoint but without using TLS for transmission. Client should
immediately bail out with a time out error.
 2017-05-31 15:51:03 -07:00
Anthony Romano
8effbda3a7 clientv3: use Endpoints[0] to initialize grpc creds 
Dialing out without specifying TLS creds but giving https uses some
default behavior that depends on passing an endpoint with https to
Dial(), so it's not enough to completely rely on the balancer to supply
endpoints.

Fixes #8008

Also ctx-izes grpc.Dial
 2017-05-31 15:01:11 -07:00
Anthony Romano
c250e7be9e clientv3: do not launch cluster on go test without explicit -run 
There's a workaround by running -run=Test but this periodically
comes up as an issue, so have `go test` only run Test* to stem
the complaints.

Fixes #8000
 2017-05-30 12:23:12 -07:00
fanmin shi
e9f464debc integration: creation of cluster now takes maxTxnOps 2017-05-24 14:48:44 -07:00
fanmin shi
ae7ddfb483 etcdserver: add --max-txn-ops flag 
--max-txn-ops allows users to define the maximum transaction operations
for each txn request. it defaults at 128.

Fixes #7826
 2017-05-24 10:32:32 -07:00
Hitoshi Mitake
4cd5e7ebb2 Merge pull request #7809 from mitake/auth-watch 
protect watch with auth
 2017-05-20 13:23:30 +09:00
Hitoshi Mitake
939912c425 clientv3, etcdserver: support auth in Watch() 2017-05-20 11:34:45 +09:00
Anthony Romano
33c375dc44 *: fill out blank package godocs 
Mostly one-liner short descriptions, but also includes some typo fixes
and some examples.
 2017-05-18 09:41:13 -07:00
Anthony Romano
63c7e9f840 clientv3: remove duplicate documentation for Do() 2017-05-11 13:25:26 -07:00
Xiang Li
b8875515a4 Merge pull request #7890 from yudai/keep_ka_loop_running 
clientv3: Do no stop keep alive loop by server side errors
 2017-05-09 11:00:21 -07:00
Iwasaki Yudai
aa85b0cea7 clientv3: Do no stop keep alive loop by server side errors 2017-05-08 15:47:34 -07:00
Luediger Reinhard
8b081ce9b3 clientv3: check IsModify 
Fix watch_test.go TestEvent

Prior to This fix the isModify case of the table driven test was never checked.
 2017-05-05 19:39:59 +02:00
Anthony Romano
2e32d2142d Merge pull request #7869 from heyitsanthony/fix-lease-require-leader-test 
clientv3/integration: drain keepalives before waiting for leader loss
 2017-05-04 08:29:16 -07:00
Anthony Romano
9fee35b02d Merge pull request #7842 from heyitsanthony/fix-switch-race 
clientv3: don't race on upc/downc/switch endpoints in balancer
 2017-05-03 13:48:00 -07:00
Anthony Romano
f6d0dda187 clientv3/integration: drain keepalives before waiting for leader loss 
500ms keepalive delay on proxy side causes client to sometimes send
a second keepalive since it waits more than 500ms for the first response.

Fixes #7658
 2017-05-03 13:22:45 -07:00
Anthony Romano
24e85b2454 Merge pull request #7852 from heyitsanthony/revert-lease-err-ka 
Revert "Merge pull request #7732 from heyitsanthony/lease-err-ka"
 2017-05-03 11:03:17 -07:00
Anthony Romano
31d5d610fc concurrency: provide old STM functions as deprecated 
semver
 2017-05-03 02:07:01 -07:00
Anthony Romano
43e5f892f6 clientv3: don't race on upc/downc/switch endpoints in balancer 
If the balancer update notification loop starts with a downed
connection and endpoints are switched while the old connection is up,
the balancer can potentially wait forever for an up connection without
refreshing the connections to reflect the current endpoints.

Instead, fetch upc/downc together, only caring about a single transition
either from down->up or up->down for each iteration

Simple way to reproduce failures: add time.Sleep(time.Second) to the
beginning of the update notification loop.
 2017-05-02 16:43:24 -07:00
Anthony Romano
6dd8fb6f24 Revert "Merge pull request #7732 from heyitsanthony/lease-err-ka" 
This reverts commit fbbc4a4979, reversing
changes made to f254e38385.

Fixes #7851
 2017-05-02 09:36:16 -07:00
Anthony Romano
c309d745a6 Merge pull request #7819 from heyitsanthony/fix-elect-compact 
concurrency: use current revisions for election
 2017-04-27 11:01:44 -07:00
Gyu-Ho Lee
f674a1b583 clientv3/integration: test client dial with expired certs 
Signed-off-by: Gyu-Ho Lee <gyuhox@gmail.com>
 2017-04-26 12:32:46 -07:00
Anthony Romano
50f29bd661 concurrency: use current revisions for election 
Watching from the leader's ModRevision could cause live-locking on
observe retry loops when the ModRevision is less than the compacted
revision. Instead, start watching the leader from at least the store
revision of the linearized read used to detect the current leader.

Fixes #7815
 2017-04-25 20:15:50 -07:00
Anthony Romano
f3053265ae clientv3/integration: use new interfaces in lease tests 2017-04-24 23:49:44 -07:00
Anthony Romano
f224d74ed7 concurrency: use new lease interface in session 2017-04-24 23:49:44 -07:00
Anthony Romano
d5f414f69b clientv3: don't halt lease client if there is a lease error 
Fixes #7488
 2017-04-24 23:49:44 -07:00