0fa059ec49
Switches all ACL caches to 2Q.
2016-08-09 11:00:22 -07:00
1e75fa0362
Moves ACL ID generation down into the endpoint.
...
We don't want ACL replication to have this behavior so it was a
little dangerous to have in the shared helper function.
2016-08-09 00:11:00 -07:00
4a931ae12e
Adds an ACL replication status endpoint.
2016-08-04 23:30:16 -07:00
c94f1e1b83
Increases the ACL cache size to 10k.
2016-08-04 18:03:07 -07:00
3906517f70
Adds a full integrated test for ACL replication.
2016-08-04 17:59:08 -07:00
f639f49cc0
Adds remaining core replication tests.
2016-08-04 16:33:40 -07:00
defb39f8d4
Removes a TODO comment.
...
Decided we don't need to log anything about the token here. If the
token is not valid then the client will get an error about that, so
anything that can happen here is related to talking to the server in
the ACL datacenter, so not specific to the token.
2016-08-04 07:46:59 -07:00
93a7fd0561
Adds tests for the ACL reconcile algorithm.
2016-08-03 21:24:09 -07:00
796933b45b
Activates fallback to replicated ACLs.
2016-08-03 21:24:09 -07:00
9cece515c0
Adds basic ACL replication plumbing.
2016-08-03 21:24:04 -07:00
a1266e4164
Adds some supplemental tests for RPC "no leader" retries.
...
This adds some extra tests for #2175 .
2016-07-11 17:32:26 -06:00
2d8cf9ef4a
consul: change tests to not expect ErrNoLeader
2016-07-10 13:24:18 -04:00
5d0a977bdf
consul: Refactor forward to hold RPC when no leader is known
2016-07-10 13:24:06 -04:00
191876f87e
consul: Add RPCHoldTimeout as tunable hold period
2016-07-10 13:23:43 -04:00
d8fd470f4f
Merge pull request #1837 from cleung2010/obfuscate-acl-token
...
Obfuscate token for lookupACL error
2016-07-05 13:56:49 -07:00
38134f1b8c
Fix substring length on obfuscated token
2016-07-05 15:53:30 -04:00
577523fc73
consul: sort source node first if at position <= 10 in PQ's
2016-07-01 14:28:58 -07:00
e9960e6c85
Merge pull request #2137 from hashicorp/f-pq-near
...
Support "near" parameter in prepared query service block
2016-07-01 12:28:48 -07:00
ccbe86d7a8
consul: mention magic _agent token in struct comments
2016-07-01 11:50:30 -07:00
ebacaa2d67
consul: send agent source data as separate query source
2016-06-30 16:51:18 -07:00
782a081925
consul: use source parameter for near prepared queries
2016-06-30 12:11:20 -07:00
270270a33a
consul: send origin node + dc when executing prepared queries
2016-06-21 15:34:26 -07:00
925915c6ac
consul: test baked-in distance sort
2016-06-21 12:54:18 -07:00
114e57fff1
consul: use the Near field instead of PreferLocal
2016-06-21 12:39:40 -07:00
8358df599d
Merge pull request #2127 from hashicorp/b-remote-consuls-locking
...
Ensure locking of `Server`'s `remoteConsuls`.
2016-06-21 10:00:04 -07:00
f9e2900692
Merge pull request #2131 from hashicorp/b-misc-microoptimizations
...
Misc micro optimizations
2016-06-21 09:59:01 -07:00
ebdb72ce0a
Ensure locking of `Server`'s `remoteConsuls`.
2016-06-20 22:59:49 -07:00
72f7a4061c
Misc comment improvements
2016-06-20 15:29:38 -07:00
9bf6e61655
Initialize a non-empty number of Consul Datacenters. No functional change.
2016-06-20 15:26:59 -07:00
b78c95d37e
Prefer rand.Int31n() over rand.Int31().
2016-06-20 15:26:27 -07:00
e81bf2a505
Fix deadlock in Consul RTT.
...
- consul/rtt.go:388: s.getDatacentersByDistance(). Acquires RLock()
- consul/rtt.go:341: sortDatacentersByDistance() RLock still held.
- consul/rtt.go:282: getDatacenterDistance() RLock still held.
- consul/rtt.go:268: getNodesForDatacenter(). Attempts to reacquire RLock(), hangs indefinitely.
2016-06-20 14:59:54 -07:00
89fe991ab7
consul: test raw PreferLocal functionality
2016-06-20 14:53:13 -07:00
1fef85cd2e
consul: support PreferLocal in PQ's
2016-06-20 14:24:40 -07:00
7482a9207d
Chase casting types.CheckID to a string into the state_store.
...
It turns out the indexer can only use strings as arguments when
creating a query. Cast `types.CheckID` to a `string` before calling
into `memdb`.
Ideally the indexer would be smart enough to do this at compile-time,
but I need to look into how to do this without reflection and the
runtime package. For the time being statically cast `types.CheckID`
to a `string` at the call sites.
2016-06-07 16:59:02 -04:00
ff45f8c8ff
Revert "Move `structs.CheckID` to a new top-level package, `types`."
...
This reverts commit 2bbd52e3b44ff1b60939a8400264d534662d6d51.
2016-06-07 16:59:02 -04:00
a4554b945c
Move `structs.CheckID` to a new top-level package, `types`.
...
Per discussion w/ @slackpad, move this type to its own top-level package
2016-06-07 16:59:02 -04:00
cd68cd3868
Move `structs.CheckID` to a new top-level package, `types`.
...
Per discussion w/ @slackpad, move this type to its own top-level package
2016-06-07 16:59:02 -04:00
0857e93d0b
Float a type balloon. Some strings are square pegs in round holes.
...
This experiment was brought about because of variable naming
confusion where name and checkIDs were interchanged. Gave CheckID
an Qualified Type Name and chased downstream changes.
2016-06-07 16:59:02 -04:00
ffcba3df58
Merge pull request #2028 from hashicorp/f-atomic-kv
...
Adds support for atomic transactions spanning multiple KV entries.
2016-05-15 13:46:05 -07:00
3756fb23a6
Remove unused peers variable from setupRaft().
2016-05-15 06:40:46 -07:00
a11f32a1da
Adds a get-tree verb to KV transaction operations.
2016-05-13 16:57:39 -07:00
0f94a7a326
Switches GETs to a filtering model for ACLs.
2016-05-13 15:58:55 -07:00
5fd99b13ef
Removes null results for deletes, and preps for more than one result from an operation.
2016-05-13 01:47:55 -07:00
2649a6336e
Adds a read-only optimized path for transactions.
2016-05-13 00:34:05 -07:00
0c34ed078c
Adds a comment for the txnKVS() function.
2016-05-12 16:11:26 -07:00
88b1c7d054
Makes get fail a transaction if the key doesn't exist.
2016-05-11 14:18:31 -07:00
3d35acaa90
De-nests the KV output structure (removes DirEnt member).
2016-05-11 13:48:03 -07:00
04a13ec3d7
Switches to "KV" instead of "KV" for the KV operations.
2016-05-11 10:58:27 -07:00
dc662f7e35
Refactors TxnRequest/TxnResponse into a form that will allow non-KV ops.
...
This isn't needed/used yet, but it's a good hook to get in there so we
can add more atomic operations in the future. The Go API hides this detail
so that feels like a KV-specific API. The implications on the REST API are
pretty minimal.
2016-05-11 01:39:10 -07:00
d980cbcd9d
Moves txn code into a new endpoint, not specific to KV.
2016-05-10 21:58:02 -07:00
James Phillips