summaryrefslogtreecommitdiff
path: root/cpp/src/qpid/cluster/Connection.h
Commit message (Collapse)AuthorAgeFilesLines
* Fix bug in cluster with authentication: nodes exit with "unauthorized-access"Alan Conway2010-07-201-0/+2
| | | | | | | | | | Adding a node to a cluster on which authentication is enabled and on which there are existing connections authenticated with mechanisms other than anonymous, may result in nodes exiting the cluster with inconsistent authorisation errors. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@965979 13f79535-47bb-0310-9956-ffa450edef68
* Fix regression in r956882, sporadic failures of ↵Alan Conway2010-06-241-1/+1
| | | | | | client_test.cpp:testBadClientData git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@957640 13f79535-47bb-0310-9956-ffa450edef68
* Fix cluster broker crashes when management is active.Alan Conway2010-06-221-4/+10
| | | | | | | | | | | | Cluser brokers were exiting with errors "modified cluster state outside cluster context" and "confirmed < (50+0) but only sent < (49+0)" Fix was to: - delay completion of incoming update till update connection closes. - delay addding new connections to managment until connection is announced. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@956882 13f79535-47bb-0310-9956-ffa450edef68
* Bug 603835 - cluster_tests.test_management failing.Alan Conway2010-06-161-1/+4
| | | | | | | | | | | Clean up connections causing extra connection objects in the mangement agent map. - update connection was not being closed. - connections belonging to members that left the cluster were not fully cleaned up Also fixed test errors making failover_soak fail sporadically. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@955370 13f79535-47bb-0310-9956-ffa450edef68
* Cluster handle connection-negotiation phase in local broker.Alan Conway2010-06-081-14/+6
| | | | | | | | | | | | | | | | | The connection negotiation phase up to the "open" or "open-ok" frame establishes whether/what encryption to use for the rest of the connection. With this patch a cluster broker completes the initial negotiation with its local clients and only then begins multicasting to other brokers. The local broker decrypts if necessary and multicasts in the clear. This replaces a problematic locking scheme that was formerly in place which caused deadlocks. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@952692 13f79535-47bb-0310-9956-ffa450edef68
* Cluster + SecurityMichael Goulish2010-05-141-1/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ----------------------------------- * initial observation of a problem was a 2% failure rate in perftests of 20,000 messages against a cluster with security enabled. Problem was occasional receit of encrypted frames before the security codec had been enabled. This is fixed with locking in cluster code (no new locks in broker code) and a callback that is fired by broker::ConnectionHandler::Handler to tell the cluster code when the opening handshake has finished. This was never a problem in the non-clustered broker before because everything happened in a single thread. * the brokers that "shadow" the connection must not have null authenticators rather than real ones, so that they go through all the motions but don't do anythig. Only the directly-connected broker can perform the security handshake. * once the directly-connected broker receives the real user ID from its callback, it mcasts that ID to all other brokers. Otherwise the shadowing brokers will al think that the user ID is "anonymous". Check this by doing a substantial perftest, and using qpid-stat -c localhost:PORT to confirm that the brokers all have the same userID for the same connection. * the user ID, negotiated during the Sasl security startup, is communicated from the directly connected broker to all other cluster brokers. * If security is *not* being used, then this code should *not* tell the brokers anything about the userID -- or it will step on the value that is being set by other code pathways. * test program at cpp/src/tests/cluster_authentication_soak is not yet fully automated -- run it with something like "sudo ./cluster_authentication_soak 500" git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@944158 13f79535-47bb-0310-9956-ffa450edef68
* Cluster: remove un-necessary code for config-seq.Alan Conway2010-03-311-2/+1
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@929616 13f79535-47bb-0310-9956-ffa450edef68
* QPID-2412: Support for EXTERNAL mechanism on client-authenticated SSL ↵Gordon Sim2010-03-051-8/+9
| | | | | | | | | | | | connections. On SSL connection where the clients certificate is authenticated (requires the --ssl-require-client-authentication option at present), the clients identity will be taken from that certificate (it will be the CN with any DCs present appended as the domain, e.g. CN=bob,DC=acme,DC=com would result in an identity of bob@acme.com). This will enable the EXTERNAL mechanism when cyrus sasl is in use. The client can still negotiate their desired mechanism. There is a new option on the ssl module (--ssl-sasl-no-dict) that allows the options on ssl connections to be restricted to those that are not vulnerable to dictionary attacks (EXTERNAL being the primary example). git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@919487 13f79535-47bb-0310-9956-ffa450edef68
* Consistent connection names across a cluster.Alan Conway2010-02-051-8/+19
| | | | | | | | | - use the same host:port for connections and their shadows. - add shadow property to managment connection to identify shadows. - updated qpid-stat and qpid-cluster to filter on shadow property. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@907123 13f79535-47bb-0310-9956-ffa450edef68
* Synchronize management agent lists during cluster update.Alan Conway2010-02-051-0/+1
| | | | | | | | | - replicate management agent lists during cluster update. - suppress management agent output during update. - on join all members force full output at next periodic processing. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@907030 13f79535-47bb-0310-9956-ffa450edef68
* Cluster: debug snapshots of queue depth at broker join, help find ↵Alan Conway2010-02-021-1/+1
| | | | | | inconsistencies. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@905674 13f79535-47bb-0310-9956-ffa450edef68
* QPID-2357 Broker boot sequence doesn't synchronize when clustered - patch ↵Alan Conway2010-01-281-0/+1
| | | | | | from John Dunning git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@904270 13f79535-47bb-0310-9956-ffa450edef68
* In clustered broker: move construction of broker::Connections to the cluster ↵Alan Conway2010-01-271-6/+32
| | | | | | | | | | dispatch thread. Constructing a connection can involve sending management information so needs to be in the cluster dispatch context. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@903864 13f79535-47bb-0310-9956-ffa450edef68
* Removed dead code - cluster::Connection::left()Alan Conway2010-01-191-3/+0
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@900860 13f79535-47bb-0310-9956-ffa450edef68
* Removed dead code: OutputTask::hasOutput and Queue::checkForMessages.Alan Conway2010-01-141-1/+0
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@899356 13f79535-47bb-0310-9956-ffa450edef68
* Fix broker crash "confirmed N but only sent M" with managed agents running.Alan Conway2010-01-111-0/+1
| | | | | | | | | | | | | | The broker's ManagementAgent caches schemas from managed agents. This cache was not being replicated to new cluster members. If an agent such as sesame was running and connected to a newly-joined broker, that broker could send schema request messages which were not sent by other brokers that had the schema in cache. This resulted in the other brokers exiting with a "confirmed N but only sent M" message. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@897955 13f79535-47bb-0310-9956-ffa450edef68
* Added config-seq counter to track config changes since cluster init.Alan Conway2010-01-061-1/+3
| | | | | | | | Config-seq is recorded persitently to help identify best store when recovering from total failure. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@896538 13f79535-47bb-0310-9956-ffa450edef68
* QPID-2296: Cluster errors when using acquire-mode-not-acquiredAlan Conway2009-12-221-1/+1
| | | | | | | Replicate consumer's queue position to new cluster nodes. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@893175 13f79535-47bb-0310-9956-ffa450edef68
* QPID-1899: Applied patch from Ken Giusti to tie in SASL enctryption to the ↵Gordon Sim2009-11-091-3/+6
| | | | | | handling of the --require-encrypted option git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@834108 13f79535-47bb-0310-9956-ffa450edef68
* Fix race condition in cluster error handling.Alan Conway2009-07-311-1/+1
| | | | | | | | | If different errors occured almost simultaneously on two different nodes in a cluster, there was a race condition that could cause the cluster to hang. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@799687 13f79535-47bb-0310-9956-ffa450edef68
* Update queue listeners in the correct order.Alan Conway2009-07-161-8/+14
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@794736 13f79535-47bb-0310-9956-ffa450edef68
* Add directory to #includeAlan Conway2009-07-141-4/+4
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@793909 13f79535-47bb-0310-9956-ffa450edef68
* Fix members joining cluster while cluster is handling client errors.Alan Conway2009-07-011-0/+2
| | | | | | | | | | | Previously cluster members could abort if a new member joins while existing members are handling a client error. Now if an update offer arrives while an error is in progress, the offering broker retracts the offer and the newcomer must try again. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@790163 13f79535-47bb-0310-9956-ffa450edef68
* Fix cluster race condition with connections closed by broker while in use.Alan Conway2009-06-301-1/+1
| | | | | | | | | | | | | | | | If a client is using a connection that is closed at the broker end because of an error, there is a race condition that allows the connection to be incorrectly re-created on replica brokers which can cause those brokers to exit with an error that does not occur on the directly connected broker. The fix: explicitly announce new connections, shadow connections are no longer implicitly created on first use. Make error-check a cluster control so it can be handled independently of the lifecycle of the connection where an error initially occured. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@789947 13f79535-47bb-0310-9956-ffa450edef68
* Make error-check a cluster-connection control rather than a cluster control.Alan Conway2009-06-181-0/+1
| | | | | | | | | Fixes bug if an error occurs during update. As cluster controls, error-checks were being processed out of sequence with the connection data they referred to. Making them connection controls ensures they are processed in the proper order. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@786294 13f79535-47bb-0310-9956-ffa450edef68
* Handle invalid AMPQ data to a cluster by closing the offending connection.Alan Conway2009-06-171-1/+1
| | | | | | | Prior to this fix, invalid data shut down the whole cluster. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@785788 13f79535-47bb-0310-9956-ffa450edef68
* Performance improvements in AggregateOutput and SemanticState.Alan Conway2009-06-161-1/+3
| | | | | | | | | | | | | | Replaced AggregateOutput hierarchy with a flat list per connection holding only the OutputTasks that are potentially active. Tasks are droped from the list as soon as they return false, and added back when they may have output. Inlined frequently-used SequenceNumber functions. Replace std::list in QueueListeners with std::vector. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@785408 13f79535-47bb-0310-9956-ffa450edef68
* Further fix to new cluster member state transfer to fix a case where unacked ↵Gordon Sim2009-06-051-0/+1
| | | | | | messages on ring policy queue cause inconsistencies in queue state between nodes. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@782075 13f79535-47bb-0310-9956-ffa450edef68
* Improved doOutput algorithm.Alan Conway2009-05-261-3/+4
| | | | | | | | | Simpler & more robust algorithm based on message count rather than byte size. Self-tuning, removes 2 hard-to-explain cluster options. Similar or marginally better performance. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@778896 13f79535-47bb-0310-9956-ffa450edef68
* PollableQueue optimization - replace deque with vector.Alan Conway2009-05-251-2/+0
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@778464 13f79535-47bb-0310-9956-ffa450edef68
* Improved & simplified cluster output algorithm.Alan Conway2009-04-291-2/+0
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@769914 13f79535-47bb-0310-9956-ffa450edef68
* Fix regression: close shadow connections belonging to brokers that have left ↵Alan Conway2009-04-141-1/+2
| | | | | | the cluster. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@764783 13f79535-47bb-0310-9956-ffa450edef68
* Fix issues when cluster is run with persistence enabled.Alan Conway2009-04-111-6/+10
| | | | | | | | | | - Handle partial failures (e.g. due to disk error): failing brokers shut down, others continue. - Enable persistence in cluster tests. - Correct message status in DeliveryRecord updates. - Remove qpid.update queue when update complete - avoid it becoming persistent git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@764204 13f79535-47bb-0310-9956-ffa450edef68
* Fix cluster TTL: replicte expiry information to newcomers.Alan Conway2009-03-091-1/+2
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@751760 13f79535-47bb-0310-9956-ffa450edef68
* Fixed race conditions in cluster.Alan Conway2009-03-081-8/+2
| | | | | | | | Execute all cluster logic in frameDeliverQueue thread, decoding only in eventDeliverQueue thread. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@751557 13f79535-47bb-0310-9956-ffa450edef68
* cluster: fix delivery-property.exchange-name set on updated messages.Alan Conway2009-03-051-1/+1
| | | | | | | | | Logging improvements, useful for debugging: - qpid/SessionState.cpp: show frame bodies with command IDs. - assign cluster-wide id number to each Event. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@750456 13f79535-47bb-0310-9956-ffa450edef68
* Replicate connection decoder fragments to new members.Alan Conway2009-03-021-7/+13
| | | | | | | | | | | | Refactoring: - Merge Decoder into ConnectionMap. - Process cluster controls in event queue thread. - Use counter not pointer for connection ID, avoid re-use. - Do all processing in event queue thread to avoid races (temporary pending performance measurements) git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@749473 13f79535-47bb-0310-9956-ffa450edef68
* Fixed issue with producer flow control in a cluster.Alan Conway2009-02-241-0/+7
| | | | | | | | | | | | | | | | | | Producer flow control uses a Timer and other clock-based calculations to send flow control commands. These commands are not predictably ordered from the clusters point of view. Added getClusterOrderProxy() to SessionState. In a cluster it returns a proxy that defers sending a command to the client until it is multicast to the cluster. In a stand alone broker it is just the normal proxy. Updated producer flow control to use this proxy. Cluster flow control is turned off in shadow connections. Only the directly connected node does flow control calculations and multicasts the commands to send. All nodes sending of the commands thru SessionState to ensure consistent session state (e.g. command numbering.) git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@747528 13f79535-47bb-0310-9956-ffa450edef68
* Cluster security support:Alan Conway2009-02-121-1/+1
| | | | | | | | - Set correct user ID on update connections. - Allow configuration of user, pass and mechanism used for update connections. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@743839 13f79535-47bb-0310-9956-ffa450edef68
* Fix memory leak in cluster code.Alan Conway2009-02-111-1/+1
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@743346 13f79535-47bb-0310-9956-ffa450edef68
* Fix cluster flow control bug: hang with large messages (>frame-max) and low ↵Alan Conway2009-02-101-1/+2
| | | | | | --cluster-read-max. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@743114 13f79535-47bb-0310-9956-ffa450edef68
* Cluster support for message time-to-live.Alan Conway2009-02-091-1/+1
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@742774 13f79535-47bb-0310-9956-ffa450edef68
* Fix for race conditions in cluster join.Alan Conway2009-02-031-3/+3
| | | | | | | | | - ConnectionDecoder: separated from Connection. - cluster/PollableQueue: stop processing frames if PollableQueue is stopped. - move state checks in event-queue handler to frame-queue handler. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@740459 13f79535-47bb-0310-9956-ffa450edef68
* Cluster rename: dump -> update, newbie -> joinerAlan Conway2009-01-271-4/+4
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@737971 13f79535-47bb-0310-9956-ffa450edef68
* cluster: Add sequence number to events & framesAlan Conway2009-01-271-3/+2
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@737968 13f79535-47bb-0310-9956-ffa450edef68
* Use special management ids for objects used in state transfer to new ↵Gordon Sim2009-01-231-0/+3
| | | | | | | | | members. This prevents the ids getting out of sync across the cluster and allows management methods to be used reliably. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@737203 13f79535-47bb-0310-9956-ffa450edef68
* QPID-1567: More changes to make clustering and federation work togetherGordon Sim2009-01-221-1/+2
| | | | | | | | | | | * replicate outgoing link traffic to all nodes * coordinate amongst nodes so that only one node actually maintains active links with the others able to take over if that node fails git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@736841 13f79535-47bb-0310-9956-ffa450edef68
* cluster: Pipeline decoding. About 10% improvement in latency and throughput.Alan Conway2009-01-211-2/+8
| | | | git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@736409 13f79535-47bb-0310-9956-ffa450edef68
* Fix --cluster-mast-max: errors in last commit.Alan Conway2009-01-091-2/+0
| | | | | | | Work around problems with CPG flow control. git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@733051 13f79535-47bb-0310-9956-ffa450edef68
* cluster: refactor multicast concerns into separate Multicaster class with ↵Alan Conway2008-12-111-1/+0
| | | | | | separate locking. git-svn-id: https://svn.apache.org/repos/asf/incubator/qpid/trunk/qpid@725853 13f79535-47bb-0310-9956-ffa450edef68
View Lorry Controller Status