| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
| |
to their own git repositories
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1751566 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
| |
tools and extras
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1740289 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Removed complex and incorrect HA+TX logic, reverted to the following limitation:
You can use transactions in a HA cluster, but there are limitations on the
transactional guarantees. Transactions function normally with the *primary*
broker but replication to the backups is not coverted by the atomic guarantee.
The following situations are all safe:
- Client rolls back a transaction.
- Client successfully commits a transaction.
- Primary fails during a transaction *before* the client sends a commit.
- Transaction contains only one message.
The problem case is when all of the following occur:
- transaction contains multiple actions (enqueues or dequeues)
- primary fails between client sending commit and receiving commit-complete.
In this case it is possible that only part of the transaction was replicated to
the backups, so on fail-over partial transaction results may be visible.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1701109 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Implements the "transactional retire and settle immediately" option for
transactions as specified in AMQP 1.0 in the qpid::messaging C++ client.
NOTE: Transactions over AMQP 1.0 require proton 0.9 or greater. With older
versions, attempting a transactions over AMQP 1.0 will raise a link-detached
exception "Node not found: tx-transaction"
1. Added descriptor list to Variant with support in Encoder and PnData.
Required to support transactions, need to be able to create described lists.
Variant changes are source and binary compatible.
A Variant now has a Variant::List of descripors which can be numeric or string.
Nested descriptors are implemented by putting multiple descriptors in the list.
Other minor changes:
- Variant refactor: don't delete impl on every assignment.
- Add Variant constructors that take a string encoding.
(new constructors, not defaulted arguments, so the change is binary and source compatible.)
- Growable buffer support for Encoder.
- Printing described Variant prints descriptors in form @descriptor value
2. Added transaction support to AMQP 1.0 client code
Added messaging/amqp/Transaction.h,cpp: transaction logic
- communicate with coordinator, send declare/dischange messages.
- add tx state info to transfers and acknowledgements.
- Sync session after discharge.
- A transactional session automatically acks any message retrieved by fetch/get
to bring them into the transaction. This is consistent the 0-10 client.
Minor fixes to existing client code:
- Fix use of pn_drain API in C++ client to work with C++ and Java brokers.
- Make amqp::Exception derive from qpid::Exception
3. Fixes to existing broker code:
- Incoming.cpp fix: start async completion before processing message.
- Delay accept of dischage message till commit is complete.
- newSession - handle failover during session creation.
4. Added tests
interop_tests.py: transaction tests that can run against an external broker, see comments.
ha_tests.py: Enable transaction tests over AMQP 1.0.
Minor test fixes:
- brokertest.py don't set default logging if QPID_LOG env vars set.
- brokertest.py Pass kwargs to broker() create function.
- qpid-receive: capacity should never be larger than message count.
- Accept user:pass as well as user/pass in Url.
- brokertest.py: Always do a ready() check on all brokers.
If proton < 0.9 is used, transaction tests will be skipped or will downgrade to
the amqp0-10 protocol with a printed warning.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1662743 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1662275 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Increased maxNegotiateTimeout to the default (10 seconds). A smaller value speeds up
detection of non-running brokers on remote hosts, but this is not necessary for the tests.
Increased some other test timeouts and added some improved error reporting.
The occasional long (> 1 second) connection delays are caused by Cyrus SASL authentication.
Not clear why this takes so long, but that is a separate issue. Here's a client log excerpt
showing the delay.
2015-02-25 08:29:37.461299895 [Network] trace RECV [[127.0.0.1:34247-127.0.0.1:45983]]: Frame[BEbe; channel=0; {ConnectionStartBody: server-properties={qpid.federation_tag:V2:36:str16(77800bff-a176-46c1-917a-32f136dee650)}; mechanisms=str16{V2:9:str16(ANONYMOUS), V2:5:str16(PLAIN)}; locales=str16{V2:5:str16(en_US)}; }]
2015-02-25 08:29:37.463116303 [Security] debug CyrusSasl::start(ANONYMOUS PLAIN)
(Note delay > 1 sec here)
2015-02-25 08:29:38.839793753 [Security] debug min_ssf: 0, max_ssf: 256
2015-02-25 08:29:38.839851781 [Security] debug CyrusSasl::start(ANONYMOUS PLAIN): selected ANONYMOUS response: 'anonymous@wallace'
2015-02-25 08:29:38.839963162 [Client] warning Connection [127.0.0.1:34247-127.0.0.1:45983] closed
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1662247 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
| |
- Fix un-necessary re-sends in amqp0_10::SenderImpl::replay.
- Throw NotFound and UnauthorizedAccess correctly from amqp0_10::SessionImpl and ConnectionImpl
- Fix ha_test wait_address and valid_address re-using a session after it is closed by NotFound.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1622592 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
with failover.
This is partly not-a-bug, there is a client error handling issue that has been
corrected.
qpid-txtest2 initializes a queue with messages at the start and drains the
queues at the end. These operations are *not transactional*. Therefore
duplicates are expected if there is a failover during initialization or
draining. When duplicates were observed, there was indeed a failover at one of
these times.
Making these operations transactional is not enough to pass, now we see the test
fail with "no messages to fetch". This is explained as follows:
If there is a failover during a transaction, TransactionAborted is raised. The
client assumes the transaction was rolled back and re-plays it. However, if the
failover occurs at a critical point *after* the client has sent commit
but *before* it has received a response, then the the client *does not know*
whether the transaction was committed or rolled-back on the new primary.
Re-playing in this case can duplicate the transaction. Each transaction moves
messages from one queue to another so as long as transactions are atomic the
total number of messages will not change. However, if transactions are
duplicated, a transactional session may try to move more messages than exist on
the queue, hence "no messages to fetch". For example if thread 1 moves N
messages from q1 to q2, and thread 2 tries to move N+M messages back, then
thread 2 will fail.
This problem has been corrected as follows: C++ and python clients now raise the
following exceptions:
- TransactionAborted: The transaction has definitely been rolled back due to a
connection failure before commit or a broker error (e.g. a store error) during commit.
It can safely be replayed.
- TransactionUnknown: The transaction outcome is unknown because the connection
failed at the critical time. There's no simple automatic way to know what
happened without examining the state of the broker queues.
Unfortunately With this fix qpid-txtest2 is no longer useful test for TX
failover because it regularly raises TransactionUnknown and there's not much we
can do with that.
A better test of TX atomicity with failover is to run a pair of
qpid-send/qpid-receive with fail-over and verify that the number of
enqueues/dequeues and message depth are a multiple of the transaction size. See
the JIRA for such a test. (Note these test also sometimes raise
TransactionUnknown but it doesn't matter since all we are checking is that
messages go on and off the queues in multiple of the TX size.) )
Note: the original bug also reported seeing missing messages from
qpid-txtest2. I don't have a good explanation for that but since the
qpid-send/receive test shows that transactions are atomic I am going to let that
go for now.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1621211 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit adds a --cluster-manager flag to qpid-ha tool.
Without this flag
- the 'promote' command is not listed in the tool help.
- using the promote command raises an error saying that it is only for cluster manager use
and mentioning the --cluster-manager flag.
With the flag: promote functions as before.
The qpid-ha help text for promote is also more clear now that it is for cluster
manager only.
Originally the idea was to split qpid-ha into two tools but I have kept one tool
with the flag and warning messages because it:
- avoids packaging changes that might trip things up.
- helps people who are already using qpid-ha promote: their scripts will
break but the error message explains how to fix it.
I think the special role of promote is sufficiently clear now even if it is
part of the same tool.
This commit also updates the following to take account of the new flag:
- rgmanager qpidd-primary script.
- qpidd tests.
- qpid book HA chapter.
NOTE: THIS WILL BREAK TEST HARNESSES that do promotion outside of rgmanager.
You'll need to add the --cluster-manager flag in the relevant places.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1619877 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The test was hanging because of a python construct not available in 2.4. It was
causing an exception in a strange place because this bit of code was imported at
runtime, and that was hanging the test. Fixed and did some cleanup
to avoid such mysterious hangs in future:
- Fixed qpidtoollibs/config.py to work with python 2.4.
- Import qpid-ha script at import time rather than runtime.
- Fix Popen.teardown logic to avoid hanging if a process can't be killed.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1591794 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
| |
- Remove some dead code.
- Removed "set" command - not ready for production. All settings in qpidd.conf.
- Removed related tests in ha_tests
- Improved help on promote command.
- Made option group for common broker connection options.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1589834 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Added timeout to qpid-ha.
- qpidd init script pings broker to verify it is not hung.
- updated documentation in qpid/doc/book/src/cpp-broker/Active-Passive-Cluster.xml.
The new results for the cases mentioned in the bug:
a] stopped ALL brokers: rgmanager restarts the entire cluster but data is lost.
Equivalent to killing all the brokers at once. This does not affect quorum because
only qpidd services are affected, not other services managed by cman.
b] stopped the primary: rgmanager restarts the primary after a timeout and promotes one of the backups.
c] stopped a backup: rgmanager restarts the backups after a timeout.
Clients that are actively sending messages may see a delay while backup is restarted.
Note you need to set link-heartbeat-interval in qpidd.conf. The default is very
high (120 seconds), it should be set lower to see recovery from sigstop in a
reasonable time.
See the updated documentation in qpid/doc/book/src/cpp-broker/Active-Passive-Cluster.xml.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1589807 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The HA tests were using only AMQP 0-10.
Modified the tests to use AMQP 1.0 if available (still use 0-10 if 1.0 is not available)
Fixed bugs uncovered both in the tests and in the AMQP 1.0 implementation.
Summary of changes:
- brokertest.py: configurable support for of swig vs. native and amqp0-10 vs. 1.0
- default to swig+amqp1.0 if swig is available, native+amqp0-10 otherwise
- qpidtoollibs/broker.py: enable use of swig client with BrokerAgent
- Swig python client:
- support for passing client_properties/properties.
- expose AddressHelper pn_data read/write as PnData helper class
- set sender/receiver capacity on creation
- limited disposition support - rejected messages.
- support for additional timeout parameters
- expose messaging::Logger, allow log configuration to be set from python.
- ha_tests.py:
- bind, delete policies not supported by AMQP 1.0, switched to using BrokerAgent QMF.
- pass protocol:amqp1.0 connection-option to c++ test clients (qpid-send, qpid-receive)
- TX tests forsce use of 0-10 protocol (but still with Swig client if enabled.)
- Broker fixes:
- Queue::Settings::isTemporary was set in the 0-10 SessionAdapter, moved to Broker::createQueue.
- broker::amqp::Session was always setting an exclusive owner in createQueue
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1585588 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
| |
limit
This is regression introduced in r1561206: CommitDate: Fri Jan 24 21:54:59 2014 +0000
QPID-5513: HA backup fails if number of replicated queues exceeds number of channels.
Fixed by the current commit. PrimaryQueueLimits was not taking account of queues already
on the broker prior to promotion.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1585507 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
| |
--- log message follows this
NO-JIRA: Remove use of python built-in 'next', not available before python 2.6.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1565382 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
channels.
The problem:
- create cluster of 2 brokers.
- create more than 32768 queues (exceeds number of channels on a connection)
- backup exits with critical error but
- client creating queues receives no error, primary continues with unreplicated queue.
The solution: Primary raises an error to the client if it attempts to create
queues in excess of the channel limit. The queue is not created on primary
or backup, primary and backup continue as normal.
In addition: raised the channel limit from 32k to 64k. There was no reason for
the smaller limit. See discussion: http://qpid.2158936.n2.nabble.com/CHANNEL-MAX-and-CHANNEL-HIGH-BIT-question-tp7603121p7603138.html
New unit test to reproduce the issue, must create > 64k queues.
Other minor improvements:
- brokertest framework doesn't override --log options in the arguments.
- increased default heartbeat in test framework for tests that have busy brokers.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1561206 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
| |
object has no attribute 'name'"
This was due to a race condition where a session was deleted while the QmfAgent
was looking it up.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1540171 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1536751 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Appears to have been fixed at this point on trunk, not clear which checkins are
responsible. Test ran for 48 hours with no failures.
Other minor changes:
- Enable test_failover_send_receive
- Increase heartbeat interval.
- Reduce capacity of senders in failover test to be more aggressive.
- Use HaBrokerTest as test base
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1522711 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
| |
- Ignore un-replicated queues when replicating transactions.
- Clean up cancel logic in QueueReplicator, causing "no such subscription" errors.
- Remove unnecessary exchange delete warnings
- ha_test.py: Shorter timeout for starting cluster brokers.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1521192 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
| |
- Set auth info on status check connections
- Clean up status check loging
- Use realm@username for authentication name (was using just username)
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1521190 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1520108 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
| |
- Backups delete transactions on failover.
- TxReplicator cancel subscriptions when transaction is finished.
- TxReplicator rollback if destroyed prematurely.
- Handle special case of no backups for a tx.
- ha_tests.py: new and modified tests to cover the new functionality.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1518982 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
| |
During a transaction:
- A broker leaving aborts the transaction.
- A broker joining does not participate in the transaction
- but does receive the results of the TX via normal replication.
Clean up tx-queues when the transaction completes.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1510678 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
| |
Backups send prepare messages to primary, primary delays completion of prepare
till all are prepared (or there is a failure).
This is NOT the production solution - blocking could cause a deadlock. We need
to introduce asynchronous completion of prepare without blocking. This
interim solution allows testing on other aspects of TX support.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1509424 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On primary a PrimaryTxObserver observes a transaction's TxBuffer and generates
transaction events on a tx-replication-queue. On the backup a TxReplicator
receives the events and constructs a TxBuffer equivalent to the one in the
primary.
Unfinished:
- Primary does not wait for backups to prepare() before committing.
- All connected backups are assumed to be in the transaction, there are race
conditions around brokers joining/leavinv where this assumption is invalid.
- Need more tests.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1509423 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
| |
test_expected_backup_timeout
Very sporadic failures so difficult to verify the fix.
- Simplified Membership, centralized status change, make it atomic.
- Fix test bug in test_expected_backup_timeout: not waiting on final status check, race.
- Remove out-of-date status info from log prefixes: Guard, ReplicatingSubscription
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1495466 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
| |
(also remove final references to dead watchdog plugin)
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1494697 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously HA code used queue sequence numbers to identify messasges.
This assumes that message sequence is identical on primary and backup.
Implementing new features (for example transactions) requires that we tolerate
ordering differences between primary and backups.
This patch introduces a new, queue-scoped HA sequence number managed by the HA
plugin. The HA ID is set *before* the message is enqueued and assigned a queue
sequence number. This means it is possible to identify messages before they are
enqueued, e.g. messages in an open transaction.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1493771 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
| |
Add support for the "amq.failover" exchange with new HA, to support migration of
clients that used this facility with the old cluster.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1485511 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Many HA tests use --port=0 to start a broker on an available port, but then need
to shutdown and restart the broker on the same port. This is not safe, on a busy
system it is possible for another process to take the port between the time the
broker is shut down and the time it is restarted.
The solution is to do bind(0) and listen in the python test framework (class
HaPort) and let the broker use the socket using qpidd --socket-fd. When the
broker is shut down the port remains bound by the python process. When the
broker is re-started it again is given access to the socket via --socket-fd.
Other changes
- move ha_store_tests into ha_tests.
- add heartbeats to avoid stalling.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1482881 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
| |
This issue evolved a bit between the original discussion and the final
commit. See https://reviews.apache.org/r/10658/ for the details.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1477112 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1456950 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
| |
ha_tests.ReplicationTests.test_auto_delete_timeout
The tests were not waiting for the cluster to be ready before starting.
Updated HaCluster to wait by default before returning.
Increase timeout in calls to wait_no_queue, the default timeout of 1 sec was the same
as the auto-delete timeout.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1441157 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
'--known-hosts-url' option but cannot be disabled
Reverts the previous commit and simplifies the semantics of setting
--ha-public-url and --ha-brokers-url. There is no longer any over-riding or
implicit updating of values. That means you must set --ha-public-url as well as
--ha-brokers-url, it will not be defaulted. Likewise if you *dont* set
ha-public-url, it will remain empty, which is the use case in this bug.
The defaulting was adding complexity without adding much value.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1421934 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Imagine a cluster with primary A and backups B and C. A queue Q is created on A
and replicated to B, C. Now A dies and B takes over as primary. Before C can
connect to B, a client destroys Q and creates a new queue with the same name.
When B connects it sees Q and incorrectly assumes it is the same Q that it has
already replicated. Now C has an inconsistent replica of Q.
The fix is to tag queues/exchanges with a UUID so a backup can tell if a queue
is not the same as the one it has already replicated, even if the names are the
same. This all also applies to exchanges.
- Minor imrovements to printing UUIDs in a FieldTable.
- Fix comparison of void Variants, added operator !=
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1409241 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
| |
descriptive text.
Also added description for --link-heartbeat-interval
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1397253 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
| |
Subscribed auto-delete queues are deleted by the backup.
Timed auto-delete queues are deleted after the timeout.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1397243 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
| |
primary.
Test test_delete_missing_response was failing with "cluster active, cannot promote".
- Fixed test bug: "fake" primary triggered "cannot promote".
- Backup: always create QueueReplicator if not already existing.
- Terminology change: "initial" queues -> "catch-up" queues.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1396244 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
replicate a deleted queue. (Based on patch by Jason Dillama)
This does not directly tackle the origin of the problem but extends Jasons's patch since
it addresses something we had to fix anyway: "leaking" queues and exchanges. It does 2 things.
1. enabled hideDeletedError on all subscription objects used by HA
This suppress the troublesome exception with a harmless no-op
2. Delete queues/exchanges missing from responses (based on Jasons patch)
Fix the "leak" of queues and exchanges possible when an object replicated
to a backup is deleted from the newn primary before the backup connects.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1393089 13f79535-47bb-0310-9956-ffa450edef68
|
| |
|
|
|
|
|
|
|
|
| |
In order to suppress unwanted warnings from certain test, the ha_test framework
was actually turning off all python logging.
This patch selectively turns off wanrnings in specific code regions and then
restores the configured logging level.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1391232 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
When re-starting a persistent HA cluster, the broker that becomes primary should
keep its store data while all the backup brokers should discard their store data
and catch up from the primary. Backups cannot simply use their own stores
because sequence numbers of stored messages will not match on all brokers. The
backup erases individual queues and exchanges as the catch-up process gets to
them.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1390123 13f79535-47bb-0310-9956-ffa450edef68
|
