From cfb1b1056e35892c04fbdafd486913bba5054587 Mon Sep 17 00:00:00 2001 From: Keith Wall Date: Tue, 21 Oct 2014 16:11:49 +0000 Subject: QPID-6154: [Java Broker] HA - Handle rollback of node when use of weak durability has allowed nodes to diverge git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk@1633407 13f79535-47bb-0310-9956-ffa450edef68 --- ...roker-Appendix-Operational-Logging-Messages.xml | 40 +++++++++++--- .../java-broker/Java-Broker-High-Availability.xml | 64 +++++++++++----------- 2 files changed, 65 insertions(+), 39 deletions(-) (limited to 'qpid/doc') diff --git a/qpid/doc/book/src/java-broker/Java-Broker-Appendix-Operational-Logging-Messages.xml b/qpid/doc/book/src/java-broker/Java-Broker-Appendix-Operational-Logging-Messages.xml index 7d3d62b075..8b84605683 100644 --- a/qpid/doc/book/src/java-broker/Java-Broker-Appendix-Operational-Logging-Messages.xml +++ b/qpid/doc/book/src/java-broker/Java-Broker-Appendix-Operational-Logging-Messages.xml @@ -346,21 +346,27 @@ BRK-1014 - Message flow to disk active : Message memory use size of all messages exceeds threshold threshold size + Message flow to disk active : Message memory use size of all + messages exceeds threshold threshold + size - Indicates that the heap memory space occupied by messages has exceeded the threshold so the flow to disk feature has been activated. + Indicates that the heap memory space occupied by messages has exceeded the + threshold so the flow to disk feature has been activated. - + BRK-1015 - Message flow to disk inactive : Message memory use size of all messages within threshold threshold size + Message flow to disk inactive : Message memory use size of all + messages within threshold threshold + size - Indicates that the heap memory space occupied by messages has fallen below the threshold so the flow to disk feature has been deactivated. + Indicates that the heap memory space occupied by messages has fallen below the + threshold so the flow to disk feature has been deactivated. @@ -1196,9 +1202,9 @@ - The node is no longer reachable. This may be as a result of the node being - stopped or a network partition may be preventing it from being connected. The node is - still a member of the group. + The node is no longer reachable. This may be as a result of the node being stopped + or a network partition may be preventing it from being connected. The node is still a + member of the group. @@ -1273,6 +1279,24 @@ node groups only. + + HA-1014 + HA-1014 : Diverged transactions discarded + + + + This node is in the process of rejoining the group but has discovered that some + of its transactions differ from those of the current master. The node will + automatically roll-back (i.e. discard) the diverging transactions in order to be + allowed to rejoin the group. This situation can only usually occur as a result of use + of the weak durability options. These allow the group to operate with fewer than + quorum nodes and therefore allow the inconsistencies to develop. + On encountering this condition, it is strongly recommendend + to run an application level reconcilation to determine the data that has been + lost. + + + diff --git a/qpid/doc/book/src/java-broker/Java-Broker-High-Availability.xml b/qpid/doc/book/src/java-broker/Java-Broker-High-Availability.xml index 25c56a8d14..8282beaa5e 100644 --- a/qpid/doc/book/src/java-broker/Java-Broker-High-Availability.xml +++ b/qpid/doc/book/src/java-broker/Java-Broker-High-Availability.xml @@ -120,9 +120,9 @@ complete, all nodes will be considered equal. + Click the Add button on the Virtualhost Panel on the Broker + tab. - Click the Add button on the Virtualhost Panel on the Broker - tab. Give the Virtualhost node a unique name e.g. weather1. The @@ -157,25 +157,23 @@ master. - -
- Creating 1st node in a group - - - - - - Creating 1st node in a group - - -
- +
+ Creating 1st node in a group + + + + + + Creating 1st node in a group + + +
+ Now move to the second Broker to be the group. Click the Add + button on the Virtualhost Panel on the Broker tab of the second Broker. - Now move to the second Broker to be the group. Click the Add - button on the Virtualhost Panel on the Broker tab of the second Broker. Give the Virtualhost node a unique name e.g. @@ -202,19 +200,17 @@ Repeat these steps until you have added all the nodes to the group. - -
- Adding subsequent nodes to the group - - - - - - Adding subsequent nodes to the group - - -
- +
+ Adding subsequent nodes to the group + + + + + + Adding subsequent nodes to the group + + +
@@ -351,6 +347,8 @@ running behind for some reason (perhaps a full-GC). If a Master failure were to occur at that moment, the replicas would elect Replica-1 even though Replica-2 had the most recent transaction. + Transaction loss is reported by message HA-1014.
@@ -378,6 +376,8 @@ >split-brain in the event of a network partition. If used to temporarily restore service from fewer than quorum nodes, it is imperative to revert it to the Default value as the failed nodes are restored. + Transaction loss is reported by message HA-1014.
@@ -397,7 +397,9 @@ It is imperative not to allow designated primary to be set on both nodes at once. To do so will mean, in the event of a network partition, a split-brain will - occur. + occur. + Transaction loss is reported by message HA-1014.
-- cgit v1.2.1