Provider position clutters the public interface and is not
required by wsrep-API v26. Remove the position from ws_meta/view
classes and deal with it internally in custom providers.
commit 3b419aa6e2cced92c07c571116aad3d55cd1e7e4
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Sun Feb 19 10:29:34 2023 +0200
Skip fetching config options if provider not loaded via wsrep-API
commit 044220cc067cbf74a956fa1a4476f6d873a78210
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Wed Jul 13 10:31:03 2022 +0300
Operation context pointer for client state
commit eeb05a92384933dd026d5ca3c6f854510bb76eed
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Mon Jul 4 09:03:23 2022 +0300
Add unit test log in gitignore
commit 92a04070fc0dc4f55dc0415442ed9003c29a9cf6
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Sun May 8 12:45:36 2022 +0300
Added convenience method prev() to seqno
commit f83ca1917e50bfe29c3912ccd3a9989ead68378a
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Sun May 1 16:37:24 2022 +0300
Pass victim context for provider on BF abort
This change is needed for custom provider implementations to
have a way to access the victim in the application context.
Helper interface operation_context to pass caller context for
service/provider callbacks in more type safe way.
commit 244eabe8cfcc86ca49981b0d8d081b740a967bda
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Wed May 25 07:39:43 2022 +0300
Handle disconnecting state in on_sync()
When disconnecting from the group, the sync event from the
provider must not change the state back to synced.
commit ba8e23df0dfec71427aac0d35036f26d908b4995
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Tue Mar 22 17:43:52 2022 +0200
Add provider position field to ws_meta and view
Provider position is needed in coordinated recovery
between application and provider. Pass the position
info from provider to application to allow making
it durable.
commit 53e60f64c953b252a47bc91e87b4131465b5f15f
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Sat Mar 19 14:45:57 2022 +0200
Reset TOI meta after releasing total order in provider
This is to keep the TOI meta available in case the provider
implementation needs it.
commit bccb9997f29e94a2d5160d388454ace5f89efb6c
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Mon Jan 3 11:19:58 2022 +0200
Fixed id ostream operator to print human readable ids
commit 6d0b37daafdb87975931b0bd95bb8ea821a4071e
Author: Teemu Ollakka <teemu.ollakka@galeracluster.com>
Date: Wed Dec 15 16:37:45 2021 +0200
Silence unused variable warning
commit 4b8616f3d13137d992b8b994baa48f2981238352
Author: Denis Protivensky <denis.protivensky@galeracluster.com>
Date: Wed Dec 15 16:43:31 2021 +0300
Fix provider loading in test for release builds
commit 6df17812d945a07a6f199e6fe83e287afbc28ed9
Author: Denis Protivensky <denis.protivensky@galeracluster.com>
Date: Tue Dec 14 20:28:56 2021 +0300
Introduce set_provider_factory() method for server_state
This allows injecting an application allocated provider into
server_state.
After this virtual provider getter is unnecessary. Made the getter
normal method and fixed unit tests accordingly.
This patch introduces a queue to store ids of transactions that failed
to send a rollback fragment in streaming_rollback(). This is to avoid
potentially missed rollback fragments when a cluster splits and then
later reforms. Rollback fragments would be missing if a node rolled
back a transaction locally (either BFed or voluntary rollback) while
non-primary, and the attempt to send rollback fragment failed in
transaction::streaming_rollback().
Transaction that fail to send rollback fragment can proceed to
rollback locally. However we must ensure that rollback fragments for
those transactions are eventually delivered by the cluster. This must
be done before a potentially conflicting writeset causes BF-BF
conflicts in the rest of the cluster.
Fixes a bug where the fact that an SR master leaves the primary view
gets missed. When two consecutive primary views have the same
membership we now assume that every SR needs to be rolled back, as the
system may have been through a state of only non-primary components.
Use iterators for scanning members vector in order to avoid
issues with integer signedness and range checks. The vector is
usually rather small and not in hot codepath, so performance
is here not an issue.
Added unit test for member_index() method.
When member joins the group and needs to receive an SST it won't
receive the corresponding menbership view event because the SST
happens after the event and will already include the effects of
all events ordered before it. The view then must be recovered from
the received state.
Minor renames and cleanups.
References codership/wsrep-lib#18
it on disconnect.
- Don't rely on own index from the view because the view may come from
another member (IST/SST), instead always determine own index from own ID.
Refs codership/wsrep-lib#13
* Added bootstrap service call to do DBMS side bootstrap operations
during the cluster bootstrap.
* Added last_committed_gtid() to provider interface
* Implemented wait_for_gtid() provider call
* Pass initial position to the server state
* Propagate server max protocol version to provider init options
* Store gtid from connected call to make cluster id and the connect
position available
