We want the statements to be removed before we ack the schema change,
otherwise it will race with all future operations.
Since the subscriber will be invoked on each shard, there is no need
to broadcast to all shards, we can just handle current shard.
Replicates https://issues.apache.org/jira/browse/CASSANDRA-7910 :
"Prepare a statement with a wildcard in the select clause.
2. Alter the table - add a column
3. execute the prepared statement
Expected result - get all the columns including the new column
Actual result - get the columns except the new column"
Currently the notify_*() method family broadcasts to all shards, so
schema merging code invokes them only on shard 0, to avoid doubling
notifications. We can simplify this by making the notify_*() methods
per-instance and thus shard-local.
If the schema_builder is constructed from an existing schema we need to
make sure that the original column ids of regular and static columns are
*not* used since they may become invalid if columns are added or
removed.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
When a column is dropped its name and deletion timestamp are added
to schema::_raw._dropped_columns to prevent data resurrection in case a
column with the same name is added. To reduce the number of lookups in
_dropped_columns this patch makes each instance of column_definition
to caches this information (i.e. timestamp of the latest removal of a
column with the same name).
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
Knowing which columns were dropped (and when) is important to prevent
the data from the dropped ones reappearing if a new column is added with
the same name.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
We want the node's schema version to change whenever
table_schema_version of any table changes. The latter is calculated by
hashing mutations so we should also use mutation hash when calculating
schema digest.
We need to track which schema version were synced with on current node
to avoid triggering the sync on every mutation. We need to sync before
mutating to be able to apply the incoming mutation using current
node's schema, possibly applying irreverdible transformations to it to
make it conform.
There is one current schema for given column_family. Entries in
memtables and cache can be at any of the previous schemas, but they're
always upgraded to current schema on access.
The verb belongs to a seaprate client to avoid potential deadlocks
should the throttling on connection level be introduced in the
future. Another reason is to reduce latency for version requests as it
can potentially block many requests.
The intent is to make data returned by queries always conform to a
single schema version, which is requested by the client. For CQL
queries, for example, we want to use the same schema which was used to
compile the query. The other node expects to receive data conforming
to the requested schema.
Interface on shard level accepts schema_ptr, across nodes we use
table_schema_version UUID. To transfer schema_ptr across shards, we
use global_schema_ptr.
Because schema is identified with UUID across nodes, requestors must
be prepared for being queried for the definition of the schema. They
must hold a live schema_ptr around the request. This guarantees that
schema_registry will always know about the requested version. This is
not an issue because for queries the requestor needs to hold on to the
schema anyway to be able to interpret the results. But care must be
taken to always use the same schema version for making the request and
parsing the results.
Schema requesting across nodes is currently stubbed (throws runtime
exception).
Schema is tracked in memtable and cache per-entry. Entries are
upgraded lazily on access. Incoming mutations are upgraded to table's
current schema on given shard.
Mutating nodes need to keep schema_ptr alive in case schema version is
requested by target node.
