It is, though I don't work on it as often as I'd like. On Tue, Oct 8, 2019 at 4:17 AM Rogerio rogerthat@users.sourceforge.net wrote: Hi, is this project still alive? HA-JDBC Releases. https://sourceforge.net/p/ha-jdbc/discussion/383396/thread/2041e133/?limit=25#ad8c Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/ha-jdbc/discussion/383396/ To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/
HA-JDBC invokes the load balancer for any JDBC method it expects to contact the database server in an immutable way. Sometimes this results in 2 calls to the load balancer for a given statement execution, sometimes more, e.g. calls to Connection.getAutoCommit() or prepareStatement(...) or getTransactionIsolation(). Sequential tests like this do nothing to test balancer behavior. Round-robin (or any balanacer logic) is pointless without concurrent activity. Perform the same test with multiple concurrent...
This is the equivalent code in HA-JDBC 3.0: Dialect dialect = new OracleDialect(); DatabaseProperties database = new LazyDatabaseProperties(new SimpleDatabaseMetaDataProvider(metaData), dialect); TableProperties table = databaseProperties.findTable(tableName); Collection<UniqueConstraint> constraints = table.getUniqueConstraints(); Paul On Mon, Jul 31, 2017 at 7:14 PM "Héctor Cadavid" hectorcadavid@users.sf.net wrote: Dear friends, I'm trying to upgrande a 8-years legacy application that worked with...
Using 3.0.x, you can create a DataSourceDatabase that references your DataSource...
Sorry for the slow response. Have you figured this out yet? I don't see HA-JDBC anywhere...
The reason for this is that in HA-JDBC we send fine/coarse durability commands per...
To use the JDBC url style of access, you must use HA-JDBC's jdbc driver as your entry...
Hi. Sorry for the delay in responding. For which JDBC url was no suitable driver...
Question... Why don't you use the ResultSet returned by PreparedStatement.executeQuery(...)...
HA-JDBC 3.0 and JGroups 3.6 are not compatible. You can either recompile with the...
Fixed. https://github.com/ha-jdbc/ha-jdbc/commit/1f42722247e80611739a18f447c92d3...
A couple of issues: Your module files should only include the module.xml, ha-jdbc...
If you remove UNICAST3 from your stack, messages can be sent concurrently - whereas...
You'd like to use JNDI from your application, correct? Just bind a DataSource to...
Can you post a test that reproduces the problem?
Your <database location="..."/> should define a jdbc url - not a jndi name.
The database ids do not need to be unique across clusters. Minimally, it is enough...
Sorry about those error messages - they were auto-generated (and probably out-of-date)...
I don't quite understand. Can you post your bean xml and HA-JDBC xml?
Can you provide more context for the InvocationTargetException? What stacktrace do...
Did you re-activate databaseB prior to attempting to stop databaseA? HA-JDBC will...
Did you re-activate databaseB prior to attempting to stop databaseA? HA-JDBC will...
I haven't gotten around to adding that yet. I'll get around to it this week. Feel...
Of your suggestions, it makes the most sense to provide the appropriate cluster context...
The NPE is coming from your membership listener - what are you doing in that method?...
Regarding #1, I verified that this is already the case, i.e. locks are acquired in...
https://github.com/ha-jdbc/ha-jdbc/commit/16534afe2789ea8b007c7a36a817c4424f11bf...
I'll sort this out today.
Because the new changes are more sensitive to suspects and address changes across...
I haven't yet - but have made some changes to the CommandDispatcher stuff in the...
I've refactored the CommandDispatcher API to draw a better distinction between sender...
https://github.com/ha-jdbc/ha-jdbc/commit/cde815332de6774345670bda80119bb31e20e3...
You are right - if a given statement requires multiple lock acquisitions, the deadlocks...
Thanks for pointing this out. I'll look into it.
For now, the easiest way to add custom MembershipListener logic is to supply your...
It is not needed in 3.0.3, because the local node is no longer excluded from calls...
Good point. https://github.com/ha-jdbc/ha-jdbc/issues/80
There isn't any split brain handling at the moment. JGroups can detect network partitions,...
https://github.com/ha-jdbc/ha-jdbc/commit/5e8634250b3884ace4a90ece422b403ce9c132...
I'll have a fix ready today.
The CommandDispatcher registers a MembershipListener with the JGroups MessageDispatcher....
In general, you would detect this via org.jgroups.MembershipListener. The viewAccepted(View)...
I see what's happening. The intended behavior is that a non-coordinator member would...
Do you still have these issues with distributable mode using version 3.0.3? IIRC,...
Agreed. In general I should add a section that includes dbms-specific notes and ...
In 3.0.0, the calls to DatabaseMetaData.getTables(...) returned tables for all schemas....
Hmm. That's doubly unfortunate. I think the only way to do this is to install some...
Hmm. That's true. As far as I can tell, Oracle doesn't support resetting the value...
In 3.0.0, the calls to DatabaseMetaData.getTables(...) returned tables for all s...
I can't flush metadata on startup using "dbuser" because HA-JDBC knows nothing about...
Actually, I don't think it was a good idea to force the user name as the schema name....
https://github.com/pferraro/ha-jdbc/commit/bd5cd86f36cd28185f666295abaa2e7771fcf...
The permissions required by the user defined in your ha-jdbc configuration is specific...
Every database in the cluster must have the same schemas. That is a requirement.
Actually - drop and recreating is much easier. https://github.com/ha-jdbc/ha-jdb...
This is fixed in the 3.0 branch already. Ill release 3.0.3 later today.
I should be able to wrap the subtraction with an eval(...) Or whatever the oracle...
OK - this is an issue with the OracleDialect.schemaPattern(DatabaseMetaData) method....
This is fixed in the latest code from the 3.0 branch. We are due for a 3.0.3 rel...
I understand, but can you run your application on server1 (without ha-jdbc), but...
It seems to me that the issue is one of 2 things: either the SQL generated by your...
On the contrary, if the sql contains any server specific qualifiers, then it would...
OK, so the problem isn't with obtaining a connection, but rather with your query....
Of course HA-JDBC will try to use both of them. That's how it was designed to work....
Can you paste the stacktrace of the exception? I'm not familiar with this jdbc d...
It looks like your definition for db2 is missing a databaseName and portNumber.
For what purpose?
Transactionally, ha-jdbc makes the same guarantees as a single database. You would...
If the databases defined in your cluster were themselves managed by the application...
If the databases defined in your cluster were themselves managed by the application...
Yes, it is. Why do you need to access this? The DatabaseMetaData object return by...
Most application servers provide a facility for defining database resources including...
Your bootstrap code should include database definitions for both db1 and db2.
I understand the question. No - the authentication information used by your application...
I have started work on this, but I'm still a long way from anything workable. It's...
HA-JDBC can handle certain non-deterministic sql functions, like sysdate(). When...
Oops - I misspoke. HA-JDBC doesn't iterate over the defined properties and look for...
Oops - I misspoke. HA-JDBD doesn't iterate over the defined properties and look for...
Correct. When configuring a javax.sql.DataSource, any configured properties are interpreted...
I'll look into this some more this week. I know we are still seeing dropped unicasts...
Your application only ever needs to reference the org.springframework.jdbc.datasource.DriverManagerDataSource,...
Don't reference your databases by jndi name. The net.sf.hajdbc.sql.Driver expects...
Don't reference your databases by jndi name. The net.sf.hajdbc.sql.Driver expect...
The concept of sticky sessions isn't really relevant to HA-JDBC. It implies that...
OK - it seems DBCP (both Commons & Tomcat) already takes care of connection rollback....
https://github.com/ha-jdbc/ha-jdbc/commit/67d1f8fe5b1e3a9b9d57431be25cb53488da33...
Ah - I see. So you're using the HA-JDBC DataSource which proxies DBCP DataSources....
Unfortunately, I don't see how I can reasonably "stop" a DataSource, given that,...
Re: the dump-restore failure using dataOnly=true, is there any other log output?...
A few comments: 1. Technically speaking, your threads are not updating their respective...
Here's how HA-JDBC works. When you call DataSource.getConnection(), the connection...
Not sure how to get which connection object? The underlying connection to each database?...
A: Yes - the relevant commit is here: https://github.com/ha-jdbc/ha-jdbc/commit/559f5cf8845e113fa76313335d523de428905531...
If the reactivation was really triggered from Tomcat 1, why would this stacktrace...
What exception is that? Are you talking about the timeout during mysql restore? I...
https://github.com/ha-jdbc/ha-jdbc/commit/559f5cf8845e113fa76313335d523de4289055...
That change is intentional - but I do see the problem now. I'll have a fix momen...
Your configuration looks fine. The balancer is invoked any time a JDBC operation...
The postgresql option for this is "pg_restore --data-only". From what I can tell,...
Is there a command line option in mysqldump that enables this behavior? If so, let's...