Some changes to the threads configuration elements
by David M. Lloyd
I was talking to Andy Miller about thread pool configuration and he
suggested a couple things. First, he thought that having count and
per-cpu arguments cause more confusion than help. I think this is
probably a good idea since people seem mostly confused by it.
Second, he proposed a flag which would cause thread pools to be
automatically pre-filled, which seems like a pretty good idea.
Questions/comments?
--
- DML
15 years
exposing Transaction timeout
by Francesco Marchioni
Dear devs,
a small proposal: would it be fine to expose transaction timeout defaults as
a property in the transactions subsystem ?
AFAIK you have to include it in the coordinator-environment element. ....
<subsystem xmlns="urn:jboss:domain:transactions:1.0">
<core-environment socket-binding="txn-socket-process-id"/>
<recovery-environment socket-binding="txn-recovery-environment"
status-socket-binding="txn-status-manager"/>
<coordinator-environment enable-statistics="false"
default-timeout="300" />
</subsystem>
kind regards,
Francesco Marchioni
15 years
attribute defaults
by Heiko W.Rupp
Hi,
in subsystem=connector : read-resource-description I see
"archive-validation-fail-on-warn" => {
"description" => "Should an archive validation warning report fail the deployment. Default: false",
"type" => BOOLEAN,
"required" => false,
"access-type" => "read-only",
"storage" => "configuration"
},
:read-resource does not return that attribute. This is no big deal, as it has a default.
BUT: for tools it is not easy to guess that default
Could the resource-description please contain the default in a parseable way
"archive-validation-fail-on-warn" => {
"description" => "Should an archive validation warning report fail the deployment. Default: false",
"type" => BOOLEAN,
"default-value" => true,
[...]
},
Thanks
Heiko
Btw.: this relates back to the question of an effective running configuration.
15 years
CLI Archive
by Brian Stansberry
Notes from a conversation today with Dimitris, John, Jesper, Jason. Not
an urgent priority, just some ideas.
Idea is to have an archive format that can be processed by the CLI. The
contents are:
1) A script containing CLI commands; the same kind of file it can read now.
2) Any content associated with the commands.
The commands that refer to the content (e.g. deploy foo.war) would use
paths relative to the root of archive.
Basically package content together with the script.
The file could be created by hand, or by the CLI itself (or the web
console) by recording commands and then instead of executing them, write
out the archive.
--
Brian Stansberry
Principal Software Engineer
JBoss by Red Hat
15 years
Accessing subsystem runtime attributes
by Francesco Marchioni
Dear devs,
I'm wondering how runtime attributes (such as **ConnectionCount for a
ManagedConnection Pool, or Thread's QueueSize) could be gathered on AS 7.
By using CLI ? Using a recursive read-resource, just dumps the resource
startup properties.....
[localhost:9999 /]
/subsystem=datasources/data-source="java:/H2DS":read-resource
(recursive="true")
{
"outcome" => "success",
"result" => {
"connection-url" => "jdbc:h2:mem:test;DB_CLOSE_DELAY=-1",
"driver-class" => "org.h2.Driver",
"jndi-name" => "java:/H2DS",
"driver" => "org.h2.Driver#1.2",
"pool-name" => "H2DS",
"use-java-context" => true,
"enabled" => true,
"pool-prefill" => true,
"pool-use-strict-min" => false,
"user-name" => "sa",
"password" => "sa",
"share-prepared-statements" => false,
"set-tx-query-timeout" => false,
"background-validation" => false,
"use-fast-fail" => false,
"validate-on-match" => false
},
"compensating-operation" => undefined
}
Thanks a lot
Francesco Marchioni
15 years
Configuration and user friendlyness
by Heiko W.Rupp
Hi,
I understand that for many configuration changes, direct editing of values (i.e. write-attribute) is not supported for various reasons
- only changing one attribute may introduce inconsistencies
- applying the change would require a restart
So the result is that many items (better name?) just allow to :remove them and then to :add
them again.
This may be fine for items that only have one or two attributes,
but for a data source as an example with 30 attributes, it is a pain for the user that
is not using the embedded console or RHQ to specify them all in a curl statement
or even in the :add(...) command in the CLI just to increase some Tx timeout or to provide
a different password.
There are of course other items in the management tree where this applies as well.
Heiko
15 years
Datasource attribute modifications
by Heiko Braun
Looks like we have a similar issue with other DS attributes:
[domain@localhost:9999 /] /profile=default/subsystem=datasources/data-source="java:/H2DS":write-attribute(name=user-name,value=test)
{
"outcome" => "failed",
"failure-description" => {"domain-failure-description" => "Attribute user-name is not writeable"}
}
It seems that all DS attributes are set to be read-only.
Like this modifications to DS resources are not possible.
Stefano, any thoughts on this?
Ike
15 years
JNDI access to JMS connection factory
by Marius Bogoevici
Hi,
I am trying to figure out whether JBoss AS7 does (or will) provide a
JNDI-accessible JMS connection factory, e.g. via "java:/JmsXA" or
similar (perhaps in the java:jboss namespace). Can anyone help me with
more details?
Thanks,
Marius
15 years
[AS7-367] exposing deployment details via the mgmt api
by Emanuel Muckenhuber
I was wondering about our requirements of exposing more detailed
deployment related information via our management API.
https://issues.jboss.org/browse/AS7-367 lists things like deployed
servlets, ejbs, ... with related invocation stats and other metrics.
In general this sounds more dynamic than other parts of our current mgmt
api and the structure mostly depend on how much detail we expose - and
maybe more importantly if we actually want to allow overriding
deployment descriptors using this mechanism right from the beginning?
Any thoughts on this are welcome,
Emanuel
15 years
Oracle datasource missing dependencies
by paulwerder＠fastmail.co.uk
Hi JBoss devs,
I'm testing Oracle connectivity using a 10.2 JDBC Driver (ojdbc14.jar).
Seems there are some issues- has anyone managed to connect to an Oracle
DB ?
This is what the server reports:
report
New missing/unsatisfied dependencies:
service jboss.jdbc-driver."oracle.jdbc.driver.OracleDriver".10.2
(missing)
And my configuration:
<datasource jndi-name="java:/OracleDS" pool-name="OracleDS"
enabled="true" use-java-context="true">
<connection-url>
jdbc:oracle:thin:@192.168.1.1:1530:ORACLESID
</connection-url>
<driver-class>
oracle.jdbc.driver.OracleDriver
</driver-class>
<driver>
oracle.jdbc.driver.OracleDriver#10.2
</driver>
<pool>
<prefill>
true
</prefill>
<use-strict-min>
false
</use-strict-min>
</pool>
<security>
<user-name>test</user-name>
<password>test</password>
</security>
<validation>
<validate-on-match>
false
</validate-on-match>
<background-validation>
false
</background-validation>
<useFastFail>
false
</useFastFail>
</validation>
<statement/>
</datasource>
I've chosen to install the JDBC driver as a module, so under the
"modules" folder I've added:
oracle
└───jdbc
└───main
module.xml
ojdbc14.jar
ojdbc14.jar.index
Last thing, the module.xml:
<module xmlns="urn:jboss:module:1.0" name="oracle.jdbc">
<resources>
<resource-root path="ojdbc14.jar"/>
<!-- Insert resources here -->
</resources>
<dependencies>
<module name="javax.api"/>
</dependencies>
</module>
Can you find any reason why the driver does not get deployed ?
(I've tried also adding META-INF/services/java.sql.Driver with the
driver class name- it didn't help however).
Thanks a lot
Paul
--
paulwerder(a)fastmail.co.uk
--
http://www.fastmail.fm - IMAP accessible web-mail
15 years
