Please see Youtube video titled: "JBoss Transactional Integrity" youtube.com/watch?v=5RiTGRCCcwg which contains a pointer to the “actual working code/test scenario” which proves that a network failure occurring in midstream during transaction processing indeed does recover correctly.
It works precisely as designed and we have shown this many times over the past 25 years! Trial by Youtube is hardly the most scientific way of doing things, but rest assured that we have done our own video and will post it very soon. Go to the JBossTS forums in future though, because that's the best place to discuss any topic in an open manner.
This is FUD because it is not true. No patch was necessary to EAP 5 and yes, it has been demonstrated. Since Trial By Youtube seems to be the depths to which IBM have sunk recently we'll obviously have to get down to their level and post a response, so watch this space.
BTW, IBM has still not presented their test case, which is something we will do when we post our response.
If this worked as designed on JBoss, then I would expect some sort of proof and so far there is none. Please let us know when you fix this so we can test it again. The EAP 5.0.1 fails this test.
One other thing - the transactions fail and get lost with JBoss when using messaging providers, not only two separate databases as shown in the demo.
@superdomethis If you have a test case that illustrates any problem with any project, take it to the relevant forum where all of the teams and communities will be more than happy to help.
Why do you say this is FUD? Did you try this? Can you post a fix to this "bug"?
The Enterprise version was tested in March 2010 and failed this test. The extensive community support was utilized. The code is simple - just update two databases via JDBC or EJB using XA JDBC drivers.
Did you bother to raise any questions on the JBoss forums. They are free after all! Also did you make your code publicly available? That might help track down user error versus issues with the application server. Until that happens and the problem is reproducible, this appears as so much FUD.
Show us the code!
tuelhostube 1 year ago
Comment removed
mharvey2561 1 year ago
Comment removed
mharvey2561 1 year ago
Comment removed
mharvey2561 1 year ago
This has been flagged as spam show
Please see Youtube video titled: "JBoss Transactional Integrity" youtube.com/watch?v=5RiTGRCCcwg which contains a pointer to the “actual working code/test scenario” which proves that a network failure occurring in midstream during transaction processing indeed does recover correctly.
mharvey2561 1 year ago
Comment removed
mharvey2561 1 year ago
It works precisely as designed and we have shown this many times over the past 25 years! Trial by Youtube is hardly the most scientific way of doing things, but rest assured that we have done our own video and will post it very soon. Go to the JBossTS forums in future though, because that's the best place to discuss any topic in an open manner.
MarkCLittle 1 year ago
This is FUD because it is not true. No patch was necessary to EAP 5 and yes, it has been demonstrated. Since Trial By Youtube seems to be the depths to which IBM have sunk recently we'll obviously have to get down to their level and post a response, so watch this space.
BTW, IBM has still not presented their test case, which is something we will do when we post our response.
MarkCLittle 1 year ago
If this worked as designed on JBoss, then I would expect some sort of proof and so far there is none. Please let us know when you fix this so we can test it again. The EAP 5.0.1 fails this test.
One other thing - the transactions fail and get lost with JBoss when using messaging providers, not only two separate databases as shown in the demo.
superdomethis 1 year ago 4
@superdomethis If you have a test case that illustrates any problem with any project, take it to the relevant forum where all of the teams and communities will be more than happy to help.
MarkCLittle 1 year ago
This was tested with the JBoss EAP v5 and still failed. Does RedHat have a fix for this or not?
superdomethis 1 year ago 5
Soooo, lemme get this straight...
1) You are using the community version - not the commercially supported, enterprise version.
2) You have not engaged JBoss/Red Hat or even the community for any assistance.
3) You haven't posted the code which you claim is failing for public review.
4) You're pimping a competitive product.
Have a missed anything?
Seriously - this kind of blatantly manufactured FUD makes IBM look *so* bad. It's childish.
tcameron1969 1 year ago 2
Why do you say this is FUD? Did you try this? Can you post a fix to this "bug"?
The Enterprise version was tested in March 2010 and failed this test. The extensive community support was utilized. The code is simple - just update two databases via JDBC or EJB using XA JDBC drivers.
superdomethis 1 year ago 8
@tcameron1969 the community edition of Geronimo don't lose the tx. Because is free we have to support the errors?
alonsov 1 year ago
Did you bother to raise any questions on the JBoss forums. They are free after all! Also did you make your code publicly available? That might help track down user error versus issues with the application server. Until that happens and the problem is reproducible, this appears as so much FUD.
MarkCLittle 2 years ago 8