The vast majority of what�s in JRun Updater 3 is what was made available to CF customers in CFMX 6.1
Updater 3 gets these fixes to JRun 4 customers.
There is a some small number items fixed in Updater 3 that aren�t in CFMX 6.1. We're working on a cumulative hotfix due in April to make those available for CFMX 6.1 Server-configuration customers.
Some other fixes available ***POST-U3*** (didn't make U3) for JRun and CFMX customers through hotfixes + other random thoughts I'll share:
(note: all of these will be in the 6.1 cumulative hotfix due in April)
- we expect to release new datadirect drivers soon to fix the oracle and sqlserver issues customers have reported (no more data...., arrayindexoutofbounds, hangs) with the 3.2 drivers shipped with JRun 4 Updater 2 and CFMX 6.1. They are still in testing mode. JRun4 Updater 3 got the 3.1+ drivers since they were the most stable. 3.1+ was also made available as a hotfix for CFMX customers with issues. We expect to replace the 3.1+ for JRun and CFMX customers with newer drivers in the next few weeks if testing goes well.
Please email me if you'd like to participate in the beta.
- some connector enhancements - including failover in clustering working as expected (we will have a hotfix for this soon - please email me). These fixes are long overdue.
- connection pooling fixes for "JRun datasources" to match the pooling fixes for ColdFusion datasources (see technote below).
http://www.macromedia.com/support/coldfusion/ts/documents/conn_pool_hotfix.htm
(for apps like AbleCommerce that use "JRun datasources" ie entered through the JMC and available to all instances). Users of JRun J2EE CFMX versions would need these if they have connection-limited datasources *AND* the underlying application used JRun datasources.
This fix will be made available in a few days.
- Client variable purge fixes for SQL and registry datasources (there were 4 bugs fixed). This often showed up when the server hung every hour on the hour for the maintenance task of purging client variables older than [N] days. This hits many customers if they use "registry" and the registry grows to say 50,000 or 100,000+ entries. SQL datasources are recommended but even if users switched, the registry purges still occurred (in addition to SQL purges) and could hang the system. The fix now respects the unchecked "purge" setting for all client datasource types.
These are the ones that come to mind that get reported most often hanging systems.
***For Dick AppleBaum***:
1) JRun and, therefore, CF use the JVM 1.3 instead of 1.4 (even with
the fix posted Christian Cantrell's blog site)
> Part of this is the installer not allowing the selection. There are also issues with using JDK 1.4 (jvm.cfg issues). We've fixed this post-Updater 3 in the launcher. (jrun stub)
Please email me and I can get you some code to test.
2) When starting a CF 6.1 JRun server instance, the scheduler does not
start (see below).
> I've heard of this sporadically but can't reproduce it. Perhaps send me your neo-cron.xml and jrun.xml and I'll see if I can repro it on my machine. 6.1 definitely massaged this .xml file on an existing server. Perhaps the one in your new instance got corrupted somehow.
Regards,
Stephen Dupre
Macromedia QA
[EMAIL PROTECTED]
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
