Ex­tremeTrans­ac­tions 5.0.86

180713: Re­fac­tor LogCloud to al­low mul­ti­ple in­stances that share the same DB schema.

You can now share the same DBMS for dif­fer­ent mi­croser­vices - so you don't need to in­stan­ti­ate mul­ti­ple DB in­stances.

To up­grade, it is best to do a clean shut­down (in no-force mode) so you are sure that there are no pend­ing trans­ac­tions.

The fol­low­ing schema changes are re­quired on your end:

Add a col­umn OWNING_RECOVERY_DOMAIN_NAME to all ta­bles

The de­fault val­ue should be what­ev­er you have for com.atom­ikos.icatch.tm_u­nique_­name in your mi­croser­vice.

Table LOG_CLOUD_STARTUP_CONFIG: com.atom­ikos.icatch.tm_u­nique_­name no longer a sep­a­rate row

The prop­er­ty val­ue com.atom­ikos.icatch.tm_u­nique_­name is now im­plic­it in part of the pri­ma­ry key columns. So you don't need to in­sert this as a sep­a­rate row any more.

180766: Re­mot­ing re­cov­ery: im­prove log­ging if getOut­come fails

Made the log files and stack traces less con­fus­ing when res­o­lu­tion of a pend­ing in-doubt trans­ac­tion fails due to con­nec­tiv­i­ty is­sues.

180765: NPE in re­mot­ing com­mit re­play if co­or­di­na­tor / root not found

Fixed a Nul­lPoin­t­erEx­cep­tion with re­mot­ing re­cov­ery to­wards a re­mote par­tic­i­pant that can­not find the trans­ac­tion in ques­tion.

180060: Im­prove log warn­ing on time­out be­fore com­mit

When a trans­ac­tion timed out we used to hint at dif­fer­ent op­tions / caus­es in the warn­ing. This has been im­proved to be ac­cu­rate now.

180154: Bug in LogCloud re­cov­ery: max_­time­out in ap­pli­ca­tion node must equal that in LogCloud

Fixed a bug that could cause prob­lems in re­cov­ery when mul­ti­ple nodes com­pete for the re­cov­ery mas­ter role.

180730: Bug in HTTP re­cov­ery

Fixed a bug where some pend­ing HTTP re­mote par­tic­i­pants were not re­cov­ered cor­rect­ly.

Only for cus­tomers. Not a cus­tomer yet?