Hi folks,
We have Axapta 2.5 SP4 running on two AOS machines with most users
connecting as 3-tier thin client. Recently we found that records
inserted, updated or deleted on tables set to EntireTable caching
method from one of the servers did not get replicated to the other by
either the usual 60-second synchronisation or the synchronise done at
midnight, as described in the document "Improved Caching.doc" found on
the Axapta 2.5 installation CD.
In order to ensure that things like the financial stuff didn't get into
a mess we shut down one of the AOS' in the production environment but
began testing the issue on the test environment with two AOS'.
Initially we were seeing exactly the same behaviour even though the
test AOS' are on physically different machines from the production ones
and the SQL Server used for the test setup is different from the
production database server as well so to all intents and purposes it
looked very much like a problem with the Axapta software.
We tried setting the "Max cache sync time" setting to 20 seconds as I
discovered on these newsgroups but that didn't work on the test
environment so I didn't attempt it on the live one. Basically while
revisiting this yesterday I noticed that the cache synchronisation on
the test environment was working again. I brought the second
production server back up and despite not having made any changes to
remedy this problem, the cache syncing seems to be working on there now
as well.
Anyone got any ideas about this? Is the synchronisation mechanism for
EntireTable caching known to be temperamental like this? Are there
other factors which can affect this? My superiors are obviously happy
that it's working again but we need to understand why it stops and if
there is anything we can do to sort it out if it breaks again in future?
|