In a non-clustered network of machines, each machine
having CFAS running, that uses a single dsn for client var
storage - are CFIDs uniquely assigned for the storage db,
or is each CFAS assigning CFIDs independently of each
other?
In both CDATA and CGLOBAL the field cfid actually
stores the cfid-cftoken values as xx-xxxxxxxx.
Is there any guarantee, in the above described
network, that each cfid in CDATA and CGLOBAL
actually points back to the originating server?
Is it possible for serverA to have created the same
cfid-cftoken pair as serverB and thus any entries
in the shared clientVar storage db might not uniquely
refer to different clients on the serveral servers?
CDATA is indexed on multiple columns (cfid,app), but
CGLOBALS has two indices ((cfid),(lvist)).
If clientP is using App_1 on ServerA and
clientQ is using App_1 on ServerB,
how likely is it, if at all, that the entries
in CDATA/CGLOBAL are conflated?
I'm thinking they could be commingled.
The CFID element seems to be a simply incremented
value each CFAS generates as need. The CFTOKEN
is not a simply incremented value - don't know how they
are generated - this may resolve the potential problem,
but I don't see how the different CFASs could inform
each other which CFTOKENs have been assigned.
???
Pan
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Structure your ColdFusion code with Fusebox. Get the official book at
http://www.fusionauthority.com/bkinfo.cfm
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists
