This should work as you suggest.
In the xxx.db file set TPRO true for both A and B.
At initialization does B get processed before A?
If so then B is processed for some other reason than A linking to it
a) PINI is TRUE
b) SCAN is not passive
c) Another link references B. Issue the command:
dblsr "<B name>",2
J. Frederick Bartlett wrote:
I have encountered a situation in which spurious (i.e. unexpected) link
alarms are being generated following the reboot of an IOC. The record
Record A has a scan rate of one second
Record A has a forward link to record B
Record B has an input link in its INP field that references record A with
and "MS" specified.
When record A is processed for the first time, record B registers an
"INVALID" alarm severity with a "LINK" alarm status. The second time that A
is processed, the "INVALID" alarm severity is cleared.
I believe that, when record B retrieves the value from record A during the
initial process cycle, the alarm severity that is propagated by the "MS"
specification is "INVALID" because record A must, somehow, retain the
original "INVALID" severity and the "UDF" status even though it has already
processed. During subsequent process cycles, record B does not generate an
"LINK" alarm status because record A no longer has "UDF" alarm status.
This behavior is counter-intuitive. The "Application Developer's Guide"
states that the forward link is not executed until the originating record
has completed processing. If this is the case, why does an input link to
that record transfer the prior (or initial) alarm state?
Finally, we are still using release 3.13.4.
- RE: Spurious link alarms generated? J. Frederick Bartlett
- Spurious link alarms generated? J. Frederick Bartlett
- Navigate by Date:
Insertion Devices Control and EPICS Rok Sabjan
Re: Insertion Devices Control and EPICS Juraj Krempasky
- Navigate by Thread:
Spurious link alarms generated? J. Frederick Bartlett
RE: Spurious link alarms generated? J. Frederick Bartlett