Re: Spurious "Channel write request failed" messages

[Ralph Lange <ralph.lange@gmx.de>]

Hi Michael,

That erroneous behavior was described in http://www.aps.anl.gov/epics/core-talk/2017/msg00503.php, and has been fixed on the 3.16 branch.

Alas, there's no bug in the issue tracker for it, and it is not mentioned on the "Known Issues" page. We should probably add the latter at least.

Thanks for pointing this out!

~Ralph

On Sat, Sep 23, 2017 at 11:47 AM, Michael Ritzert <michael.ritzert@ziti.uni-heidelberg.de> wrote:

Hi all,

since upgrading from 3.14 to 3.16.1, I see the following messages on the console of an IOC:
DB CA Link Exception: channel "dest1"
DB CA Link Exception:  native  T=DBR_DOUBLE, request T=DBR_DOUBLE, native N=1, request N=1,  access rights {RW}
DB CA Link Exception: "Channel write request failed", context "dest1"

This is from a dfanout record that simply pushes a value to another IOC.

However, the write succeeds.

It's totally reproducible with these two DBs run in softIOCs:

sender.db:
record(calc, "count") {
        field(SCAN, "1 second")
        field(VAL, "0")
        field(PINI, "YES")
        field(INPA, "count")
        field(CALC, "A+1")
}

record(dfanout, "send1") {
        field(DOL,  "count CPP")
        field(OMSL, "closed_loop")
        field(OUTA, "dest1")
}

record(dfanout, "send2") {
        field(OMSL, "supervisory")
        field(OUTA, "dest2")
}

receiver.db:
record(ao, "dest1") {
}

record(ao, "dest2") {
}

There's one messages every second about dest1, and one on every write to send2.

Cheers,
Michael
--
Dr. Michael Ritzert                     Tel: +49 621 181 2883
Schaltungstechnik und Simulation        Fax: +49 621 181 2734
Technische Informatik, Uni Heidelberg   michael.ritzert@ziti.uni-heidelberg.de
68131 Mannheim, Germany                 http://sus.ziti.uni-heidelberg.de