ioc re-connect problem (additional info)

[Matt Bickley <bickley@cebaf.gov>]

> --- Begin Message ---
> The communications failure was between databases, not involving sequences.
> In particular, the problem was in the task named "dbCaInput", which manages
> database CA input links on the host IOC.  Another solution (which wasn't
> widely known at the time of the failure Chip referred to) was to delete the
> "dbCaInput" task and respawn it.  When the dbCaInput restart solution was
> applied, roughly 20 of our IOCs had suffered the communications failure
> simultaneously.
>
> > Here's a typical problem at Jefferson Lab (taken from log):
> > 
> >  The global sequence on iocin1 claimed the signals from iocin3 and iocwa4
> >  were monitored but not connected even though the iocs were indeed up and
> >  the signals reachable by caget.  iocin3 and iocwa4 also were not providing
> >  signals to iocin1.  The solution was to reboot iocin1 to force the se-
> >  quence to re-establish connections to various viewer iocs. Rebooting the
> >  data source (iocin3 or iocwa4) did not fix the problem.
> > 
> > Anyone have any suggestions for how to find or fix this?
> > 
> > Regards,
> > 
> > Chip
>
> ----
> Matthew Bickley                                       Email: bickley@jlab.org
> Computer Scientist                                   Voice mail: 804-249-7347
> TJNAF
>
>
> --- End Message ---