OAK SCSI 4, User Manual

Page: 49 / 54

Power up delays
To cope sensibly with drives spinning up, without excessive timeout
periods for real errors, the following is done
The first disc operation after a hard reset, which will nearly always be a
mount operation on the default drive, will retry for up to a configurable
amount of time if any error is reported (e.g. SCSI timeout or not ready
errors).
This copes with the differences in response to non-ready drives with all
Oak SCSI drives.
The timeout period is held in EEProm, and the default value is 20
seconds. The optimum value is a couple of seconds longer than the
spin-up/initialise time of the slowest drive in the system.
This avoids delays in reporting subsequent errors, or attempts to mount
non-existent additional drives, in particular when doing *Bye, as
FileCore Mounts drives before parking them (as it has to find the park
address from the disc record).
The only possible problem with this strategy is where there are
secondary drives which take longer to spin up than the default drive -
attempts to access them whilst spinning up will result in errors, and as
FileCore remembers errors when mounting drives, any subsequent
attempt to access the drive will return the same error (e.g. drive not
ready) that was obtained from the initial attempt to mount, despite the
fact that the drive is now ready. The only way round this is to either do
a hard reset, or *RMReinit SCSI.
Page 42