Minutes of the Feb-18-1997 conference call on Stardust Interoperability
test results:
Present: Jay Martin
Bob Pentecost
Lloyd Young/Evan/Matt
Gail Songer
Ron Bergman
Peter Zeller
Stuart Rawley
Chuck Adams
Harry Lewis
Review of MIB WALK
hrDeviceID - 2 of 6 vendors returned 0.0 rather than a device OID from the
private enterprise. Jay emphasized that, if drivers are to self configure
based on the SNMP information, the hrDeviceID needs to be unique to the
printer model. One way to think of it is if a different PPD (Postscript)
file would be required than the device should have a different OID. OID's
should not be so granular as to represent configurations which can be
discerned from the Printer MIB itself, such as how many input trays are
installed.
hrDeviceDesc - Serial and Parallel are considered optional interfaces
since the Printer MIB is largely network oriented. 1 or 6 vendors
described their Serial and Parallel interfaces.
hrDeviceStatus - One vendor indicated "down" because the interface "slot"
existed (and presumably was pre-configured for a particular NIC card)
but the card was not installed. Matt will clarify that subunit status
for hrDeviceStatus should not indicate down if the interface is not
installed.
hrDevice - Do we need a table for storage if we have a disk? Only one
vendor tested with a hard disk installed. Can the storage table be
indexed to the printer or should it be indexed to a separate device
(storage) type? Unresolved. Perhaps vendors should retest, individually,
and report what they do for HDD?
hrPrinterDetectedErrorState - One vendor appears to have reported 4
bytes while the rest reported 1 byte. One vendor used a different tool
however. Vendors should individually determine if they send 1 or 4 bytes.
hrSystemDate - Wide differences. No realtime clock. Not useful. Two
printers have exact same value for hrSystemDate and hrInitialLoadParms.
One vendor's value is nontrivial. Unexplained.
hrSysUpTime - Half did it. System group has sysUpTime. Used by applications
to correlate Power Recycles with config change counter.
prtChannelCurrentJobCntLangIndex and prtChannelDefaultPageDescLangIndex -
Since these are indices (not enums), what if you don't have a current
job control language (possible) or a default PDL (hard to imagine).
Matt will clarify that (0) means don't have.
prtChannelIfIndex - Most implementations can't distinguish between
two NICs in terms of ifIndex. At least one printer can. It was decided
to accept this and let it pass. It's not clear if printers that don't
distinguish NICs with separate indexes actually report only the services
for one NIC or if they lump all the services together.
prtChannelStatus - Some of the definitions for subUnitStatus are pretty
ambiguous... like Idle vs Standby and Active vs Busy. Some printers assume
different Available or Unavailable states under, otherwise, similar conditions.
Chuck will try to clarify. One option is to indicate 0, 4 and 6 are all
acceptable states when the printer is Available.
prtGeneralConfigChanges - Wide variance here. Need more examples. What
actually constitutes a config change? Matt to clarify.
prtInputStatus - Wide variance (if taken at face value). What if the
Input subUnit is available but the whole printer is in standby? In general,
subUnit status is part of the Alert top 25 project and this topic should
be addressed in a separate call.
Next call will take up with prtInputDefaultCharSet.
Harry Lewis - IBM Printing Systems