(See SLDPM conference Item 28)
Status: Believed fixed....please report if any more occurrences seen.
We think this has been tracked down to a change that occurred when we switched silo control from VM to UNIX. The code on VM which emulated the interface to the silo was not behaving in quite the same way as the old code, in particular SETUP was granting access to a TAPE even when it was "in transit". Bill Weeks has changed the code on VM to workaround this.
Status: Understood, fix in progress
This turned up two bugs. The first was that the exit handler in CARTCMD was not working on the Alpha. This has now been fixed (fix is in DEV CART)...but not installed in PROD yet as it will make things worse until second part of fix is ready.
The second problem was in JIOFTO. Len is working on a fix for this (In the case where a new tape is mounted after a previous TCLOSE the flag which indicates that Jazelle should not dismount the tape needs to be copied from the previous TCLOSED file to the newly opened file).
Status: Waiting for feedback from DEC
New code in the SLD staging system has added time-stamps when this error occurs. This enabled Charlie to associate this error with an informational message in the VMS error log. This new information has been sent to DEC.
Status: Unknown
Occasionally after a CART MOUNT control never returns to IDA. This can sometimes be worked around by Charlie dismounting and re-mounting the tape from UNIX.
Status: unknown
The DCSC server has crashed several times in recent weeks with the message:
%DCSC-E-ILLOGICAL_EVENT, Illogical event/state combination
Status: Ongoing
Not very scientific evidence suggests tape throughput is lower that it used to be. Log files do not show heavy silo usage from UNIX (at least for last two days) so this is unlikely to be the problem. Need to investigate tape throughput and compare to number obtained after tune-up of staging system last year.