| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 339.1 | Find is known and fixed here. Unit Attention is an error. | NABETH::alan | Dr. File System's Home for Wayward Inodes. | Wed May 07 1997 15:24 | 30 |
| re: Find.
Fixed that months ago... As I recall it was a stupid error
on the part of person that did the GUI find implementation.
I would pound his head against the wall for it, but my head
still hurts from when I did that yesterday for another stupid
bug...(*) Look through the closed problems for the full description.
I think you can get it not to hang by defining MRU_WAIT to a
small value.
re: Wildcards.
Don't support them. I'll mention this to our technical writer.
Bear in mind that the MRD Find support doesn't use Send Volume
Tag/Request Volume Element Address, which is the part that does
support wildcards. When the TL820 firmware is fixed, we'll
think about it.
re: Unit Attention.
Okay. If the robot is going to report back errors when we do
commands it seems incumbant upon us to pass the error back
to the user. MRU has a simple error reporting mechanism;
It work or It Didn't Work. Thus, we take a very narrow view
of something which works; no error AT ALL. For us, the Unit
Attention was an error. The IOD on the TL820 does the same
thing.
(*) This is a joke I made yesterday by pounding my own head
against the wall. You had to be there...
|
| 339.2 | Need new version, I guess.. | SUBSYS::TRAN | Straight <Left> Hitter.. | Thu May 08 1997 08:24 | 12 |
|
Thanks, Alan..
I'll get new T1.2 and try it.
RE: Unit Attenion.
This is a SEP SCSI requirement for robot to report 6-28-00 when
external media changes has occured, ie. insert magazine. It's OK
if MRU will report any failure, just want consistency here. As a
user I can deal with that.
T.
|
| 339.3 | | SSDEVO::ROLLOW | Dr. File System's Home for Wayward Inodes. | Thu May 08 1997 08:45 | 4 |
| Unless you want to pick up the current build piece by piece
there isn't a more current OpenVMS kit. If you don't mind
a piece-meal build, let me know via mail and we can work
something out.
|
| 339.4 | | SSDEVO::ROLLOW | Dr. File System's Home for Wayward Inodes. | Thu May 08 1997 08:50 | 8 |
| re: Check Condition
We are consistent. If an underlying SCSI command completes in
such a way that we feel compelled to do a Request Sense, we
treat that command as having failed. If a Digital UNIX I/O
control fails with EIO set, we do Request Sense. If an OpenVMS
$QIO fails with SS$DRVERR or the GK driver tells us there was
a SCSI error we do a Request Sense.
|