Conference ssdevo::hsz40_product

    Dilx can only be run from terminal, not hszterm, which is what the
    error message appears to be?
    
    

	The Digital UNIX version of hszterm (when I last worked on it)
	had two lists of commands; those that it would NEVER run and
	those that it could be coerced into running.  How to add commands
	to the 2nd list is documented somewhere, or was when I turned
	the distribution over to someone else...

Are these controlers a dual pair? I have found running dilx is only
possible from the controller serial port on which the JBOD unit is
currently preferred to. (can't remember the message as I am at home
but it sounds familiar)

dilx needs a JBOD disk with a unit defined. Defining a unit 'binds'
the JBOD to a preferred controller in the pair and dilx can only
test units on the controller on which it is executing.

I am not sure, but you may also be prevented if the unit is reserved
by a host. (SCSI reservation - show unit full) 

Phil
[Posted by WWW Notes gateway]

    I have two sites that both report the same symptoms. The disks are all
    configured correctly (Or as correctly as we can see) and the HSZ50s are
    singular ones. The system sees the disks and you can do everything you
    would normally do with them. However DILX cannot be run on them as it
    consistantly says there are no disks available for testing. The systems
    have even been shut down to ensure that there is no conflict and the
    drives are reported as being AVAILABLE. The are no bad states reported.
    
    The second anomaly is that despite the fact the controller is singular
    and it does not have preferred targets set the disk drives all show
    a state of PREFERRED PATH THIS CONTROLLER. When trying to clear this
    state the command is rejected. If you type "SET <unit> ?" the commands
    given do not include the NOPREFERRED_PATH. This part is very worrying
    as in the V5.1 Release Notes is says that if a dual-redundant
    controller has a battery failure is will crash but the units wont
    become available on the other controller if preferred pathing is set
    and this must be cleared to get them out of mount verify. How is this
    achieved without a working command?
    
    The controllers are both new installs running V5.1 HSOF.
    
    Any comments would be appreciated as we start to raise an IPMT.
    
    Regards
    
    		Keith Shergold

>                                              When trying to clear this
>   state the command is rejected. If you type "SET <unit> ?" the commands
>   given do not include the NOPREFERRED_PATH.

    I  suspect  that  HSOF  is  being  smart;  since  you  only have one 
    controller, all spindles are preferred  to  it  (no  point  allowing 
    anyone to set preferred pathing on a non-redundant coontroller!!) ;-)

>    This part is very worrying
>    as in the V5.1 Release Notes is says that if a dual-redundant
>    controller has a battery failure is will crash but the units wont
>    become available on the other controller if preferred pathing is set
>    and this must be cleared to get them out of mount verify. How is this
>    achieved without a working command?

    Your  test is with a non-redundant controller.....get it a buddy and 
    try the `test' again....betcha it'll work!!

>    Any comments would be appreciated as we start to raise an IPMT.

    Don't raise an IPMT case.

    More information on .4
    
    I managed to get access to an HSZ50 with (As luck would have it) bothe
    V5.0 and V5.1. I set up the disks under V5.1 and what do knoe;;; DILX
    failed to find any disks to test. I the shut it down, inserted the V5.0
    PCMCIA and checked the state of the disks. All being well I ran DILX
    and guess what....you've got it; DILX worked. It did nothing but
    exchange the cards. I went back to V5.1 doing a simple change of cards
    and DILX still refused to find any testable unit. I have therefore come
    to the conclusion that V5.1 DILX is bust. 
    
    Is this a known problem? Is this a bad batch? Should we do something
    mysterious to get DILX to run under V5.1?
    
    Sorry, Chris, I haven't checked the other part of the problem as we do
    not appear to have two copies of either version.
    
    Keith

    More on .4
    
    I managed to get my hands on an HSZ50 and V5.0 and V5.1 HSOF. I set the
    disks up from scratch under V5.1 and all was well until I ran DILX and
    like the two sites I am dealing with it reported that there were no
    disks available for testing. I then shut it down and brought it back up
    with V5.0 all was well again so I ran DILX and low and behold it ran
    with all disks; no questions asked. I then re-installed V5.1 and it
    just did not want to know. My conclusion therefore is that DILX is bust
    under V5.1
    
    Is this a known problem? Is it a batch issue. Do I raise an IPMT?
    
    I could not progress the second part of the problem as we do not have
    two PCMCIA cards with the same version of HSOF on them.
    
    Regards
    		Keith Shergold