[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference ssdevo::hsz40_product

Title:HSZ40 Product Conference
Moderator:SSDEVO::EDMONDS
Created:Mon Apr 11 1994
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:902
Total number of notes:3319

740.0. "UNIX 3.2f and 3.0-2 OK ?" by BACHUS::WAUTERS (Note written by a human being) Tue Jan 21 1997 07:04

T.RTitleUserPersonal
Name		DateLines
740.1SSDEVO::T_GONZALESFri Jan 24 1997 19:153
    use 3.2g 3.2f not qualified with hsz40
740.2ThanksBACHUS::WAUTERSNote written by a human beingMon Jan 27 1997 04:293
    Thanks
    
    Philippe
740.3GEM::SHERGOLDA FOOL'S living PARADISESun Feb 02 1997 06:145
    As for the upgrade from V2.7 to V3.0-2 there are problems with it,
    especially if you have WRITE BACK CACHE and/or devices with
    SAVE-CONFIGURATION on them. I would keep an eye on IPMT case CFS.48075.
    
    Keith
740.4Unable to go to 3.0-2 under DECsafe ?BACHUS::WAUTERSNote written by a human beingThu Feb 06 1997 05:2128
    Here are the results of the upgrades :
    
    	- we have a total of 8 HSZ40's connected to 5 different systems
    	- all the HSZ's are configured in the same way :
    		- no raid 0, 1, 0+1 or 5 are used
    		- writeback cache is enabled on all disks
    		- the configuration is saved on the first disk of every
    		  BA356.
    
    We had no problem to upgrade 4 HSZ40's : procedure nicely written,
    fast, easy. A dream !
    
    Then we had to upgrade the 4 HSZ's on the production systems and we got
    tons of errors when the systems rebooted. Errors were "Illegal request
    or _CDB parameter". The disks were seen by the system on the console (
    >>> show device ) but were unusable when UNIX started. The major thing
    that differs between the production systems and the other ones is that
    they are installed in a DECsafe (ASE) cluster. I checked the ducumented
    revisions before the upgrade and I think that everything is OK (
    KZPSA's are at revision A10 by example ).
    
    So, we had to turn back to 2.7 ! This is a real pity because the
    production cluster cannot be stopped easily.
    
    A call is logged in our local subsidiary.
    
    
    Philippe