| T.R | Title | User | Personal Name
 | Date | Lines | 
|---|
| 1277.1 | Some gotchas | AIMTEC::WICKS_A | It wasn't supposed to end this way | Thu Aug 20 1992 17:22 | 10 | 
|  |     Yukiji,
    
    Make sure OA$PRIMARY_NODE is defined with the '::' on the end and in
    EXEC mode. Also ensure that the password in UAPASSWRD is in uppercase.
    
    Apart from that I'd have to see the entries.
    
    Regards,
    
    Andrew.D.Wicks
 | 
| 1277.2 | Thanks, but colons are there. | EWBV51::NAGURA | Yukiji Nagura /ALL-IN-1 CSC/Tokyo/Japan | Fri Aug 21 1992 04:52 | 14 | 
|  | 	Andrew,
>    Make sure OA$PRIMARY_NODE is defined with the '::' on the end and in
>    EXEC mode. Also ensure that the password in UAPASSWRD is in uppercase.
    
	Parhaps you mean SYS$NODE or SYS$CLUSTER_NODE. Yes, the logical
     has defined with trailing colons. In A1V24START.COM, OA$PRIMARY_NODE
     is to be defined as SYS$NODE minus two colons.
	BTW, Next week I'm going to attend to the Advanced Techinical Seminar,
     do you happen to be there, Atlanta's most --famous Welshman ? 
							Yukiji
		
 | 
| 1277.4 | my two-penny's worth... | WAYOUT::CLARKE | The Cat in the Hat comes back. | Wed Aug 26 1992 14:29 | 20 | 
|  | Are you talking about the infamous;
'User agent object is unkown in the mail directory'
I spent the best part of a day tearing my hair out over one of these, turned
out to be that the UAPASSWRD data file was sequential rather than indexed (shown
by a dir/full).
Fix was to recreate file using <CREATE UAPASSWRD from within ALL-IN-1 then 
re-adopt mail directory.
Anther common gotcha with this type of problem are duplicate records in the 
UAPASSWRD data file. Do <FORM UAPASSWRD and press find should only be one record.
Sometimes a blank record may occur and it this is the case you will have to 
recreate file as above.
Another one for my Stars article to write list!
Aston
UK CSC
 | 
| 1277.5 |  | EWBV51::NAGURA | Yukiji Nagura /ALL-IN-1 CSC/Tokyo/Japan | Thu Oct 29 1992 02:45 | 9 | 
|  | 
	After a long time of investigation (much more struggle), this problem
      finally got fixed.
	The point was several DUMP_CACHEs individually to SUBSCRIBER, PROFIL
      USERAGENT DSABs. Sometimes we'd got LIB$FREE_VM errors before. So some
      in-memory blocks might be corrupted after USERAGENT_POST.SCP.
							Yukiji
 |